Accelerating food production in Sub-Saharan Africa

Material Information

Accelerating food production in Sub-Saharan Africa
Mellor, John Williams, 1928-
Delgado, Christopher L
Blackie, M. J
University of Zimbabwe -- Dept. of Land Management
International Food Policy Research Institute
Place of Publication:
Published for the International Food Policy Research Institute by Johns Hopkins University Press
Publication Date:
Physical Description:
xix, 417 p. : ; 24 cm.


Subjects / Keywords:
Agricultural productivity -- Congresses -- Africa, Sub-Saharan ( lcsh )
Food supply -- Congresses -- Africa, Sub-Saharan ( lcsh )
Agriculture and state -- Congresses -- Africa, Sub-Saharan ( lcsh )
Landbouw ( gtt )
Voedselproductie ( gtt )
Politieke aspecten ( gtt )
Agriculture -- Productivité -- Congrès -- Afrique noire ( rvm )
Aliments -- Approvisionnement -- Congrès -- Afrique noire ( rvm )
Politique agricole -- Congrès -- Afrique noire ( rvm )
Africa south of the Sahara -- Agricultural industries -- Productivity -- Improvement
bibliography ( marcgt )
conference publication ( marcgt )
non-fiction ( marcgt )


Bibliography: p. 377-396.
General Note:
Papers and commentaries from a conference sponsored by the Dept. of Land Management of the University of Zimbabwe and the International Food Policy Research Institute, held at Victoria Falls, Zimbabwe in Aug. 1983.
General Note:
Includes index.
Statement of Responsibility:
edited by John W. Mellor, Christopher L. Delgado, Malcolm J. Blackie.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
13581540 ( OCLC )
86010684 ( LCCN )
0801833906 (alk. paper) ( ISBN )

Full Text


in Sub



Accelerating Food
Production in
Sub-Saharan Africa

Other Books Published in Cooperation with
The International Food Policy Research Institute:

Agricultural Change and Rural Poverty: Variations on a Theme
by Dharm Narain
Edited by John W. Mellor and Gunvant M. Desai

Crop Insurance for Agricultural Development: Issues
and Experience
Edited by Peter Hazell, Carlos Pomareda, and Alberto Vald6s

Accelerating Food
Production in
Sub-Saharan Africa

Edited by
John W. Mellor
Christopher L. Delgado
Malcolm J. Blackie

Published for the International Food Policy Research Institute
The Johns Hopkins University Press
Baltimore and London

1987 The International Food Policy Research Institute
All rights reserved. Published 1987
Printed in the United States of America

Second printing, hardcover, 1988

The Johns Hopkins University Press, 701 West 40th Street, Baltimore, Maryland 21211
The Johns Hopkins Press Ltd., London

coThe paper used in this publication meets the minimum requirements of American
National Standard for Information Sciences-Permanence of Paper for Printed Library
Materials, ANSI Z39.48-1984.

Library of Congress Cataloging-in-Publication Data

Accelerating food production in Sub-Saharan Africa.
Papers and commentaries from a conference sponsored by the Dept. of Land
Management of the University of Zimbabwe and the International Food Policy
Research Institute, held at Victoria Falls, Zimbabwe in Aug. 1983.
Bibliography: p.
Includes index.
1. Agricultural productivity-Africa, Sub-Saharan-Congresses. 2. Food
supply-Africa, Sub-Saharan-Congresses. 3. Agriculture and state-Africa, Sub-
Saharan-Congresses. I. Mellor, John Williams, 1928- II. Delgado,
Christopher L. III. Blackie, M.J. IV. University of Zimbabwe. Dept. of Land
Management. V. International Food Policy Research Institute.
HD2117.A33 1987 338.1'9'67 86-10684
ISBN 0-8018-3390-6 (alk. paper)


List of Tables and Figure ix
Foreword Denis R. Norman xiii
Preface xvii

Part I
Defining the Food and Nutrition Problem 1

chapter 1
Strategic Issues in Food Production in Sub-Saharan Africa
Christopher L. Delgado, John W. Mellor, and
Malcolm J. Blackie 3
chapter 2
The Evolving Food Situation
Leonardo A. Paulino 23
chapter 3
The Nutrition Situation and Its Food Policy Links
Shubh K. Kumar 39
chapter 4
Commentary on Defining the Problem
Alexander P. Vamoer 53

Part II
Potential and Practice in Food Production Technology
Development 57

chapter 5
The West African Semiarid Tropics
Peter J. Matlon 59
chapter 6
Eastern and Southern Africa
Michael Collinson 78


chapter 7
The Humid and Subhumid Tropics
Coenrad H. H. ter Kuile 97
chapter 8
Identifying Technological Potentials
Frits W. T. Penning de Vries and Cornelis T de Wit 109
chapter 9
Technological Change and Agricultural Labor Use
Christopher L. Delgado and Chandrashekhar G. Ranade 118
chapter 10
Women's Role and Agricultural Technology
Shubh K. Kumar 135
chapter 11
Development and Extension of Agricultural Production Technology
Guy Vallaeys, Pierre Silvestre, Malcolm J. Blackie, and
Christopher L. Delgado 148
chapter 12
Commentaries on Technology
T. Ajibola Taylor, Robbie M. Mupawose, and
Michael Collinson 161

Part III
Support Systems for Agricultural Development 171

chapter 13
Input and Output Marketing Systems: A Nigerian Case
S. O. Olayide and Francis S. Idachaba 173
chapter 14
Restructuring Marketing Systems for Smallholders: Cases in Zimbabwe
Malcolm J. Blackie 187
chapter 15
Commentaries on Marketing Systems
Gunvant M. Desai, Abiodun O. Falusi, Michael Schluter, and
Robinson L. Gapare 199
chapter 16
Agriculture and Central Physical Grid Infrastructure
Michael Lipton 210
chapter 17
Commentaries on Infrastructure
Sudhir Wanmali and Francis S. Idachaba 227


Part IV
Food Policy in the Context of National Development Strategies

chapter 18
Growth Strategy and the Agricultural Sector
Ojetunji Aboyade 241
chapter 19
Commentary on Growth Strategy
Anne 0. Krueger 253
chapter 20
Food Policy and the Choice of Trade Regime
T. Ademola Oyejide 257
chapter 21
Commentaries on the Trade Regime
Achi Atsain and Tshikala B. Tshibaka 274
chapter 22
Food Price Policy and Equity
Dharam Ghai and Lawrence Smith 278
chapter 23
Commentaries on Price Policy and Equity
Carl K. Eicher and Dunstan S. C. Spencer 290
chapter 24
Domestic Political Considerations for Food Policy
Henry Bienen 296
chapter 25
Commentaries on Political Considerations, Nationalism, and Development
Benson Kabeta Muleya, Philip Ndegwa, and
Kazushi Ohkawa 309

Part V
Donor Assistance Strategies 319

chapter 26
Growth of Foreign Assistance and Its Impact on Agriculture
Uma Lele 321
chapter 27
Commentaries on Foreign Assistance
Moise C. Mensah, Anne de Lattre, and Bruce F. Johnston 343

viii Contents

Part VI
Conclusions 351

chapter 28
Priorities for Accelerating Food Production Growth in Sub-Saharan Africa
John W. Mellor, Christopher L. Delgado, and
Malcolm J. Blackie 353

References 377
Contributors 397
Index 405

Tables and Figure
p 0


2.1 Population and food production in sub-Saharan Africa 24
2.2 Population growth and agricultural labor force in sub-Saharan Africa,
various periods, 1961-2000 26
2.3 Production and area harvested of major food crops in sub-Saharan Af-
rica, various periods, 1961-80 28
2.4 Distribution of major food crops, 1961-65, and production increase,
1961-65 to 1976-80, in sub-Saharan Africa 29
2.5 Average annual growth rate of domestic utilization of major food crops
in sub-Saharan Africa, 1966-80 31
2.6 Trends in exports and imports of basic foods in sub-Saharan Africa,
1966-70 to 1976-80 32
2.7 Production and consumption of major food crops in sub-Saharan Africa,
1976-80 and projections to 2000 34
2.8 Cropland area in relation to selected indicators in sub-Saharan Africa
and selected non-African countries, 1976-80 36
2.9 Annual growth in arable land, cropland, agricultural labor force, irri-
gated area, fertilizer consumption, and agricultural tractor use in sub-
Saharan Africa, 1961-65 to 1976-80 37
3.1 Preschool child malnutrition in selected African and Asian countries
3.2 Maize purchased by official marketing organizations in Zambia
3.3 Seasonal distribution of weight-to-height ratios as percent of reference
standards, Chipata District, Zambia, 1981-82 48
3.4 Crop area for 1980-81 in relation to month in which previous food
stores were depleted, Chipata District, Zambia 49
3.5 Crops planted in two consecutive years, Chipata District, Zambia
6.1 Monthly long-term rainfall in Machakos, Kenya 79

List of Tables and Figure

6.2 Cattle ownership and relation to cultivated land, crop, and income in
Chibi South, Zimbabwe, 1980-81 83
6.3 Main crops in eastern and southern Africa, 1980 89
6.4 Annual yields, livestock and major crops in Ethiopia, Kenya, Tanzania,
and Zambia 94
7.1 Climatic characteristics of four locations in Nigeria 98
7.2 Daily production and annual yield for phaseolus beans, maize, and
cassava in four locations in Nigeria 99
7.3 Land suitable for growing five crops in four agroecological zones of
tropical Africa 100
9.1 Productivity of labor and land in five areas of sub-Saharan Africa
9.2 Differences in the average product of labor for rice and maize in Africa
and Asia 123
9.3 Rainfall and seasonality of labor input to farming in sub-Saharan Africa
and semiarid India 125
9.4 Distribution of labor input to farming in Burkina Faso and semiarid
India 127
9.5 Increase or decrease in inputs and returns from use of improved tech-
nology for four crops in Nigeria, the Philippines, and India 131

10.1 Effects of technological change on women's role in production of three
crops in four areas of sub-Saharan Africa 142
13.1 Fertilizer delivered by recommended application date for northern
Nigerian agricultural development projects 175
13.2 Subsidies on two farm inputs, Nigeria, 1979-82 182
16.1 Government expenditure in 28 countries of sub-Saharan Africa

17.1 Budget allocation for infrastructural development in Manicaland
Province, Zimbabwe, 1984-89 230
17.2 Road densities by states in Nigeria, 1978-80 235
17.3 Selected rural infrastructures in Borno State, Nigeria, 1979-80
17.4 Selected rural infrastructures in Lagos State, Nigeria, 1979-80
20.1 Structural change in the Nigerian economy, 1960-81
20.2 Movement of sectoral prices in Nigeria, various years, 1973-
81 267
20.3 Nominal and real exchange rates for the Nigerian naira, 1970-80

List of Tables and Figure

20.4 Estimated rates of protection for Nigeria's agricultural products, various
periods, 1960-79 270
21.1 Implicit nominal tax rate to farmers in Zaire, five-year averages,
1971-82 277

22.1 Annual change in real producer prices in the 1970s for three cash crops,
twenty sub-Saharan African countries 284
22.2 Equity effects of rises in food prices on food purchasers, 11 sub-Saharan
African countries 287
25.1 Industrial and agricultural growth in Japan, 1887-1904 and 1904-
1919 316


3.1 Household disposal and retention of maize, Chipata District, Zambia,
1981 45


Denis R. Norman

Sub-Saharan Africa is an enormous region comprising 39 countries with a
combined area of more than 22 million square kilometers. It represents nearly
three-quarters of the African continent and about 15 percent of the land mass of
the entire world. The population in mid-1979 numbered about 344 million,
slightly more than three-quarters of the total for all of Africa and 8 percent of the
global total. During the sixties, the population of sub-Saharan Africa grew 2.5
percent per year, and by the end of the seventies the rate had risen to 2.7
percent. The population growth rate in the industrialized countries during the
seventies was less than 1 percent. By the end of this century, less than a decade
and a half away, the sub-Saharan population will number 639 million, a fright-
ening increase of 86 percent over 1979.
The overriding challenge facing agriculture is the need to increase produc-
tion of food enough to feed our ever-growing population. Achievement of food
security throughout the region will be of paramount importance during the
remaining years of the twentieth century. In addition, agriculture must become
a significant supplier of raw material to urban industry, while at the same time
becoming an important market for urban products. Agriculture must also in-
crease its earnings of foreign exchange, of which there is often a chronic
As might be expected in a region as vast as ours, the climate ranges from
tropical rain forest to arid desert. The question arises as to whether agriculture
can in fact reasonably be expected to fulfill its required role. Fortunately, Africa
has great agricultural potential. A recent report claims that Africa's 51 countries
could feed three times their present population, despite the fact that 47 percent
of the continent's land is unsuitable for growing crops. Since an appreciable
portion of unsuitable land is in the North African desert, the major area of
potential is in the sub-Saharan region.
Nevertheless, the current situation is far from satisfactory. In 1982, 26
countries in Africa faced acute food shortages and nutrition problems. During
the sixties, the continent's food production per person fell 7 percent, and during


the seventies an additional 15 percent. According to the World Bank, agri-
cultural growth in sub-Saharan Africa during the seventies was only 1 percent
per annum in both the oil-exporting and low-income countries. The region's
middle-income countries, on the other hand, achieved a rate of a little more than
2 percent per annum, approximately in line with the growth rate of their rural
population. In only eight countries, which account for less than 15 percent of
the region's population, did per capital agricultural production actually increase
in the 1970s.
In view of the expectation that the urban population will quadruple during
the last quarter of this century, present trends are intolerable and must be
reversed. Another disturbing problem that needs urgent attention is the rapid
depletion of vital topsoil, estimated by the director-general of the Food and
Agriculture Organization of the United Nations to be a frightening six million
hectares per annum.
This volume provides up-to-date definitions of the food and nutrition
problems that face us now and in the medium-term future-clear understanding
is the first major step to the resolution of our difficulties. The book also deals
with the policy needs for promoting growth of food production. The concept of
solving our food shortages by increasing production is the policy of the South-
ern African Development Coordination Conference (SADCC), of which my
country is a member. And it is the approach I personally favor.
The book also examines food policy in the context of national develop-
ment strategies, a topic vital to our future. Though potentially rich, our region is
at present extremely poor. In 1979, per capital gross national product in the
whole area was only $411. Compare this with a GNP per capital of $9,440 in the
countries of the industrialized world. The demand for improved living stan-
dards presses from virtually every quarter, while available resources are ex-
tremely limited. Without exception, each country of the region faces the prob-
lem of determining food policy in the context of its overall national
development strategy.
The second half of the twentieth century has seen the evolution of interna-
tional development aid from the wealthier parts of the world to less-fortunate
regions. This recognizes that all humanity is interdependent and that it is
impractical and intolerable for one segment to enjoy a high standard of material
welfare while the remainder lives in abject poverty and hunger. However, this
aid-welcome as it is-is limited and largely uncoordinated. In some cases,
through misallocation by donor or recipient, it has not been used to best
advantage. Several chapters in this volume, by representatives from both sub-
Saharan Africa and donor countries and agencies, analyze past experiences and
suggest aid strategies for accelerating agricultural growth in our part of the
world. Several stress the point that the aid most urgently required is that which
would assist our region in greater utilization of its agricultural base in order to
increase overall production and lessen dependence on food aid.

Foreword xv

Perhaps the most telling message of the book is the urgent need for
immediate action. Time is not on our side. A permanent solution must be found
now, if we are to halt the rapidly deteriorating situation. These lines from The
Rubaiyat of Omar Khayyam say it most poignantly:
The bird of time has but a little way to fly-
and, lo, the bird is on the wing.


In the face of mounting food crises in sub-Saharan Africa, the Department of
Land Management of the University of Zimbabwe and the International Food
Policy Research Institute brought together 44 scholars and policymakers from
African countries, bilateral donor nations, and international organizations, for
a conference on accelerating food production in Africa. The conference was
held in August 1983 at Victoria Falls, Zimbabwe.
The conference focused on policies to accelerate the rate of food produc-
tion growth in sub-Saharan Africa. It sought to determine the current state of
knowledge with respect to policy needs, identify gaps in this knowledge, and
draw conclusions as to the immediate policy actions required. These elements
define the agenda for research to improve food production policy. Policy re-
search is of particular importance to a region where conditions are different
from much of the rest of the world; where food supply problems are in-
creasingly critical, intractable, and prevalent; and where so relatively little
applied research on food policy has been undertaken.
The papers and commentaries presented at the conference stimulated a
wide-ranging discussion, moderated by senior African policymakers and pol-
icy analysts. The papers grew out of intense prior discussion of the issues by the
conference contributors and their colleagues. Following the conference, the
papers were intensively developed in the light of the discussions and proceed-
ings and through a continuing dialogue with the editors and other colleagues.
Following identification of important gaps in the agenda at the conference,
participants were asked to prepare new papers on the specific food issues
involving labor constraints, women's role in food production, and additional
topics on infrastructure.
The final papers-the chapters herein-provide a wide-ranging, up-to-
date, and varied approach to key issues, which reflect the diverse physical,
cultural, economic, and political environment of Africa. Recognition of this
diversity is essential to grasping the magnitude of the problem of accelerating
food production in Africa and to setting the essential priorities. In Africa,
differences in perception as to developmental needs and the initial conditions
necessary for development are greater than elsewhere because of the wide gaps


in knowledge that characterize the region. Many of these differences are re-
flected in the chapters in this book. Food policy is, of course, the product of
political processes. And the environments of these political actors are subtly
reflected in their points of view. We have attempted to catch these nuances in
this volume.
Our introductory chapter sets forth the key strategic issues as a guide to
placing the individual chapters in perspective. The concluding chapter attempts
to cut through the diverse insights and knowledge gaps and draw conclusions
for immediate action. These conclusions are built solidly on the papers and
commentaries but inevitably reflect our experience, our exposure to the liter-
ature and practice, and our best judgment as to the general conclusions to be
Thus, the introduction and conclusion owe much to the authors of the
papers, but these contributors bear only modest responsibility for the final
product. Policy formulation proceeds and decisions are made, based, at best,
on what is now known. We have tried to assist this process by distillation,
extrapolation, and organization. As researchers, we are naturally interested in
the agenda for future research. The problems will still be with us three years
from now, when much current research should be completed. For this reason
our conclusions are inevitably of a tenuous nature. This must leave pol-
icymakers uneasy-and peihap, more timid than hindsight will justify. But,
today's decisions will assist research in paving the way for better decisions as
the policy process evolves.
The editors owe much to many. Our greatest debt is to the African officials
who took the time from pressing immediate problems to contemplate longer-
term strategies. These cabinet-level officials, civil servants, prominent aca-
demics, and other policy analysts provided professional expertise with an
awareness of the political and economic realities essential to useful definition of
longer-term problems and strategies. In particular, we wish to thank two partici-
pants in the Victoria Falls conference, Denis Norman, who was then minister of
agriculture, and Robbie Mupawose, permanent secretary, ministry of agri-
culture, government of Zimbabwe. Their presence and active participation in
the conference not only set the tone for businesslike discussions but under-
scored the seriousness with which the government of Zimbabwe viewed not
only food policy but its role as host and facilitator of the conference.
We are also grateful to the United Nations Development Program (UNDP)
and most particularly to Bill Mashler for the continued support, both intellec-
tual and financial, so necessary to an evolutionary effort of this type. UNDP
made a major initial grant and then, as the scope of the effort and participation
grew, an important supplementary grant to pay for added travel of participants
from developing countries. A contribution of the government of the
Netherlands made possible additional African participation and the planning
and follow up that made this volume possible.


The staff of the International Food Policy Research Institute and the fac-
ulty of agriculture of the University of Zimbabwe engaged in lengthy discus-
sions that honed the topics. These groups also provided the practical logistical
support so essential to a successful intellectual endeavor. In particular, we wish
to note the assistance provided by Peter Hopkins, Michael Hallam, Sheila
Sheldon, and Michelle MacGregor of the University of Zimbabwe. Particular
thanks on the IFPRI side are due to Robert Bordonaro and Laurie Goldberg. The
staffs of the Victoria Falls Hotel, the University of Zimbabwe, and IFPRI, who
organized the conference deserve much credit for providing the atmosphere
conducive to a complex intellectual effort.
The book has benefited greatly from comments on drafts by a number of
colleagues. Cheryl Christensen deserves special thanks in this regard for
providing a wealth of detailed and highly insightful comments on the entire
manuscript. These proved invaluable to the editors in the revision process. We
would also like to acknowledge the helpful comments of Ray Hopkins and John
Strauss. Michael Lipton and Peter Matlon provided a series of detailed and
highly relevant suggestions for the revisions of specific parts of the manuscript.
Finally, we wish to thank Irene Pereira for her persevering help in keeping
track of massive editorial correspondence and in typing endless manuscripts,
Wayne Dexter for invaluable assistance in shortening and clarifying the prod-
uct, Lien-Huong Tran and Theresa D'Addio for editorial research assistance,
and the members of the IFPRI Information Services Department for their con-
tinuing efforts to bring this book to fruition.

Part I

Defining the Food
and Nutrition Problem

chapter 1

Strategic Issues in Food

Production in Sub-Saharan Africa

Christopher L. Delgado, John W. Mellor, and Malcolm J. Blackie

Accelerating food production growth rates is now the central issue in African
development, a point eloquently made by Denis Norman in the foreword to this
book. The famines that periodically strike sub-Saharan Africa appear to have
become increasingly severe, and 24 African countries faced acute food
emergencies in 1984 and 1985, several involving large-scale human privation.
The conventional wisdom suggests that sub-Saharan Africa, formerly a food
exporter, became a net food importer in the 1970s, and by the mid 1980s, the
long-run rate of growth of net food imports seemed still to be increasing. In the
mid-1980s, sub-Saharan Africa was the only major region of the world where
population growth rates were still increasing and per capital food production
appeared to be declining. Even though urban food prices have been rising, the
rate of rural to urban migration has been high, with cities in the 1960s and 1970s
growing at over 6.5 percent per annum (World Bank 1981a).


This book starts with the premise that African food problems must and can be
solved in large part by increased national food production. That this premise is
clearly shared by the political leadership of African nations is demonstrated in
numerous public statements at national, regional, and pan-African levels. It is
reflected in the 20 contributions in this book by nationals of African countries.
Why this view has not been implemented is the subject of several chapters.
The centrality of domestic food production is also shared by the small
farmer, who gives priority to assuring the family food supply in his cropping

Defining the Food and Nutrition Problem

choices. Robinson Gapare, president of a smallholder farmers' organization in
Zimbabwe, observes that those "who buy food are shunned as lazy farmers."
This emphasis on home food production is strengthened by generally poor
transportation and distribution networks for food, which increases dependence
on local production for food security.
The emphasis on food in this book does not imply that every region of
every country or even the continent should be self-sufficient in food staples, or
that the objectives of increased food production can be dealt with separately
from nonfood agricultural production or from general economic development.
In some areas, topography, temperature, rainfall, and soil provide very substan-
tial comparative advantage to nonfood export crops; and everywhere, non-
agricultural employment must grow faster than agricultural employment. In
some cases, food and nonfood agriculture are so intertwined that they cannot be
treated separately, and in other cases they must be considered in conjunction
with the whole set of development processes. Nevertheless, because domestic
food production is so widespread in area and employment and because the
potentials for resource productivity increase are so great, it has a central place in
a development strategy virtually everywhere in sub-Saharan Africa. Successful
overall development is generally dependent on moving the food sector.
The centrality of food in African development arises from four forces.
First, domestic food production occupies a high proportion of the labor force in
Africa, frequently at very low levels of productivity. The soil and climate
conditions in some areas are such that productivity of labor cannot be expected
to increase. Major improvement in the well-being of these people will require
them to shift to other occupations. But for most of these people it is unreason-
able to expect such a shift in the foreseeable future. The capital requirements as
well as the organizational requirements for absorbing them in other sectors are
so large that massive numbers of people must remain in the food sector for
generations. Thus the vast majority of rural people will continue to contribute
directly to the food sector for the foreseeable future.
Second, in Africa-as has been convincingly demonstrated in Asia and
parts of Africa-improved production technology in the dominant food sector
can provide major increases in aggregate national income. Growth in real
productivity in the food sector increases real income of small farmers, and their
expenditures stimulate production and employment in other sectors of the
economy. This book emphasizes the first stage of such growth: how to achieve
increased output and productivity in domestic food production. It does not treat
the comparably important issue of how the income and employment multipliers
from -food production can be enhanced in Africa. That part of the story will be
pursued as food production growth is accelerated, though some of these issues
are touched upon, particularly in the concluding chapter. An important reason
why food production has not been given primacy in African development
strategies is insufficient evidence of its multiplier role in overall development.

Strategic Issues in Food Production

The case needs to be made that food production can effectively drive the
development process rather than being an essential, but unfortunate, diversion.
Third, while agricultural exports are important in Africa and generate the
foreign exchange necessary to the growth of the domestic food sector, expan-
sion of export commodity production depends more directly on improved price
ratios than is the case for food. Many export commodities have had a com-
parative advantage despite unusually unfavorable producer price relationships
imposed on them; furthermore, infrastructure is typically better in the areas
where they are grown. Also, research on export commodities has been under
way far longer than on food commodities, having usually started in the colonial
period. Thus, the export crop sector can probably be put back on track more
easily than the domestic food sector can be developed, although the food sector
continues to command much of the economic resources. Nevertheless, expan-
sion of export commodity production does require scarce seasonal labor re-
sources and thus may compete with what the farm family sees as its primary
job-producing food supplies for home consumption. Thus, the export crop
problem and the food production problem will continue to be interrelated-and
neither can be ignored.
Finally, projections of present trends point to food imports on the order of
40 million tons by the year 2000. While such projections are always of doubtful
accuracy both because of the questionable nature of underlying data and of
ongoing structural changes affecting outcomes, they do serve to illustrate the
relative magnitude of impending problems in the absence of structural change.
The foreign exchange requirements and physical facilities for handling such
imports seem at best difficult to deal with. Although African food imports in the
1970s represented a small percentage of world food trade, that will soon no
longer be the case. It is anomalous that, in this vast, land-abundant continent,
where more than three-quarters of the population is engaged in smallholder
food production, over one-fifth of food staples consumed are imported and that
the percentage is rising.


The second premise of this book is that technological change in agriculture is a
necessary condition for achieving sustained increases in food production. The
dynamics of rapid population and urban growth in a continent that is primarily
poor and rural suggest that per capital food production (or foreign exchange)
must grow rapidly if food consumption levels are to be maintained. A simple
arithmetical illustration is that if three quarters of the population are in agri-
culture, overall population growth is 3 percent per annum, and urbanization 6
percent, per capital productivity in agriculture must increase 1 percent, ceteris
paribus, to maintain constant consumption. However, a 3 percent overall real

Defining the Food and Nutrition Problem

rate of income growth coupled with a unitary income elasticity of demand for
food, a realistic assumption in Africa, would require a 4 percent rate of growth
per capital in food production-or very high growth rates in foreign exchange
outlays for foods if choking food consumption by higher prices is to be avoided.
This contrasts sharply with the outcome portrayed by Paulino, which shows
little or negative per capital productivity change in agriculture. The argument is
strengthened when the depressing effects of population pressure on the margin-
al productivity of agriculture are considered.
Without doubt, policy changes in some parts of Africa would result in
immediate growth in food production. Examples would be where market pol-
icies provide strong disincentives to producers compared to what would obtain
in the absence of these policies. The chapters by Olayide and Idachaba and by
Blackie illustrate clearly that not only can output marketing systems hold
production back but that failure to develop effective smallholder input delivery
systems is a major constraint on both technological change and production
growth with existing technologies.
It could be argued that policy interventions to raise food prices further than
their presently high levels, relative to world prices, would induce producers to
accumulate capital that could and would be invested in technological change.
Furthermore, the price increases, on this view, would send messages to private
and public sectors that would induce needed ancillary policy changes. Thus,
the argument goes, the key policy issue is less technological change than
incentives to set in motion autonomous processes of change.
The problems with this argument are twofold. First, as Delgado and
Ranade show, the structure of African (especially West African) agriculture is,
on the whole, relatively less propitious for that strategy than the agricultural
structure of the less egalitarian Asia. Returns to agriculture, on this view, are
relatively less concentrated on a landowning class in Africa; since they are low
in any case, accumulation is difficult in the absence of specific policy interven-
tions to this end. Second, public sector implementation of policies in response
to farm lobbies is probably slower in most of Africa than in Asia, for the same
reason. Compared to Asia, higher food prices in Africa may induce a return of
labor to the farm. But that is counter to the general historical experience of
development, involving rising agricultural labor productivity and declining
agricultural population.
In any event, the time pressures on African governments are such that not
only must improved producer incentive policies be put in place, but also the
much more complex issue of needed policy complements must be addressed.
The ongoing, complex, and disaggregated nature of these interventions sug-
gests that both technological change and recommendations for specific policy
reforms must be generated primarily by African institutions to ensure con-
tinuity, constituency, and relevance of the results. This will require not only
institutional development but greatly increased numbers of people to work

Strategic Issues in Food Production

therein and much greater central resources for investment in agriculture.
Once technological change in African food production is seen as central, it
is necessary to weigh the relative importance of three sets of policy-relevant
questions. First, what is the nature of the technological problems that science
must overcome to boost productivity on a sustained basis, and how adequate are
current research and extension processes? Second, what are the roles of other
support services and infrastructure, including provision of inputs, output mar-
keting systems, roads, and institutional knowledge? Third, what incentives
would influence at least some farmers to produce more once the technological
potentials are available and the necessary inputs and output marketing channels
are in place? The third question requires knowledge of how incentives are
influenced by the various agricultural and nonagricultural policies.
Improved knowledge of how technology, support services, infrastructure,
and incentives interact, and how priorities among them should be defined,
would set the framework for specific policy improvements, such as for fertilizer
stocks or processing facilities. Finally, it would lead to a more effective division
of functions among central and local governments, research establishments,
semiautonomous organizations, private companies, farmers' organizations,
and external donor agencies.
Given that solution to Africa's food problems must involve a sustained
process of technological change in production, the book's first five parts sug-
gest the means to generate such a process. Separately, they convey key pieces of
the puzzle. Part I is diagnostic and confirms the emphasis on production. It
discusses recent trends in the African domestic food sector, current thinking
about the relation of food production to nutrition, and the need for policy
action. Part II reviews current practices and potentials in agricultural tech-
nology development with emphasis on the major geographic and ecological
zones. Particular attention is devoted to the adequacy of current efforts to
generate a sustained stream of technological innovation. In part III, attention is
focused on the major support systems other than research for accelerating
agricultural development. These include input supply systems, marketing sys-
tems, and infrastructure. Armed with this knowledge, we can then examine
national development strategies to see why the requisites for growth in the food
sector have not been provided. This takes us to part IV and an examination of
growth strategy, trade policies, price policy, and a broad range of equity and
nation-building concerns. Insights into the long-run concerns actually faced by
policymakers in deciding on development strategies are presented by a senior
Kenyan and a senior Japanese commentator. Finally, in part V the role of
external donors is examined in the context of the powerful influence they
continue to exert in determining African food development policy. The final
part of the book brings the disparate and sometimes conflicting elements to-
gether and states priorities for policy action.

Defining the Food and Nutrition Problem


Analysis of the African domestic food problem requires information on the
production record of the past few decades, the degree of deterioration over
time, how it compares with other parts of the developing world, and the extent
of variation within Africa. It is especially important to know if some regions or
commodities are doing better than others and how these food production in-
creases have been achieved. Specific attention should go to the importance of
yield increases compared to area increases, the role of irrigation, fertilizer, and
machinery, changes over time in the relative importance of the various produc-
tion-increasing forces, how Africa's record compares to other areas, and the
implications of this record for the future.
Paulino's chapter, in viewing food production outcomes over the 1960s
and 1970s, documents a markedly poorer production record than Asia's. It
suggests that the rate of area expansion, the most important source of food
production growth in the past, has recently fallen below that of land-con-
strained Asian countries.
Paulino's analysis clearly demonstrates the weight of West Africa, es-
pecially Nigeria, in these overall results. This suggests the priority that should
be attached to better knowledge of the issues in this region from a food policy
perspective. It also cautions that uncertainties about the accuracy of the aggre-
gate figures for Nigeria should be extended to consideration of the regional
figures. The same might be true for the central African area, due to the weight of
Zaire, and to the influence of Ethiopia and the Sudan on aggregate results for
eastern and southern Africa. Nevertheless, the wideranging similarity of prob-
lems demonstrated across countries in Paulino's analysis strongly suggests that
the overall impression of a technologically stagnant agricultural sector is no
statistical aberration.
On the other hand, there are sources of hope in the results. Rice production
appears to have grown at a quite respectable rate, largely through area expan-
sion, and has contributed more to increments in output than the vastly more
widespread sorghum. Maize has also contributed considerably more to growth
than its production base would suggest. A second source of hope is that the use
of modern inputs has grown rapidly, though from a very low base. Paulino
pursues this analysis to the limits of the weak data bases currently available.
Doubts about the adequacy of the data base even for these very rough purposes
underlines the urgency of improving the institutional capacity of African coun-
tries to generate the data needed for assessing and modifying food policy.
Kumar confirms that the deterioration in the underlying food situation
described by Paulino is from a very low and inadequate nutritional base. The
acuteness of the humanitarian issue is clear. She also demonstrates that, for a
rather well-endowed agricultural area, labor scarcities are rooted at least partly
in inadequate food production and consequent deficiencies in food intake and

Strategic Issues in Food Production

capacity to work. The special nutrition problems in Africa relate not only to the
initially low levels but also to the seasonality and the high year-to-year vari-
ability of food intake. Both she and Vamoer confirm that hunger is primarily a
rural phenomenon in Africa; Vamoer also notes a rising tendency for malnutri-
tion in urban areas. Given the ubiquitous nature of smallholder farming (com-
pared to landless laborers), acute malnutrition is seen as closely related to
decline in food availability. The relation between production and consumption
is much more direct than in most of Asia, where food production affects the
poor primarily through its effects on rural employment and prices.
Ghai and Smith, on the other hand, point out that Asian-like situations are
becoming increasingly common in parts of Africa and urge caution in raising
food production through higher prices. As Vamoer points out, the importance
of production to the nutritional status of at-risk populations in Africa corrobo-
rates greater smallholder production as the key element in improving food
Furthermore, Kumar suggests that the decline in nutritional status in Zam-
bia in poor crop years may result in increased migration from rural areas and
reduced availability of agricultural labor, which in areas of relatively low
population density lowers planting levels and production in succeeding years.
Thus a vicious circle results-low labor productivity in food production results
in poor nutritional status, which further reduces labor productivity.


Wherever there are even the most rudimentary agricultural research systems, it
is claimed that technology is already available to greatly increase production. It
is often asserted that all that is needed is to alert the farmer to these oppor-
tunities, or to change public policies with respect to prices, or perhaps to
streamline input delivery systems. This view was widespread in Asia in the
1950s, but major yield increases did not occur until agricultural research break-
throughs were achieved in the mid- and late 1960s. By then it was widely
recognized that, while research station yields were higher than farmers' aver-
ages, they were not above those of the best farmers. Experiment stations often
had different physical and economic conditions from farmers. Most particu-
larly, they sought higher yields rather than higher profits. There was also a
tendency to note those experiment station plots with high yields and to ignore
those with low yields-a luxury not permitted the practical farmer. It is proba-
bly correct that testing of varieties for profitability and safety in farmers' fields
is still more the exception than the rule in Africa today.
Part II deals with the vital question as to what extent low adoption rates for
technology are explained by the inappropriateness of recommended technolo-
gies to farmers' conditions and to what extent they are explained by lack of

Defining the Food and Nutrition Problem

incentives and information. The issue is surrounded by considerable controver-
sy and a paucity of facts. The eight chapters in part 11 are written by persons
with widely differing backgrounds and extensive experience in the matter.
The chapters by Matlon, ter Kuile, and Collinson, who are associated with
three of the international agricultural research centers in sub-Saharan Africa,
adopt the perspective of farming systems analysis. Penning de Vries and de Wit
approach the same set of issues from a more strictly agricultural science per-
spective on a continentwide basis, and discuss the positions of the three region-
ally oriented farming systems chapters in that light. Delgado and Ranade
explore the key role of labor constraints in African agriculture and the particular
problems they pose for technology design in Africa compared to better-known
cases of technological progress in South Asia. The special role of women in
Africa, their role in production, the neglect of their needs in the current institu-
tions of technological change, and the implications thereof, are briefly explored
by Kumar. Vallaeys and his colleagues focus on the issues involved in the
development and extension of improved technologies under the conditions
outlined by previous authors. Finally, the commentaries by Taylor and Mupa-
wose deal with these issues from the perspective of direct experience with
developing national systems of agricultural research in Nigeria and Zimbabwe,
The farming systems chapters for the three major regions (part III) inves-
tigate the nature of soil resources, climate, potential for different kinds of
irrigation, the genetic potential for various food crops in African environ-
ments, and the interaction of crops and livestock in complex systems. They
show how these factors interact with available knowledge to define the poten-
tial for technological change to increase resource productivity. They include
reviews of major past and current research in Africa, judgments as to some of
the most promising fields of inquiry, and thoughts upon some of the most
difficult physical and technical constraints to overcome.
Generally, these chapters illustrate the considerable constraints placed on
African agricultural production by the continent's difficult soils, climate,
groundwater, disease, and pests. They reflect generally poor progress in the
provision and adoption of research results, despite a massive increase in re-
search investment, albeit over a relatively short period of time. They illustrate
the need for input-intensive strategies (because of the low nutrient content of
soils) and the inappropriateness of many available technologies and inputs,
especially in zones with low agricultural potential.
All three chapters emphasize a major issue raised by Norman, namely the
loss of cropland due to erosion. Soil fertility management in both the drier and
more humid zones emerges as an issue with implications for technology design,
land tenure policies, fertilizer delivery and extension systems, and the promo-
tion of integration of livestock and cropping systems. While growth in fertilizer

Strategic Issues in Food Production

use is clearly key to boosting the productivity of old and leached soils, the
fragility of some is such that better ways to improve soil structure must be
considered at the same time. Furthermore, the exact mix of interventions, as
Matlon shows, is likely to be quite location specific.
Matlon also argues cogently for a change in conceptual approach to agri-
cultural research in Africa, away from strict adherence to green revolution
strategies for plant improvement to boost yields per hectare. He feels that
research strategies should include (1) more balance between immediate produc-
tion gains and environmental impacts; (2) packages better suited to limited farm
resources, with each component capable of being viable by itself; (3) more
region-specific recommendations; and (4) more off-station research, with more
direct participation by farmers. We attempt in the concluding chapter to recon-
cile this thoughtful position with the urgency of major acceleration in food
production growth rates.
All these chapters stress the need for a farming systems approach to
technology design and implementation, which would permit innovations ap-
propriate to the constraints faced by smallholder farmers. A major constraint is
the widespread seasonal labor shortages and low labor productivity, which limit
expansion of land area. Delgado and Ranade show that, while innovations in
seed fertilizer have typically increased demands for labor input in critical
periods in both Asia and Africa, this was more likely to cut into farm prof-
itability in Africa than in Asia because of Africa's sharply rising labor supply
schedules. Thus technology design must consider seasonal labor impacts,
which implies that farming systems research may have an even greater adjunct
role to play in agricultural research in Africa than in Asia. Nevertheless, it is
probably still the case that the most efficient means of raising labor productivity
in Africa will involve raising yields per hectare.
Matlon and Collinson make the crucial point that both West and East
Africa are progressively losing arable land. This suggests that the need for
technological change is being driven not only by a need to increase food
production but also by a need to intensify production on a shrinking resource
base. This should be considered in light of Penning de Vries and de Wit's
contention that the biological capacity for food production in Africa is high,
given sufficient intensification. In this context, they raise the issue of single
versus intercropped systems, the former being a key component of purchased
input-intensive development strategies, whereas the latter is a traditional solu-
tion to the need to intensify and diversify at the same time.
The chapter by Vallaeys and others dealing with the longer-run issues in
technology improvement explores the processes by which technological im-
provements are generated in Africa. These contributors devote particular atten-
tion to ways in which institutions can strengthen a two-way flow of information
between farmers and scientists. This is especially relevant to the division of

Defining the Food and Nutrition Problem

labor among national, regional, and international research institutions, to the
size of national institutions, and to allocation of funds among basic research,
adaptive research, and extension functions.
The conclusions for the present are clear and sobering. Existing tech-
nology packages have limited potential for most of the arable area outside
irrigable and high-altitude zones. However, maize in subhumid lowland areas
may also present significant potential. In general, African research institutions
are not yet at a stage to continuously generate the needed technology.


Technological change in agriculture is typically associated with increased com-
mercialization of output and rapid increase in use of purchased inputs, most
notably fertilizer. Paulino's chapter confirms that current input rates for fertil-
izer are exceptionally low in sub-Saharan Africa. They are less than 5 kilo-
grams per hectare in more than half of the 39 countries of sub-Saharan Africa
and over 15 kilograms per hectare in only 5 countries. Even the latter rate is low
by developing country standards. Experience elsewhere in the world indicates
that the averages for Africa probably reflect moderate dosages of fertilizer on a
small proportion of the land and none on the vast bulk of it.
Rapid expansion of input delivery systems, especially for fertilizer, would
be expected to be a major immediate source of growth in production where
inadequate policies in this area have held growth back. Experience in Zim-
babwe since 1980 has shown how rapidly maize production could increase in
smallholder agriculture, given an assured market and satisfactory delivery
systems for inputs, especially fertilizer. Paulino points out that fertilizer use has
been growing 10 percent a year for sub-Saharan Africa as a whole, but from a
very low base. Rates in Asia ranged between 15 and 25 percent in early stages
of growth.
However, the issue of returns to fertilizer in Africa is in fact very complex.
Matlon discusses severe technical problems with fertilizer use in the semiarid
tropics of Africa, including low response of local varieties and possibly long-
term declines in yields. Penning de Vries and de Wit suggest that the explana-
tion may lie with deficiencies in phosphorous or trace elements. Low response
to fertilizer may explain nonadoption of available technologies and contribute
to an inelastic response of aggregate agricultural output to increased relative
food prices. On the other hand, Olayide and Idachaba as well as Falusi point to
situations where fertilizer use has grown rapidly to substantial levels. This has
been typically associated with ample supplies and good delivery systems; these
conditions seldom occur in most of Africa. We conclude that there is a potential
for increased fertilizer use in specific areas, but not everywhere, and that where

Strategic Issues in Food Production

and when to emphasize improved input delivery systems requires careful
A recurring theme is that the public sector is unable to effectively manage
existing input and output marketing systems. Desai, drawing on Asian experi-
ence, notes that the private sector manages output marketing with greater ease
than public entities. But he also notes that the private sector in India was slow to
provide reliable input supply systems, especially in the earlier phases of agri-
cultural development. He suggests that African governments draw back from
parastatals in output marketing in favor of a more intensive effort on the input
Several chapters address the different ways that governments do, and
should, intervene in the provision of agricultural services. Blackie raises the
issues stemming from centralization of decisionmaking. He also reviews the
costs and benefits of full control by parastatal bodies versus regulation of
private sector agents. He draws on experience in Zimbabwe, which is more
positive about the role of parastatals than most other African countries, and
addresses the issue of how to make them work as well in smallholder areas as in
commercial farming areas. Olayide and Idachaba discuss the important issue of
intervention through intensive package approaches, as in integrated rural devel-
opment projects, versus single-function programs, such as insecticide provi-
sion. They also review Nigerian experience with input subsidy schemes at the
national level. Throughout the discussions of marketing, it is clear that the
optimal role for government will vary greatly among countries and that ideo-
logical positions of the left and right should be viewed cautiously.
Poor transport systems are a major marketing problem in Africa. Agri-
cultural marketing margins in five major African countries studied by Ahmed
and Rustagi (1984) are typically twice as high as in four major Asian countries;
higher transport costs account for 40 percent of the difference. In large part, this
may be due to the scarcity of rural roads. In their sample, road densities
averaged between 0.03 to 0.10 kilometer per square kilometer in sub-Saharan
Africa, compared to 0.40 kilometer per square kilometer in South Asia. Ida-
chaba's chapter on infrastructure illustrates the low density of roads in rural
Nigeria and the tremendous variability of infrastructure from state to state and
within states.
Despite poor infrastructure, there is disagreement both in policymaking
circles and in this book as to the priority to attach to resource allocation for it.
Lipton argues that in many areas the potential for technological progress is
limited, an argument in accord with the thrust of the technological chapters. For
this reason, he objects to the emphasis on centrally provided grid infrastructure,
which he feels absorbs too many resources. This raises the issue of whether
technological change must precede provision of infrastructure. However, Lip-
ton does not accept as given the current low priority for foreign exchange

Defining the Food and Nutrition Problem

allocations to agriculture, contrary to those who believe that because low use of
purchased inputs and lack of infrastructure are endemic in Africa they are not
policy issues. Idachaba adds the twist that it is particularly difficult to set
investment priorities between infrastructure and other agricultural uses since
any one element, such as service centers, may require another, such as roads, to
pay off.
Wanmali raises issues about how to allocate infrastructural expenditures
among functions and locations and over time. He stresses the linkage effects of
government investment in infrastructure to private sector provision of services.
He also argues forcefully for a view of centrally provided infrastructure beyond
just roads and other grid elements. Thus provision of institutional knowledge
and decisionmaking capacity is key to getting agriculture moving. This is
central to the need evoked earlier of continuity and constituency in policy
analysis. Provision of what he terms soft infrastructure, such as better agroser-
vices, is important also. This becomes an issue for direct government interven-
tion in the context suggested by Desai, in which the private sector might be slow
at filling needs, especially in areas with low potential.
Although Wanmali and Idachaba, in particular, support massive invest-
ment in infrastructure, the three substantive chapters also suggest that in-
frastructure is very sparse in Africa and that vast resources would be required to
build it up to a state comparable to other parts of the world where the revolution
in seed fertilizer has prospered. Lipton argues that more attention should be
focused on production-specific infrastructure, such as irrigation, within a well-
defined technological context. The issues are whether resources of that magni-
tude can be raised domestically and from foreign sources and, if so, should they
be allocated so massively to rural development. If so, it is then important to
know how to secure maximum effect from mobilization of the private sector
and the time lags involved.
Given the large role of government in provision of support services and
infrastructure, Falusi and Schluter stress the constraints on improving govern-
ment performance in these areas. Lack of decisionmaking capacity at the local
level is singled out as a key problem, which is exacerbated by acute shortages of
skilled manpower. Gapare, arguing from the standpoint of a lifetime of direct
experience, suggests ways to secure the active participation of smallholder
farmers in decisionmaking. Such a strategy goes far beyond local resource
allocation decisions. It is in effect a plea for more rural political input and
power, a plea echoed in a much broader context by Bienen and Ndegwa. It
crops up in several other chapters, also, and was continually emphasized in
discussions at the Victoria Falls conference. How to promote structures that
effectively promote rural participation is a central issue from both the technical
and national political standpoints.

Strategic Issues in Food Production


Successful pursuit of specific production policies is possible only in the context
of consistent overall development policies. Four chapters plus four related
commentaries are devoted to these broader policy issues. They contain many
examples of how larger strategic objectives constrain food production policy.
The very real commitment to increased food production in many African
governments is often frustrated by the complexity of the problem and the
paucity of knowledge for dealing with it. Extreme pressure from foreign and
international agencies may also alter policies and basic strategies in the face of
powerful contrary forces of ideology, political constituency, equity concerns,
and vested interests-these, too, conditioned by the paucity of knowledge,
particularly with respect to the details of policy implementation. Pressure from
external sources is most likely to occur in times of misfortune, when govern-
ments are desperate because of collapsing terms of trade or climatic disaster.
Once the short-run crises end, however, policies seem all too likely to revert if a
national consensus has not been built. For this reason, in the concluding chapter
we emphasize building a national policy analysis capability so that the appro-
priate changes in policy can be fully analyzed and persuasively argued.
When faced with conflicting strategic choices, agricultural policymakers
may need to accommodate a second-best solution. They will have to accept
certain policy constraints to agricultural development and substitute compen-
satory policies that work around the initial distortion. In any event, exploration
of the links between overall incentives for agricultural production and overall
development strategies is essential to clarify options for strategic change and to
better understand what agriculture is up against.
Agricultural development requires massive investments; governments
may coopt the resources for this from within agriculture or from other sectors by
such policies as favorable price relationships that are substantially determined
by foreign exchange rates, related monetary and fiscal policies, and other
favorable policy decisions. However, these all have short-run political costs.
The sense of public strategy with respect to agriculture that is necessary to
resolve these conflicts in favor of agriculture has in practice been absent in
Africa. Perhaps this has been due to lack of confidence in the potential of
technological change and by doubts as to the capacity of agricultural develop-
ment to move other sectors.
Aboyade makes three important points concerning this problem. First, he
suggests that agriculture may have been given a more central role than is
generally recognized, but that those resources were grossly misallocated and
hence ineffective in achieving the desired development. This is consistent with
a view that a vast amount of detailed knowledge is necessary to successfully
implement effective policy. Furthermore, it underscores the point that the in-

Defining the Food and Nutrition Problem

stitutional base for gathering and assimilating factual knowledge is grossly
underdeveloped in Africa.
Aboyade raises another important issue. Once a long-run commitment is
made to emphasize agriculture and the nature of the effort is understood, the
problem remains as to how to stimulate agriculture in the short run. At this
point, unit costs of production may be high and resource productivity low, yet it
is politically important to demonstrate the benefits that a technologically dy-
namic agriculture will produce. Aspects of this serious question are discussed
in this book and underlined in the conclusions, but unfortunately only tentative
and provisional answers are possible at this time.
Aboyade emphasizes the importance of local government bodies to rural
development. Resource requirements are immense, and resources must be
raised, in part, locally. Programs must be tuned to local conditions. Further-
more, a rural constituency is essential if political support is to be given to a
rural-oriented strategy. Aboyade shows a distinct lack of enthusiasm for the
constantly shifting fads of the donor community and a pessimism for traditional
strategies based on agricultural commodity exports.
Even more than other developing areas, Africa relies on agricultural ex-
ports, which are a crucial source of foreign exchange for imports of items vital
to the growth of the domestic food sector, such as fertilizer and transport
equipment. Oyejide shows clearly how the trade regimes of African countries
have prejudiced not only the agricultural export sector but food production for
domestic consumption. This is a technical area in which knowledge has ex-
panded particularly rapidly in the past decade, and hence Oyejide's chapter lays
out the relevant conceptual framework.
Oyejide examines food policy within the larger context of the choice of
trade regime and policies toward exchange rates, tariffs, and nontariff barriers
to trade. These factors are especially important in the relatively small, open
economies of Africa. He examines how trade regime policies influence the
structure of incentives for agriculture relative to other sectors; the way that trade
regime policies affect the movement of resources among production of domes-
tic food, cash crops, nonagricultural goods, and services; and the implications
for strictly defined food policies. Finally, in a section especially relevant to
Nigeria, but which may be applicable to a growing number of African coun-
tries, he looks at the way that the presence of a dominant nonagricultural export
sector, such as oil, can discourage agricultural activity. In Nigeria, this occurred
through the exchange rate, which lowered returns in domestic currency to
producers of agricultural tradables, and through the way oil wealth was spent,
which gave a positive incentive to move resources into nontradables, such as
the service sector. It should be noted that foreign assistance-which in several
African countries is as large relative to gross national product (GNP) as oil is in
Nigeria-has a similar effect, particularly if it is used largely for local cost

Strategic Issues in Food Production

Both Aboyade and Oyejide conclude, from different approaches, that
serious consideration should be given to raising domestic agricultural prices in
Nigeria relative to world prices, even when they are above world market levels.
Aboyade bases his argument on the need for governments to give a strong,
immediate, positive signal to agricultural producers and to the political system
of their intention to respond to the problems of the sector. Oyejide's argument is
based on the fact that, although effective protection rates are currently high for
Nigerian agriculture, they are even higher for many nonagricultural items. In
effect, he argues for equal effective protection for the two sectors.
Krueger introduces a note of caution on the short-run protectionism for
food production suggested by Aboyade, although in a context of market-
aligned exchange rates. She notes that agriculture should be able to stand on its
own feet. Atsain and Tshibaka argue that care must be exercised when applying
infant industry arguments to agriculture, because protection can backfire if it
prevents productivity-increasing innovation. In effect, the short-run solution
can interfere with the long-run solution.
Oyejide and Tshibaka are both clear on a point that occurs in many of the
chapters-and that we comment on at greater length in the conclusions. In
discussing the effect of exchange rates on prices, Oyejide sees wage rates and
labor migration as central to agricultural problems in Africa; in other words,
rising production costs, as reflected in labor costs growing more rapidly than
productivity, are more critical than relative producer prices per se. In effect, the
villain of the food story is neither low output prices nor the draining of resources
into export crops but high labor costs fostered by rapid growth in the non-
agricultural, urban sector. The trade regime's capital flows and exchange rates
play a role in creating this situation, just as they do in relative output prices, but
the specifics of the cure may differ. Tshibaka also points out that the main way
the trade regime affects incentives is through rural labor markets, and adds that
we know very little about how these markets work.
Oyejide makes the point, as does Aboyade, that African food problems
cannot be understood without looking at the structural factors that favor the
nonagricultural sectors over agriculture in the policy process. These include the
political economies of emerging African nations, the greater feasibility of
centrally managed policy intervention in their nonagricultural sectors, and a
desire to protect against external threat.
The political economy of food-pricing policy is a recurring issue in Af-
rican policy debates. Ghai and Smith examine the impact of high food prices on
consumers, assuming that they are not net sellers of food. Based largely on East
African evidence, they argue that the assumption that all poor people are net
sellers of food at all times is no longer valid. In fact, it seems clear that the
proportion of poor people who are sellers of food varies substantially by year.
Ghai and Smith also focus on the rapidly growing mass of unemployed poor in
the cities. From their viewpoint, production policies requiring higher food

Defining the Food and Nutrition Problem

prices would have negative equity implications, at least in the immediate
future, and thus would run afoul of important strategic considerations. The
validity of their argument will depend on how "distorted" domestic prices are
in the first place and on the feasibility of inserting a wedge between consumer
and producer prices. As Eicher suggests, these factors will vary widely among
countries. The equity concern with higher food prices is a widely accepted view
wih respect to Asia but a more recent concern in Africa. It strengthens the case
for increasing domestic food production by cost-decreasing technological
change, the only solution that reconciles the interests of both producers and
In support of the above point, Spencer and Eicher point out that technolo-
gies are at least as important to incentives as output prices. They note that, even
though the domestic terms of trade between food and nonagricultural goods
may have shifted in favor of food producers in the 1970s, as Ghai and Smith
argue, aggregate agricultural production increased little. This could be ex-
plained by lack of appropriate technology, and appears to suggest that while
better output prices for food producers may be an important component of a
food production strategy, it is not a sufficient condition. Both Spencer and
Eicher stress the problems of looking at relative price incentives using aggre-
gate data that disguise a great deal of spatial and seasonal variation and that are
misleading in many cases. Both are critical of a general, price-based approach
to equity, claiming that this issue can only be analyzed by detailed country
Ghai and Smith, while focusing on poverty, do add evidence for which
there is increasing concurrence-that whatever the intent of governments may
have been, relative food prices have been trending up in much of Africa. Their
evidence is based on shaky data for consumer prices and on official food prices
that refer to only a portion of marketing. These ratios, then, represent a
measure of the effect of the trade regime on the relevant domestic exchange rate
and bear at least some relation to the "real exchange rate" discussed by Oye-
jide. The controversy about the Ghai-Smith chapter drives home how depen-
dent useful policy prescription is on accurate factual information, which is still
largely missing from the African scene. In addition, the Ghai-Smith chapter
drives home how much policy results vary from country to country in Africa.
Thus we must be as careful in generalizing about policy as in generalizing about
the physical environment.
Why governments do not give more emphasis to food production and why
the analyses they provide seem so much in error concerns Bienen as well as
Aboyade. Bienen points out that vested interests in urban development and the
better organization of urban people may result in policies inimical to rural
development. Military governments have an interest in overvalued exchange
rates, which hold down the currency costs of military equipment and imported
goods consumed by the military but which discriminate against agriculture.

Strategic Issues in Food Production

This again emphasizes the need for political organization and representation of
rural people if effective rural development is to occur.
Using a series of country case studies, Bienen shows that politicians can
make rational choices that are nevertheless contrary to overall economic effi-
ciency-and to successful food production policies, in particular. This stems in
part from the fact that urbanites are typically better organized than rural people.
However, coalitions of large-scale farmers or rural leaders also can obtain
government programs favorable to them. For example, subsidies can more
easily be targeted to subgroups of politically powerful farmers than price-based
interventions can be.
In commenting on Bienen's paper, Muleya points out that political vari-
ables not only affect agricultural incomes but that the reverse is also true. He
emphasizes that expatriate advisers are important to development of food pol-
icies, while overall development strategies are the outcome of a national politi-
cal process, and that the two may not have the same objectives. Furthermore,
expatriate advice is notoriously subject to the intellectual fashion of the mo-
ment. Aboyade also calls for greater reliance by African governments on
indigenous intellectual expertise to help provide continuity between policy and
strategy formulation.
Ndegwa, a major African policymaker, cogently argues for pan-African
nationalism as a means of mobilizing resources for development without gloss-
ing over the immense difficulties. He emphasizes that food policies are second-
ary and that national unity is and should be the overriding objective of national
development strategies in Africa. Nationalism is seen as a means of minimizing
ethnic and class rivalries, which can have disastrous effects on human welfare.
Problems resulting from nation-building policies, such as imperatives for re-
gional equity, can be overcome somewhat through increased regional economic
Okhawa looks at the long-run development processes of Japan and other
countries of Southeast Asia in terms of their relevance for Africa. He stresses
the long-run nature of development paths. While Japan's postwar growth was
rapid, he points out that it was preceded by a very long period of modest results.
During this time, government intervened extensively in the economy to build
the "social capability" of the population to accept foreign innovations and then
to create indigenous ones. He stresses the particular problems that African
countries face as very latecomers in the development process. Here, high wage
differentials between a "modern" nonagricultural sector and a "traditional"
rural sector appeared much earlier in the development process than they did
in Southeast Asia. Okhawa's central notion is that African food problems can-
not be well understood without taking into account the nature of the world
system into which Africa is emerging. This clearly differs from the earlier era,
when Southeast Asian countries were at comparable stages of economic

Defining the Food and Nutrition Problem


It is clear that donors have played a major role in African food policy. Aboyade
cites changing fads in donor assistance as a reason why African food policies
are in such disarray. Much of the external discussion of African food problems
appears to assume that it is easy to determine the food policy problems in any
given situation and that made-to-order policy solutions automatically follow.
On the other hand, an emerging school of thought among some senior African
analysts, as reflected in parts of the Lagos Plan of Action, postulates that
African problems are peculiar to Africa and that insights from other areas are of
limited value. The correct position is probably in the middle ground.
Clearly, many specific food policies and technologies are not directly
transferable to Africa. Yet it is plausible that the processes followed to develop
these policies and technologies, and perhaps the specific types of institutions
developed, are transferable. At the very least, experience elsewhere should
yield valuable insights into the constraints on creating similar institutions in
Lele reviews the impact of foreign assistance on African agricultural
production strategies in the past and assesses the potential for useful future
contributions. Her assessment of what donors have done previously confirms
that their influence was major. In the early 1980s, total annual foreign as-
sistance to sub-Saharan Africa was approximately $10 billion, of which per-
haps a third went to agriculture. There were around 80 thousand external
technical assistants residing in 40 countries. Nevertheless, by her account the
donor impact on food production growth over the past 20 years was probably
small and possibly counterproductive.
Lele raises the important issue as to the desirability of donors focusing on
final growth or equity objectives in the short run, as distinct from creating the
capacity in the nations to achieve these objectives in the long run through the
establishment of institutions, trained manpower, procedures, policy analysis,
and information capacity. She gives the example of Asia at its earlier stage of
development, in which fewer and larger donors were able to help develop such
indigenous capacity. She points out that, in contrast, in Africa the proliferation
of donors, their conflicting views about development goals and means, and
competition among them has failed to create such capacity. It is clear that for the
donors to be successful in Africa, the conditions obtaining in Asia would have
to be simulated through aid coordination. This requires a more cohesive do-
nors' view as to the most effective agricultural strategies to achieve the long-
term goal.
Because of the political nature of food policies, Mensah suggests that
donors cannot have much of a direct role in setting them. However, they do
have an important role in increasing host-country capacity to do a better job,
especially in planning, budgeting, and monitoring public investment in agri-

Strategic Issues in Food Production

culture. Thus, Mensah and Lele agree on the fundamental approach donors
should adopt. However, Mensah adds that, where donors make specific recom-
mendations, the costs need to be carefully assessed in advance and then at-
tributed in an agreed fashion among parties.
Johnston also stresses the importance of distinguishing between the facili-
tation of policy formulation and the provision of services in foreign assistance.
He adds that this distinction can be applied within a basic, human needs
approach. For example, the proper policy would be to strengthen the institu-
tions providing national health services rather than to provide health services, a
subtle but key difference. The policy facilitation perspective was not particu-
larly noticeable in the basic human needs era of foreign assistance to Africa in
the 1970s, but it is not necessarily incompatible with a strategy for immediate
improvement of human welfare through foreign assistance.
De Lattre corroborates Lele's contention that much of the resource flows to
agriculture in the 1970s did not attack the key problems. Half of official devel-
opment assistance to agriculture in the Sahel went to irrigated cropping, but 95
percent of domestic food supplies come from rainfed areas, much of it without
good irrigation potential. De Lattre adds that, since domestic policies are
clearly central to the success of aid, it is unrealistic to assume that donors will
refrain from taking a position on them.


The contributors to this book reflect a wide range of backgrounds and view-
points. They come from five continents and range from African cabinet mem-
bers to a representative of smallholder farmers and from staff of donor agencies
to African academics. While they could not be expected to agree on all points, a
surprising number of common themes emerge.
First, they clearly put special emphasis on increasing the capacity of
farmers to produce rather than on just the will or incentive to do so. Tech-
nological change is a key consideration. Perhaps the greatest diversity of views
concerns the policies toward marketing systems, input supplies, and infrastruc-
ture. This may be due to the fact that the right policies are highly time- and
location-specific. In any case, such policies also are likely to depend to a great
extent on available technology for the area in question.
Second, the demand side question that determines the domestic terms of
trade between agriculture and nonagriculture is viewed as a strategic issue with
high political content. Output price policy as conventionally thought of is
considered to be but one of several key policy variables affecting farmers'
incentives to produce. Furthermore, it is probably correct to view relative
output prices, like many other aspects of food policy, as products of decisions
quite outside the food sector. Thus nonagricultural policies such as pegged

Defining the Food and Nutrition Problem

exchange rates and industrial tariffs may frequently have greater repercussions
on agricultural incentives than agricultural policies, strictly defined.
Because the capacity of farmers to produce increasingly requires massive
resource allocation to agriculture, and agricultural incentives are a strategic
and political issue, efforts to get agriculture moving will affect the overall
development process. Therefore, the centrality of domestic food production to
development is an especially important theme in this book. This theme derives
primarily from the structure of African countries. In many areas, growth in
food production is frequently the only feasible means of assuring food se-
curity. In a longer-term perspective, it is probably the only means to insure the
active participation of the majority of the population in the development pro-
cess, given the relative importance of agriculture and the overwhelming pre-
ponderance of food in agricultural output.
It is clear from the many contributions to the Victoria Falls conference
that there is an urgent need both to mobilize scarce resources and to set pri-
orities for their use. The consensus at the conference was that, in general,
current processes for setting priorities in agriculture are inadequate. Because
of the size and nature of the public goods involved-research, infrastructure,
education, and so forth-it is clear that governments will have to have a big
role in getting African food production moving. Therefore, this book puts
particular emphasis on the need to improve the quality of policy analysis and
conditions for government decisionmaking. This will involve both training
and the support of improved institutions. While this is primarily the responsi-
bility of national governments, there is a clear consensus that external donors
could-and should-become much more deeply involved.
The following chapters vividly portray the complex problems of accelerat-
ing food production in Africa. They also constitute an urgent call to action. But
governments can do only a few things at a time. That is the dilemma of public
policy. In the concluding chapter, we offer prescriptions simple enough to be
put into practice. Such an exercise is unavoidably risky. And, of course, tomor-
row's priorities will necessarily differ from today's. Nevertheless, we are confi-
dent that our colleagues here have provided the basis for an initial agenda that is
not only correctly set but lends itself to evolution toward a far broader and more
complex system, as the necessary formation of human capital occurs.

chapter 2

The Evolving Food Situation

Leonardo A. Paulino

Over the past two decades, of all the major regions of the Third World, sub-
Saharan Africa has had the most rapid growth of population and the slowest
growth of food output.' And its population growth rate has accelerated, while
the food production growth rate has decelerated. For the Third World as a
whole, production of major food crops outpaced population growth by more
than a half percentage point per year over the past two decades, while in sub-
Saharan Africa the food production growth rate was a full percentage point
short of its population growth rate. Food production growth has depended more
on growth in crop area and labor force and much less on yield increase and
modern inputs compared to other parts of the world. Thus African agriculture is
characterized by poor and deteriorating performance and little modernization.
Per capital food performance was particularly poor in West Africa, which
accounts for nearly half of both the people and the production of basic food
staples of sub-Saharan Africa (table 2.1). Among its subregions, West Africa
had the slowest growth in food output. Population growth outpaced increases in
production of major food crops by 2 percent a year in West Africa and 0.5
percent a year in eastern and southern Africa. In contrast, the food production
increase in central Africa averaged 3 percent a year, which exceeded population
growth significantly but fell far short of offsetting the declines of per capital
output in the other subregions.

I. Much of the basic information in this chapter is from the Food and Agriculture Organization
of the United Nations, which relies heavily on national statistics. The agricultural statistical systems
of many countries in sub-Saharan Africa are still in early stages of development. Doubts as to the
quality of crop and livestock statistics have been expressed and inconsistencies noted. In this
chapter, data are used to support a level of generalization believed to be consistent with the accuracy
of the data. In the belief that the errors tend to be random, the data are generally used at a high
degree of aggregation of countries and, to some extent, of commodities. Thus the grouped data tend
to be stable and to provide a more reliable basis for indicating trends.

Table 2.1 Population and food production in sub-Saharan Africa


Eastern and southern
GNP per capitale
Under $250
$500 and over
Self-sufficiency ratio
Under 95%
105% and over

Population" Total food production Per capital
food production
Millions Percent Million tons Percent (kilograms)

Food production
annual growth

Total Per capital

139.1 44
55.8 17
122.8 39

88.7 28
86.3 27
142.7 45

191.1 60
105.3 33
21.2 7

Sources: U.N. 1983; FAO data tapes 1975, 1979, 1981, and 1983.
"1976-80 averages.
b1976-80 averages; food includes cereals, roots and tubers, pulses, groundnuts, and bananas and plantains. For purposes of aggregation, rice is in husked form,
and noncereals are converted to cereal equivalents based on calorie content.
cMeasured from the midpoint of 1961-65 to the midpoint of 1976-80.
dCountries composing the regions are as follows: West Africa-Benin, Burkina Faso, Chad, the Gambia, Ghana, Guinea, Guinea-Bissau, Ivory Coast, Liberia,
Mali, Mauritania, Niger, Nigeria, Senegal, Sierra Leone, and Togo. Central Africa-Angola, Burundi, Cameroon, Central African Republic, Congo,
Gabon, Rwanda, and Zaire. Eastern and southern Africa-Botswana, Ethiopia, Kenya, Lesotho, Madagascar, Malawi, Mauritius, Mozambique, Namibia,
Reunion, Somalia, Swaziland, Tanzania, Uganda, Zambia, and Zimbabwe.
e1980; based on trend estimates of real GNP during 1961-80 (1979-81 = 100).
fRatio of production to consumption, 1976-80.

The Evolving Food Situation

Judging from other regions of the world, we expect lower population
growth rates and higher food production growth rates to be associated with
higher incomes. The opposite has been the case in sub-Saharan Africa. Thus
between 1961-65 and 1976-80, food production in very low-income (GNP per
capital of less than $250 in 1980) and low-income ($250-$499) countries in
sub-Saharan Africa expanded 2 percent or more annually but expanded by less
than 1.5 percent in the middle-income ($500 and over) countries. Population
increase greatly outpaced the growth of food production in the last two upper-
income groups. Hence, unlike most other regions, the rates of growth in per
capital production show a pronounced inverse relationship with income. Per
capital food production in the middle-income group had a rapid decline of 1.7
percent a year.
Self-sufficiency in food is a rare phenomenon in Africa. Countries with
1976-80 self-sufficiency ratios (SSR) of 105 percent or more for basic food
staples represent only 7 percent of the population and 9 percent of food produc-
tion in sub-Saharan Africa. However, these countries expanded output of basic
food staples 2.6 percent a year, or nearly as fast as population growth, while the
average annual production per capital was over half as high again as for the most
food-deficit group. Population growth was twice as fast as growth in food
output in the food-short countries (ssR of less than 95 percent), which account
for more than half of both population and food production.


Based on United Nations projections, population growth rates in sub-Saharan
Africa will further accelerate in more than 60 percent of the countries, and the
regional rate will substantially exceed 3 percent per year for the period 1980-
2000 (table 2.2). In the 1970s, eastern and southern Africa experienced the
fastest population increase, followed closely by West Africa. Population
growth in central Africa has been slower than the other regions but showed the
fastest acceleration from the 1960s to the 1970s.
Much of the future acceleration in population growth rates will come from
the very low-income countries, which averaged only 2.4 percent annually for
1961-80 but which are projected to accelerate sharply, to 3.0 percent, for the
rest of the century. The projected average rates for the low-income and middle-
income categories for 1980-2000 both exceed 3 percent a year and are about
equal to their rates during the 1970s.
Consistent with rapid rates of urban growth, sub-Saharan Africa's agri-
cultural labor force expanded at a rate of only 1.7 percent annually from the
early 1960s to the late 1970s. That rate, much less than the population growth
rate, was revealingly about the same as the growth rate of food production.
The proportion of the total population in the agricultural labor force (active

Defining the Food and Nutrition Problem

agricultural workers relative to total population) decreased 5 percentage points,
reaching 30 percent by the late 1970s. This decline, reflecting the effect of an
increasing dependency ratio and out-migration from agriculture, requires either
increased labor productivity in agriculture, decreased per capital consumption,
or large increases in imports. Of course it is the latter that has ruled in Africa.
The lowest percentage of the population in the agricultural labor force was in
West Africa, whose ratio of agricultural workers to total population was down
to 27 percent in 1976-80, compared to about 33 percent in central Africa. The
biggest relative decline was, however, in eastern and southern Africa, where
the proportion declined by 7 percentage points from close to 40 percent in the
early 1960s. All three subregions showed nearly equal growth rates of the
agricultural work force.
The inverse relationship between the relative size of the agricultural labor
force and its per capital income is seen clearly from data among income cate-
gories for both reference periods. In 1976-80 the proportion of agricultural
workers in the population ranged from over 35 percent in the very low-income
countries to 25 percent in the middle-income group. Thus we see in Africa
a rapid increase in the relative size of the nonagricultural labor force but little
or no capacity to support that labor force from increased domestic food

Table 2.2 Population growth and agricultural labor force in sub-Saharan Africa, various peri-
ods, 1961-2000 (percent)

Agricultural labor force

Population growth rate Ratio to total
Growth rate population
Population 1961-70 1971-80 1980- 1961-65 to
characteristic 2000" 1976-80 1961-65 1976-80

West 2.7 3.1 3.3 1.7 33 27
Central 2.0 2.7 3.0 1.6 36 32
Eastern and 2.9 3.1 3.4 1.7 38 31
GNP per capital
Under $250 2.2 2.6 3.0 1.6 40 36
$250-$499 2.6 3.2 3.3 1.8 37 32
$500 and 3.0 3.3 3.4 1.7 31 25

Sources: U.N. 1983; FAO 1952-82.
"Based on medium-variant projections.
bEstimates are for midpoints of indicated periods.
cSee note d to table 2.1 for countries in each region.
d1980; based on trend estimates of real GNP during 1961-80 (1979-81 = 100).

The Evolving Food Situation


Nearly half of the countries in sub-Saharan Africa achieved annual increases in
food output, averaging 2 percent or more for the period 1961-80 (table 2.3).
The poor performance for the region as a whole was largely due to growth of
less than 1 percent a year in West Africa. More than one-half of the total
increase in food production came from eastern and southern Africa, about one-
fourth from central Africa, and 22 percent from West Africa (table 2.4). Root
crops dominated the production increases of West Africa and central Africa,
while maize provided much of the growth in eastern and southern Africa. Half
of the increase in rice production came from West Africa and about 40 percent
from eastern and southern Africa.
The two relatively bright spots in food production were maize and rice:
maize contributed about 30 percent of the increase in food production, and rice
12 percent. The contribution of rice was striking, since this crop represented
only 5 percent of production of the major food crops in the early 1960s. But the
maize contribution of nearly a third of output growth from one-fifth of the area
is a point to which later chapters will return.
Millet and sorghum were the poor performers. Largely influenced by West
African trends, regional production of these leading crops of sub-Saharan
Africa increased only 0.7 percent annually for the 1961-1980 period. These
two crops accounted for 27 percent of major food crop production in 1961-65
but provided less than 10 percent of output growth through the 1970s. The
relative contribution of millet and sorghum to production growth was far below
the 1961-65 percentage output share of these crops in all subregions-about
half that in eastern and southern Africa, two-fifths in West Africa, and less than
a fifth in central Africa.
The noncereal food crops, as a group, maintained their share of production
growth. Roots and tubers alone accounted for one-third of the total increase in
output, somewhat more than their initial share in output. But that still repre-
sented a decline in per capital production and probably no increase in factor
In keeping with generally poor performance in export crops in Africa, the
output of groundnuts in West Africa, where 80 percent of the region's crop was
produced in the early 1960s, declined drastically, by about 2 percent per year.
Nearly one-fourth of this decrease was offset by output gains in central Africa
and eastern and southern Africa. Thus this export crop did especially poorly in
the region in which food crop production did worst and relatively better in
regions where food crop production did better.

Table 2.3 Production and area harvested of major food crops in sub-Saharan Africa, various periods, 1961-80 annual growth (percent)

Total West Central Eastern and southern

Area Area Area Area
Crop Production harvested Production harvested Production harvested Production harvested

All major food crops" 1961-80 1.7 1.4 0.8 1.1 2.9 3.3 2.4 1.2
1961-70 2.2 2.4 1.1 2.2 4.4 4.8 3.0 2.2
1971-80 1.6 0.8 1.9 1.1 1.8 2.3 1.4 0.3
Cerealsb 1961-80 1.8 1.2 1.0 1.1 2.2 2.7 2.6 0.9
1961-70 1.9 1.8 0.5 1.2 3.4 3.3 3.2 2.4
1971-80 1.8 0.9 2.6 1.8 1.6 2.3 1.2 -0.6
Millet and sorghum 1961-80 0.7 0.9 0.4 1.1 1.0 0.9 1.5 0.0
1961-70 0.0 1.2 -0.7 1.2 0.2 0.4 2.3 1.7
1971-80 1.9 0.9 2.2 1.5 2.4 1.3 1.0 -1.8
Maize 1961-80 2.7 1.7 1.3 0.9 2.5 3.4 3.1 1.7
1961-70 3.4 2.7 2.6 0.6 4.7 3.9 3.4 3.3
1971-80 1.4 1.3 3.8 3.3 0.8 2.9 0.8 0.1
Rice 1961-80 3.6 2.8 3.9 2.7 7.2 7.2 3.1 2.3
1961-70 5.1 3.2 5.3 3.0 < < 4.4 2.0
1971-80 2.8 2.3 3.5 2.2 4.8 1.5 1.8 2.7
Noncereals 1961-80 1.7 1.8 0.7 1.1 3.2 3.7 2.3 1.9
1961-70 2.7 3.7 1.8 4.0 4.8 5.8 2.6 1.7
1971-80 1.4 0.7 1.0 0.0 1.9 2.3 1.8 0.7
Roots and tubers 1961-80 2.3 2.4 1.9 1.3 2.9 3.8 2.5 3.1
1961-70 2.9 2.3 2.5 1.1 4.4 6.0 1.9 1.2
1971-80 2.3 2.3 1.7 1.4 2.3 2.6 3.7 3.6
Pulses 1961-80 2.7 2.6 2.7 3.3 3.7 2.8 2.4 1.4
1961-70 4.6 6.1 5.3 9.0 6.1 4.6 3.5 2.0
1971-80 2.4 0.7 5.6 1.3 0.8 1.7 0.9 -0.8
Groundnuts 1961-80 -1.9 -0.6 -3.7 -2.0 4.4 4.9 0.7 0.6
1961-70 0.7 2.4 -0.7 1.7 7.6 6.9 3.9 3.0
1971-80 -3.6 -2.6 -3.8 -3.9 -0.3 2.3 -5.6 -2.9

Source: FAO data tapes 1975, 1979, 1981, and 1983.
"Does not include bananas and plantains, for which data on harvest area are not available.
bIncludes wheat and other cereals.
cMore than 10 percent.

Table 2.4 Distribution of major food crops, 1961-65, and production increase, 1961-65 to 1976-80, in sub-Saharan Africa

Production increase 1961-65 to 1976-80b

Distribution 1961-65





Eastern and

Eastern and
Total West Central southern Mmt Percent Mmt Percent Mmt Percent Mmt Percent

All major food crops 100 100 100
Cereals 55 53 27
Millet and sorghum 27 40 11
Maize 18 8 14
Rice 5 4 1
Wheat and other 5 1 1

Root crops
Bananas and

45 47 73
27 27 52
9 14 5
5 4 6
4 2 10

16.70 100 3.75 100 4.28 100 8.67 100
9.20 55 2.20 59 0.74 17 6.26 72
1.45 9 0.59 16 0.08 2 0.78 9
4.94 29 0.54 14 0.49 11 3.91 45
1.96 12 1.02 27 0.18 4 0.77 9
0.85 5 0.06 2 -0.02 < 0.80 9
7.50 45 1.54 41 3.55 83 2.41 28
6.29 33 2.54 33 2.34 55 1.41 16
-1.19 -1.62 0.38 8 0.09 1
1.53 8 0.54 7 0.35 9 0.62 7
0.86 4 0.09 1 0.48 11 0.29 4

Source: FAO data tapes 1975, 1979, 1981, and 1983.
"For purposes of aggregation, rice is in husked form and the noncereals are converted to cereal equivalents based on calorie content.
bAbsolute production increase is in million metric tons; parts may not add up to totals due to rounding.
cNegative; positive contributions within the group are adjusted accordingly.

Defining the Food and Nutrition Problem


About 80 percent of the growth in production of basic food staples in sub-
Saharan Africa during the past two decades arose from an average increase of
1.4 percent a year in harvest area. The picture is one of a growing population, a
major portion of the growth leaving the countryside for the city, and the larger
rural population expanding into the land area about proportionally, with little
change in labor productivity. Growth of food output in West Africa and central
Africa was due entirely to expansion of agricultural area, since output per
hectare declined in both subregions during that period. Even in more productive
eastern and southern Africa, nearly half of output growth came from increases
in harvested area.
Among the major food crops, expansion in harvest area in sub-Saharan
Africa was the source of almost two-thirds of the growth in cereal output and all
of the increase in noncereals. However, the individual cereals showed varied
trends. With declining crop yields, the small increase in the production of
millet and sorghum came wholly from the 0.9 percent annual expansion in
harvest area. More than three-fourths of the output growth of rice was due also
to increases in area. Even in the case of maize, for which yield increase was
relatively most important, area expansion was the source of 65 percent of
production growth.
Food production increases in sub-Saharan Africa dropped from 2.2 per-
cent a year during the 1960s to 1.6 percent a year in the 1970s, as growth of
harvest area fell from 2.4 percent to less than 1 percent. These changes appear
to reflect trends in central Africa and eastern and southern Africa more than in
West Africa. Despite an overall improvement of output per hectare in eastern
and southern Africa, the major slowdown of the growth of harvested area in this
subregion led to a major reduction in production growth from the 1960s to the
1970s. In contrast, growth rates in the output of basic food staples in West
Africa, although still slow, rose in the 1970s because of increases in the produc-
tion of maize, millet, and sorghum. Despite a much slower growth of harvested
area and the decline of crop yield in central Africa, production expanded close
to 2 percent a year in the 1970s.


During 1966-80, total domestic utilization of basic food staples in sub-Saharan
Africa increased an average of 2.2 percent a year (table 2.5).2 That is over one-
fifth faster than the food production growth rate, but still one-sixth slower than

2. Data on domestic utilization, which includes seed, feed, and waste, together with its
consistent companion set of statistics on food trade, are from FAO 1982a.

The Evolving Food Situation

the population growth rate. Imports financed commercially and through food
aid account for the difference between production and consumption. Consump-
tion of major food crops directly for food expanded at a slightly faster rate of 2.5
percent, which raised its share of total domestic utilization in 1966-70 by two
percentage points to 73 percent in 1976-80.
Although use of major food crops for animal feed expanded at more than 3
percent a year, it remained at only 6 percent of total domestic utilization. This
confirms that livestock feeding in the region is still largely dependent on open
range and waste products. Feed use in central Africa expanded at a fast 3.7
percent a year, but its relative share of total domestic utilization remained at 4
percent, which was the smallest among the subregions.
With a rate of growth of 2.7 percent a year, the proportion of major food
crops used as feed in eastern and southern Africa remained unchanged during
the 1960s and 1970s. Although the share of feed in total domestic utilization
was still low compared to other areas in the Third World, it was significantly
higher than in West Africa and central Africa. Eastern and southern Africa
accounted for half of the basic food staples used for animal feed in sub-Saharan
Africa in 1976-80.


Sub-Saharan Africa has moved rapidly from a net exporter to a substantial and
rapidly growing importer of basic food staples. Net exports of basic food
staples by sub-Saharan Africa in 1966-70 averaged 1.3 million tons a year
(table 2.6). Noncereals, including groundnuts, accounted for almost 75 percent
of exports, and cereals 95 percent of imports. Slow growth of production and
rapid increase in consumption reversed this favorable trade position. Food
exports declined 7 percent a year, and imports rose 9 percent a year. By the late
1970s, sub-Saharan Africa was a net importer of 4.4 million tons of staple food
annually. Net imports had surged to some 10 million tons by the mid-1980s.

Table 2.5 Average annual growth rate of domestic utilization of major food crops in sub-
Saharan Africa, 1966-80 (percent)

Region Total Food Feed" Other/

Sub-Saharan 2.2 2.5 3.1 1.3
West 2.3 2.7 3.6 1.1
Central. 2.7 2.7 3.7 2.8
Eastern and 1.9 2.0 2.7 1.3

Source: FAO 1982a.
"Includes bran and cakes.
bSeeds, allowances for waste, and amounts for nonfood uses.

Table 2.6 Trends in exports and imports of basic foods in sub-Saharan Africa, 1966-70 to 1976-80

Relative change
(percent) 1966-70 to

Annual growth rate
(percent) 1966-70 to

Annual net exports of basic foods
(million metric tons)


Imports Cereals

Noncereals Cereals



Sub-Saharan Africa
Eastern and
GNP per capital
Under $250
$500 and over

-52 +140



Source: FAO 1982a.
Note: Includes trade in bran and cakes for feed use.
"Calculated between midpoints of periods.
b1980; based on trend estimates of real GNP during 1961-80 (1979-81





The Evolving Food Situation

The shift in the food trade position can be traced particularly to West
Africa, whose food exports, principally groundnuts and pulses from Nigeria
and Niger, declined sharply, while food imports tripled. West Africa contrib-
uted virtually all of sub-Saharan Africa's net exports of basic food staples in
1966-70, but by 1976-80 it accounted for more than half of the region's net
food imports.
Central Africa increased its net food imports more than sixfold from the
late 1960s to the late 1970s. Food trade was much smaller than in the other
subregions; however, net cereal imports were up nearly fourfold, while net
noncereal food exports decreased to one-fourth of 1966-70 levels. Eastern and
southern Africa was just self-sufficient during 1966-70, but shifted its food
trade position to net imports by 1976-80. Net imports of basic food staples
averaged 1.3 million tons during the later period. Annual rates of decrease in
food exports (4.4 percent) and increase in food imports (6.2 percent) were
slower than in West Africa and central Africa. Unlike in the other subregions,
cereals predominated in eastern and southern Africa with respect to food
From 1966-70 to 1976-80, net food imports of the very low-income
countries tripled, while the two upper-income groups shifted from net food
exporters to net food importers. Food imports of the very low-income countries
expanded at 10 percent a year, while their food exports remained about un-
changed. Food aid was a substantial part of food imports of this group. Food
imports also grew at 10 percent a year in the middle-income countries. This was
accompanied by an equal rate of decline in food exports, largely as a result of
the major drop in output of groundnuts. The food trade position of the middle-
income category was reversed from being a net exporter of a half million tons a
year in 1966-70 to a net importer of 2.7 million tons in 1976-80. The low-
income countries had more moderate growth rates in food trade than the mid-
dle-income group. Their food imports increased 8 percent a year, and exports
declined 6 percent. During 1976-80, the group's net exports of noncereal food
staples offset more than half of their net cereal imports.


Simple projections of past trends serve two useful purposes. First, by magnify-
ing and illustrating current imbalances, they indicate where the past is taking us
if we do not succeed in changing the underlying forces. Second, they probably
give a better indication of the present situation than current numbers-the latter
being subject to the vagaries of individual years, and the trend representing a
smoothing over a substantial period of time.
If one simply projects the slow production growth trends of 1961-80,
food output in sub-Saharan Africa would reach only 114 million tons by 2000

Table 2.7 Production and consumption of major food crops in sub-Saharan Africa, 1976-80, and projections to 2000 (million metric tons)

1976-80 averages

Projections to 2000



Production Consumption Gross Net Productionb Consumption" Gross Net

Sub-Saharan Africa
Eastern and
GNP per capitald
Under $250
$500 and over

5.7 3.0 113.7

"Negative values denote surpluses.
bBased on extrapolations of country rates of production during 1961-80.
cAssuming the continuation of the 1966-80 trends in GNP per capital and using the U.N. medium-variant population projections to 2000.
"1980; based on trend estimates of real GNP during 1961-80 (1979-81 = 100).





53.7 46.8

36.3 36.3
3.7 0.1
13.7 10.4

6.4 2.8
13.3 11.1
34.0 32.9

The Evolving Food Situation

(table 2.7). Output of basic food staples in the region would fall about 47
million tons short of projected demand by the end of the century, and the net
deficit would be 15 times the estimated annual average for 1976-80.3 Produc-
tion shortfalls in 36 of the countries would total over 50 million tons, more than
45 percent of which is for Nigeria. Assuming that the past trend of regional
trade in noncereals has continued, data on cereal trade indicate that net imports
of basic food staples in 1984 were roughly on their "deficit" projection line to
the year 2000.
More than three-fourths of the net deficit in 2000 would be in West Africa,
where the projected production shortfall of 36 million tons is 11 times the
annual average in the late 1970s. Central Africa's aggregate surpluses of over
3.5 million tons would almost equal the aggregate shortfalls of that subregion.
In eastern and southern Africa, the projected gross surplus of another 3.5
million tons would offset only a fourth of the subregion's projected aggregate
deficits, leaving a net shortfall of more than 10 million tons.
About 70 percent of the region's projected net food deficit by the end of the
century would be in the middle-income group of countries, again dominated by
Nigeria's production shortfall. The very low-income group would face a net
deficit of about 3 million tons. Projected food demand by the low-income
category would exceed projected food output by about 11 million tons. Both the
very low-income and low-income groups had slight food surpluses in 1976-80.


Trends in the use of agricultural imputs in sub-Saharan Africa provide clues to
the slow increase of food output in the past two decades and suggest pos-
sibilities for accelerating production growth in the future. Compared to other
regions in developing countries, sub-Saharan Africa has little cropped area per
worker, little irrigation, little use of fertilizer, and few tractors. It is no wonder it
has performed poorly compared to other regions. The following four points,
drawn from tables 2.8 and 2.9, summarize the input situation.
First, despite the abundance of land relative to population, the number of
hectares of cropped area per farm worker is small compared with that in other
developing areas. Estimates for 1976-80 are 1.5 hectares in West Africa and
1.2 hectares in the two other subregions. About 30 percent of the countries had
less than 1 hectare per agricultural worker. Chapter 9 discusses the labor
productivity relationship that helps to explain this situation. Second, the area

3. Consumption projections by country assume a continuation of the trends in per capital
incomes during 1966-80, the income elasticity of demand for meat as a proxy for that of feed use,
and the U.N. projections of population growth rates. Output projections, also by country, are based
on extrapolations of the 1961-80 trends, except for countries with declining growth rates, for
which no further decrease from the 1980 trend estimate of production is assumed.

Defining the Food and Nutrition Problem

under irrigation had expanded 3.5 percent a year from 1961-65 to 1976-80.
Eastern and southern Africa contain about 70 percent of the total irrigated area
in the region. However, only about 1 percent of the total cropland in sub-
Saharan Africa area was irrigated during 1976-80.
Third, fertilizer consumption was only about six kilograms per cropped
hectare in the late 1970s, although it increased 10 percent a year from 1961-65
to 1976-80. Eastern and southern Africa, which accounted for about 60 percent
of reported use in 1976-80, had an annual increase of 7.5 percent. About a third
of the countries used less than a kilogram per hectare. Fourth, use of agri-
cultural tractors in sub-Saharan Africa was less than 1 per thousand hectares of
cropped area in the late 1970s. Tractor numbers in use per thousand hectares
averaged 1.5 in eastern and southern Africa, 0.7 in West Africa, and only 0.3
in central Africa, with half of the countries showing less than 0.5. However,
the number of farm tractors increased over 6 percent a year from 1961-65 to
Utilization of modern agricultural inputs in sub-Saharan Africa is much
lower than in other developing areas. India, for example, has a cropland area
1.3 times that of sub-Saharan Africa, 22 percent of which is irrigated. Fertilizer
use is 26 kilograms per hectare, and there are two tractor units per thousand
hectares-five times and two times, respectively, the number for sub-Saharan
Africa. The very low levels of input use help explain the relatively low produc-
tivity of agriculture in sub-Saharan Africa. At this initial low base of use, even

Table 2.8 Cropland area in relation to selected indicators in sub-Saharan Africa and selected
non-African countries, 1976-80

Agricultural Agricultural
labor force Irrigated Fertilizer tractors
(hectares per area consumption (per thousand
Region or country unit)" (percent) (kgm per hectare) hectares)

Sub-Saharan Africa 1.3 1.1 5.6 0.8
West 1.5 0.7 4.1 0.3
Central 1.2 0.1 2.7 0.7
Eastern and 1.2 2.1 8.9 1.5
Other countries
India 1.0 21.8 26.3 1.9
Indonesia 0.7 27.9 42.0 0.6
Algeria 3.8 4.1 23.7 6.4
Turkey 2.7 7.3 46.6 12.9
Brazil 3.8 2.6 58.7 5.3
Mexico 3.3 19.6 48.1 5.4

Sources: FAO 1952-82; FAO 1961-81.
Note: Cropland area includes arable land and land under permanent crops.
"Inverse of calculated ratio.

Table 2.9 Annual growth in arable land, cropland, agricultural labor force, irrigated area, fertilizer consumption, and agricultural tractor use in sub-Saharan
Africa, 1961-65 to 1976-80 (percent)


Sub-Saharan Africa
Eastern and

Arable Agricultural
land Cropland" labor force

Irrigated Fertilizer Agricultural
area consumption tractors

1.1 1.3

Sources: FAO 1952-82; FAO 1961-81.
"Includes arable land and land under permanent crops.

38 Defining the Food and Nutrition Problem

the moderately rapid growth rates of input utilization has yet to improve overall
crop yields, which have remained stagnant or have even declined during the
past two decades.


Food projections to the year 2000 suggest that filling the projected gap of basic
food staples in sub-Saharan Africa from domestic production would require
more than twice the 1.8 percent annual growth rate of 1961-80. Such a rate is
unlikely to be achieved.
The data suggest the need for radical acceleration in the growth rate for the
inputs of modern agriculture; a special problem of accelerating growth in West
Africa; a major role for maize and rice in future production growth; and a major
need for improvement in sorghum and millet production because of their heavy
weight in production-and despite their poor past record.

chapter 3

The Nutrition Situation

and Its Food Policy Links

Shubh K. Kumar

Malnutrition is generally not perceived to be a pervasive problem in much of
Africa, except during famines. There is little doubt that in "average" years,
malnutrition in Africa is a smaller problem than it is in Asian countries such as
Bangladesh, India, Nepal, or Sri Lanka. Another reason why nutritional prob-
lems do not receive much attention in Africa is that in recent years episodes of
severe drought, famine, and starvation have been relatively frequent, and atten-
tion has focused more on the shorter-run alleviation of food problems than on
longer-term underlying problems. Finally, the proportion of the undernourished
has been almost uniformly lower in urban centers than in rural areas due to a
variety of income subsidies, better health, water supply, and sanitation facili-
ties. Because of dispersed habitations and less variation in living standards, the
rural problem is seldom visible except when reports of severe famine surface.
Although numerous community-level nutritional surveys are available in
Africa, differences in sampling and measurement make assessment of the
distribution and nature of the nutritional problem and comparisons between
countries over time difficult. However, several recent national nutritional sur-
veys for individual countries have identified some broader dimensions and
characteristics of this problem.
Table 3.1 summarizes information from the 12 African countries for
which surveys are available. The data indicate that for preschool children,
chronic undernutrition affects nearly a fourth of the population, but tends to be
less variable across countries than acute undernutrition. Chronic undernutrition
occurs when nutritional deficits extend over time; in children, it can reduce
growth rates and attained height for age. Acute undernutrition results from a
substantial nutritional deficit of short duration, in which growth patterns may or

Table 3.1 Preschool child malnutrition in selected African and Asian countries (percent)

undernutrition" Acute undernutritionb Underweight'
Country and
survey year Rural Urban Rural Urban Rural Urban

23.8 12.7

22.4 15.7

20.2 13.8

Sierra Leone
Sri Lanka

9.9 5.2

0.7 23.0"

1.7 25.5d

0.8 16.5

2.4 32.4

.. 9.1
.. 8.1


... 16.1 to 26.4



. 21.6

. 47.7 to 62.3/'

. 10.9g

. 42.0



Sources: USAID 1975, 1976a, 1976b, 1976c, 1977a, 1977b, 1977c; USHEW 1974, 1975; Kenya
1979; Philippines 1978; FAO 1974.
"Children below 90 percent of reference height for age.
bChildren below 80 percent of reference weight for height.
cChildren below 75 percent of reference weight for age.
dChildren below 80 percent of reference weight for age.
eSurveys covered only the rural sedentary population of that part of each country estimated to be
most affected by the drought. The affected zone varied from a relatively small part of
Burkina Faso to nearly all of Niger. Geographical coverage for the area was about one-third.
/Range of children below 70 percent of reference weight for age.
gSurvey covered rural and urban children and showed that 10.9 percent of the total were under-





20.5 11.4 2.3

26.6 13.8 3.2

... 11.4

S- 10.7
.. 5.3

.. 9.9
.. 6.1

.. 12.1

The Nutrition Situation and Its Food Policy Links

may not be normal, but weight-for-height ratios are reduced. Increased physical
growth and voluntary physical activity are both by-products of improved diet
and health standards.
It is misleading to state that physical growth can occur at the cost of
physical activity, or vice versa, since over extended periods of time the two
should be synergistic. In the short run, the degree to which dietary or health
deprivation affects growth rates is influenced by change in physical activity.
However, the measures of acute undernutrition or shorter-run deprivation are
found to be closely related to year-to-year food availability in sub-Saharan
Africa. This is in contrast to countries with a large proportion of rural landless
and a relatively diversified rural economy. Surveys made in the Sahel in 1974
and Zambia in 1970-72, when food shortages were severe, show sharply
higher levels of undernutrition than in less-favorable years.
The national surveys also indicate that rural areas are consistently worse
off than urban areas. This observation is contrary to a widespread belief that
urban malnutrition is more severe.' Furthermore, none of the surveys that were
conducted during severe food shortages covered urban areas, and rural-urban
differences in nutrition are likely to be much more pronounced during these
periods. The surveys also suggest that hard core, chronic undernutrition in most
sub-Saharan countries is relatively lower than in the South Asian countries for
which comparable figures are available.
While Africa's stagnant economic growth and the pressures to eliminate
heavy urban subsidies may lead to deteriorating urban nutrition, it may be just
as likely that these factors have a dampening effect on rates of rural-urban
migration. Any answer to the question of rural-urban distribution of the nutri-
tional problem and changes therein is likely to center on the characteristics of
migration and the relative attraction of urban areas. If, in contrast to parts of
Asia and Latin America, rural poverty is less severe in Africa and the pull of
cities greater than the push from rural areas, then might one expect the rural
nutritional problem to remain larger than the urban problem? This is a ques-
tion that requires further investigation, and there are many factors associated
with urbanization that influence nutritional levels apart from migration pat-
terns and rural-urban income differentials.
Cross-sectional nutritional surveys do not reflect year-to-year and seasonal
fluctuations in nutritional status. These temporal fluctuations are of crucial
importance in Africa. Not only are they interrelated and the result of agri-
cultural production characteristics, but there may also be a causal link between
high seasonal and year-to-year production fluctuations contributing to the rela-
tive stagnation in production trends. The problem of seasonal variability is

1. This may be the case if only low-income urban households or clinic samples are taken, as is
often the case in smaller studies.

Defining the Food and Nutrition Problem

especially pronounced in single-crop-cycle agriculture. A single bad harvest
not only deprives the household of adequate food that year but may also limit its
production capacity for the next year. Migration and cash incomes help to
stretch food supplies but may also deplete the household labor supply essential
for production.
Diseases such as onchocerciasis, bilharzia, and malaria also take a heavy
toll on human resources. The seasonality of malaria coincides with periods of
heavy labor demand, creating a serious problem especially in those areas with
single-crop-cycle agriculture. The tsetse fly also continues to impede livestock
development in large areas of Africa, further accentuating the dependence on
human labor for agricultural production.


The relation of undernutrition (or overnutrition) and food availability, while
best demonstrated at the individual consumption level, also holds true in vary-
ing degrees at area and national levels. As in all biological and social rela-
tionships, there are many intervening variables: disease and individual meta-
bolic variations at the individual level; purchasing power, its composition, and
intrahousehold allocation at the household level; and marketing and food dis-
tribution mechanisms at regional and national levels. The greater the influence
of intervening factors, the lower the potential for meeting food needs of people
by solely increasing aggregate food supplies. Endemic diseases, highly skewed
income distributions, and inadequate food distribution mechanisms, for exam-
ple, could lead to a situation in which aggregate food self-sufficiency overlaps
with a high prevalence of undernutrition. Despite these caveats, however, the
underlying relationship between food supplies and nutrition cannot be argued
Low levels of aggregate food supplies per capital raise levels of acute
malnutrition, other things being equal. Analysis of large-scale nutritional sur-
veys for 11 African countries indicates that greater aggregate food availability,
under the prevalent conditions of smallholder food production, is typically
correlated with improved nutritional status for the population as a whole
(Kumar 1981). There is also evidence that the nutritional status of populations is
affected by large year-to-year fluctuations in food production. Surveys con-
ducted during drought years in the Sahel and in Zambia show very high preva-
lence of child malnutrition (table 3.1). By 1975, when the Sahel crisis was
easing and large-scale food distribution was under way, the problem of acute
malnutrition diminished. Scattered evidence from village studies in semiarid
areas of Africa also indicates large year-to-year fluctuations in the prevalence of
child malnutrition.
Even if good national-level statistics were available, they would tend to

The Nutrition Situation and Its Food Policy Links

understate variability, since acute shortages tend to be localized. But even at the
national level, variability is high: Koester (1984) estimates that for both the
Sahel and the countries of the Southern African Development Coordination
Conference (SADCC) collectively (Angola, Botswana, Lesotho, Malawi,
Mozambique, Swaziland, Tanzania, Zambia and Zimbabwe) the degree of
instability in cereal production during 1960-80 was about 50 percent higher
than in other parts of sub-Saharan Africa, and more than three times that of the
rest of the world. Production instability was found to be even higher for
individual countries in the Sahel and SADCC.
For individual countries, the type and quality of food statistics generally
reflect the national policy environment for the agricultural sector. In Zambia,
for example, state marketing for maize, the major staple food crop, has been
emphasized. Here, good year-to-year statistics are available on purchases by
the marketing parastatal and its affiliated cooperatives. In addition, annual
surveys have been conducted since the 1970s to obtain production and market-
ing information from the noncommercial sector of agriculture. Though private
maize sales do exist, they are insignificant in urban food supply as a result of
Zambia's pricing and marketing policies for maize (Kumar 1984). The degree
of interyear variability in Zambia's food availability is reflected in the consider-
able fluctuation in year-to-year purchases of maize by the marketing parastatal
and cooperatives. During the seven-year period preceding 1980-81, fluctua-
tions from year-to-year range from -43 percent to 48 percent (table 3.2).
Traditional farmers in Zambia characteristically rely little on purchase of
food grains. Marketed production goes almost entirely for urban consumption.
Even though commercialization has been increasing, the majority of farmers
are expected to retain a high proportion of their production. In the early 1970s,
household purchases accounted for only 16 percent of maize consumption in
rural areas; the situation probably has not changed significantly. A survey of a
maize-producing region in Eastern Province in 1981-82 found that 62 percent

Table 3.2 Maize purchased by official marketing organizations in Zambia

Change from
Marketing year Maize purchased previous year
(ending 30 April) (90 kilogram bags) (percent)

1973-74 4,435,019
1974-75 6,534,333 48
1975-76 6,216,455 -5
1976-77 8,333,022 34
1977-78 7,738,347 -7
1978-79 6,462,847 -16
1979-80 3,732,879 -43
1980-81 4,247,404 14

Source: Zambia 1980.

Defining the Food and Nutrition Problem

of rural maize purchases were made by small farm households. Defined as those
with less than three acres of total planted area, they constituted nearly half of
rural households. However, as a group they purchased only 4 percent of their
annual maize consumption. It is only along the major transport routes that a
secular decline in on-farm storage and an increasing dependence on purchased
maize meal can be observed.2 Therefore, most of the year-to-year shortfalls in
purchased grain by the parastatals and cooperatives shown in table 3.2 is
reflected in reduced supplies for urban areas. While this poses a potentially
serious food security concern for urban populations, shortfalls are generally
foreseen many months ahead, and hence imports usually close the deficit.
In rural areas, on the other hand, interyear variability of food production is
a real problem. A relatively small proportion of farmers market the bulk of
maize sold in any given year (as indicated by a collaborative survey in Zambia's
Eastern Province by the International Food Policy Research Institute, the Na-
tional Food and Nutrition Commission, and the University of Zambia). The
Eastern Province is the third largest maize-producing province in Zambia but
has a very small proportion of large-scale "commercial" farms, usually defined
as operating 40 acres or more. In 1981-82, only 16 percent of farmers, those
with cultivated farm sizes over 7 acres, marketed 93 percent of all maize sold in
one year. Only one-fourth of maize produced was sold. Figure 3.1 shows the
pattern of maize sales and retentions with increasing household maize produc-
tion. Households producing less than 2.5 tons of maize-close to the subsis-
tence requirement of the average household-sell practically no maize.
It is likely that for the aggregate of Zambian smallholder agriculture, both
the price elasticity of marketed surplus and the elasticity of marketed surplus
with respect to production are higher than the corresponding elasticities for
household retentions, as seen for small farmers elsewhere (Krishna 1967;
Toquero et al. 1975; Haessel 1975). However, since a majority of farmers
market very small quantities, for them even complete elimination of sales
would hardly buffer the nutritional effects of a poor harvest. Furthermore, for
this lower-income majority, cash incomes and assets are low to begin with,
which makes production shortfalls harder to deal with. For this group, the
elasticity of home consumption with respect to food production will be close to
unity, and hence their consumption will be closely related to production
Larger farmers, who contributed 93 percent of marketed maize, will have
an elasticity of marketed surplus with respect to production substantially great-
er than unity-and much larger than that for the small farmers-implying that
they are in a better position to buffer their home consumption from production

2. The original rail line bisects the country north to south down the central plateau, connecting
the mining areas of the copper belt to other parts of southern Africa. It is along this rail line that
agriculture was originally promoted; it is well situated for access to markets and infrastructure.

v Maize retention 1981
o Maize sales 1981

0 1 10 100 1,000 10,000
Production (kilograms)
Source: Collaborative IFPRI study in Zambia. Production (kilograms)
Note: The majority of households produced less than 2.5 tons.

Figure 3.1 Household disposal and retention of maize, Chipata District, Zambia, 1981 (thousand tons)




U 100

- O

0, 10
l -

Defining the Food and Nutrition Problem

It is unfortunate that, though considerable effort has gone into analysis of
aggregate response and farm-size-specific response of marketed surplus in
Asia, relatively little attention has been paid to household subsistence strategies
and the link between retentions, retail access, and off-farm incomes (Behrman
1966). In the African context, these should be of major importance in explain-
ing marketing behavior, and neglect of them precludes the full recognition of
the subsistence consequences of production fluctuations.


Intrayear fluctuations in nutritional status are particularly relevant in Africa
because of the large areas under a single annual crop cycle. Most of the northern
and southern savannas, composing the bulk of arable land in Africa, depend on
monocropped agriculture due to a unimodal rainfall pattern. Since irrigation
facilities are not widespread outside the large-scale commercial farming sector,
a single crop must provide food needs for the entire year. Livestock (in tsetse-
free areas), trading, and other nonagricultural activities can improve food
security between harvests. Because of the long storage period required and
significant losses from pests, adequate farm storage of grain is problematic for
many households.
Schofield (1979) analyzed available data from 25 in-depth studies from
various parts of western Africa to examine seasonality in food consumption.
She found that villages with a unimodal rainfall distribution had relatively
higher levels of caloric and protein intakes than those with a bimodal rainfall
distribution but also had much more seasonal fluctuation. For all villages, per
capital calorie intakes decreased from a postharvest dry season level of 2,012
calories to a wet season low of 1,885. However, practically all of the seasonal
fluctuation in the sample was accounted for by the 15 villages (in northern
Nigeria, Senegal, and Burkina Faso) where there was unimodal rainfall. Here,
dry season calorie consumption was 2,458 calories per capital, but wet season
consumption fell to 2,191 calories, an estimated 88 percent of requirements.
Although seasonal fluctuations may be lower in the bimodal rainfall areas,
excessive reliance on root crops and the associated problems of dietary bulk and
palatability contribute to very low aggregate consumption of calories and
The decline in consumption in the wet season in the unimodal rainfall
areas occurs at a time when nutritional requirements are increasing due to the
peak labor demand for agriculture. Increased disease incidence during the rains
can also raise nutritional requirements. It has been observed that when house-
hold food stores are adequate, food consumption increases to meet the higher
seasonal energy requirements. Simmons (1981) found this to be the case in
households of all farm sizes in northern Nigeria. Her study indicates that

The Nutrition Situation and Its Food Policy Links

despite the increase in consumption of well-off households, energy needs to
meet seasonal labor demands in some cases still exceeded consumption, lead-
ing to a net shortfall. Cattle ownership, however, made households less suscep-
tible to consumption shortfalls. This is probably because livestock and dairy
sales provided a source of wet season income. Also, livestock ownership
typically is correlated with overall wealth and asset status.
Current IFPRI work in Zambia also indicates that heavy work requirements
can be a cause of decline in nutrition. Though the Eastern Province is one of the
better agricultural areas of the country, the unimodal rainfall pattern creates a
period of food scarcity for many households during the wet season. Food stocks
are drawn down during the planting and weeding seasons (December through
March), leading to a stretching of household food supply during this period of
heavy labor demand. This reduction in food stocks is reflected in a reduction in
body weight of working-age household members, despite an increase in their
caloric intake. Table 3.3 gives the seasonal distribution of weight-for-height
ratios by age groups as a percent of reference standards.
September, January, and May represent postharvest, rainy season, and
preharvest conditions, respectively. For both rural and the peri-urban samples,
average nutritional status declined between September and January and was
partially regained by May.3 For those under 15 years old, the pattern is similar
in urban and rural areas. However, for adults-especially the working-age
group of 15-55 years-urban and rural patterns differ, with only the rural
sample showing a decline. Between September and January, the incidence of
low weight for height (below 80 percent of expected weight for height) among
rural adults nearly doubled, increasing from about 16 percent to about 29
percent, while it was relatively unchanged at about 16 percent for the urban
adult population. There is an increased incidence of disease in both urban and
rural areas, especially malaria and diarrhea, during the rainy season. This
appears to be the main factor responsible for the decline in children's nutritional
status. Heavy farm work coinciding with disease and reduced food availability
appear to be responsible for the difference in rural and urban patterns for adults.


Seasonality in food availability and consumption coinciding with heavy work
and debilitating diseases have implications for year-to-year fluctuations in agri-
cultural production on the one hand and for prospects for longer-run production
increases on the other. Anthropologists and biological scientists have observed
the association of food shortages both with labor migration and with capacity to

3. This was a cross-sectional rural and low-income urban sample. Both groups had annual per
capital incomes approximately half of the national GNP per capital for 1981.

Table 3.3 Seasonal distribution of weight-to-height ratios as percent of reference standards, Chipata District, Zambia, 1981-82

Urban" Rural

Age group Sample size September January May Sample size September January May

6-17 months 12 101.7 97.9 99.9 17 95.2 89.1 91.5
18-35 months 16 93.5 93.4 94.9 16 95.4 95.0 95.6
36-59 months 20 103.3 94.2 96.8 39 99.3 92.9 92.6
5-10 years 39 92.4 87.4 89.5 43 94.3 89.2 92.3
10-15 years 17 99.3 93.6 93.6 56 98.5 92.5 95.6
15-20 years 12 102.5 103.0 103.6 32 100.0 94.6 96.1
20-55 years 40 103.7 100.5 100.5 104 95.5 90.7 92.8
55 years and over 3 89.6 83.5 85.9 21 91.5 85.2 86.8
Total 159 98.9 94.7 95.9 328 96.5 91.2 93.3

Source: Collaborative IFPRI study in Zambia.
"Low-income housing areas in the provincial town, which is classified as peri-urban due to its small population.

The Nutrition Situation and Its Food Policy Links

carry out heavy agricultural work (Richards 1939; Fox 1953).
Current IFPRI research in Zambia also indicates an association between
household food availability and capacity for agricultural production. Details
of current planting were obtained in 1981 for all the rural households in the
survey. In table 3.4, the size of food crop, cash crop, and total area planted are
seen to be related to the duration of the previous year's maize supply. The total
area planted by a household was about twice as much if maize supply lasted
beyond January than if it did not. Households whose maize supply lasted
beyond January also planted substantially more cash crops.
It should be noted that at the time there was no land constraint for any of
these households. They brought new land under cultivation at will, and most
households kept some land fallow, even though continuous cultivation of near-
by land was almost universally practiced in this area. Therefore, while the
previous year's plantings and performance would be reflected in duration of
stocks, there is no a priori reason why current and previous planting should be
related unless mediated by food and labor supply or requirements. This argu-
ment is further illustrated in table 3.5, which compares plantings in the first
survey year with those in the following year. Of the rural sites surveyed, those
in valleys faced an extremely poor harvest in 1981 due to excessive rains, while
the plateau sites reported average to good harvests. It was found that area
planted for 1982 declined by nearly 20 percent for the valley households but
remained constant on the plateau. The decline was associated with a 10 percent
decrease in food crop area and nearly 40 percent decrease in cash crop area per
household in the valley.
An analysis of factors affecting cropped area shows that household labor
supply and cash income from the sales of the previous year's crop are the best
predictors of area planted by the household (Kumar 1984). Availability of a
plow or tractor for land preparation by itself does not increase area planted,

Table 3.4 Crop area for 1980-81 in relation to month in which previous food stores were
depleted, Chipata District, Zambia

Acres planted per household for 1980-81 crop
Percent of Food crop area Cash crop area
Chipata's 133 Per
Month food households Total capital Per Per
stores depleted in sample area area Total capital Total capital

October 13.5 4.04 0.81 2.82 0.58 1.08 0.20
November to 42.9 3.97 1.17 3.20 0.98 0.72 0.18
February to 37.6 7.88 1.23 4.70 0.95 3.10 0.26
After April 6.0 8.11 2.47 5.39 1.22 2.71 1.24

Source: Collaborative IPPRI study in Zambia.

Defining the Food and Nutrition Problem

possibly due to labor bottlenecks for weeding. An alternative explanation is that
mechanization may reduce labor supply by releasing surplus labor for out-
migration and through the substitution of other activities for agricultural labor
(Mellor and Ranade 1984).
There are several mechanisms by which the depletion of household food
stores can adversely affect production of next year's crop. First, seasonal em-
ployment opportunities during the dry season are severely limited by the virtual
absence of irrigation. If migration to urban areas takes place in response to crop
failure, it is unlikely to be of short duration if employment is found. Hence, the
household labor supply is reduced; this may or may not be accompanied by a
proportional reduction in demand for food and other items, depending on the
extent to which dependents migrate. Demand for farm labor and farm wages are
highest during the rainy season planting and weeding period. Hence, house-
holds who meet their food production deficit in any year with farm employment
can either work for extremely low wages during the dry season or sacrifice their
own planting and weeding time to earn supplementary income during the rainy
Finally, health and nutritional factors have implications for labor input. It
is evident that weight loss occurs for working members despite an increase in
caloric intake and is related to the magnitude of household labor input. A
weight loss of, for example, four to five kilograms over a short period of time,
such as one to two months, may not be inconsequential from the standpoint of

Table 3.5 Crops planted in two consecutive years, Chipata District, Zambia

Acres per household
Totala Plateaub Valley"
Crop 1980-81 1981-82 1980-81 :1981-82 1980-81 1981-82

Local maize 1.75 2.16 1.98 2.78 1.53 1.54
Hybrid maize 1.33 1.07 2.45 1.98 0.22 0.18
Groundnuts 0.84 0.72 1.28 1.20 0.40 0.25
Cotton 0.41 0.13 0.12 0.11 0.69 0.15
Sunflowers 0.21 0.33 0.22 0.29 0.20 0.38
Sorghum 0.52 0.50 0.12 .01 0.91 0.97
Total household 3.44 3.55 3.60 4.14 3.28 2.96
food area
Total cash area 2.05 1.56 2.98 2.42 1.14 0.72
Total area 5.49 5.11 6.58 6.56 4.42 3.68

Source: Collaborative IFPRI study in Zambia.
aN = 109.
bN = 54. Only two out of three sites are included here, since 1981-82 planting data from one
site had not been processed.
cN = 55.

The Nutrition Situation and Its Food Policy Links

worker productivity, since not only fat but also muscle mass is lost (Brooks et
al. 1979; Lipton 1983).
Although Eastern Province in Zambia is relatively well off, seasonal stress
is observed despite an increase in calories consumed. In most semiarid areas,
there is an actual decline in wet season consumption. Clearly, the influence of
decreased consumption on labor input and capacity will be more pronounced in
these areas.


Scattered information on other characteristics of the nutritional problem in
Africa has been reviewed elsewhere (Schofield 1979; Kumar 1981). Three main
sets of issues arise from this literature.
First, food-deficit areas and economically depressed areas are generally
the worst off, with rural areas faring relatively poorly. While variability in food
supply poses critical problems in the unimodal rainfall areas, large parts of the
bimodal rainfall belt have a lower annual level of food availability and diets
with less calories and protein because of excessive reliance on root crops (with
associated problems of dietary bulk and palatability).
Second, the nutritional status of households is related to household food
production and storage in rural areas and to cash incomes in urban areas. The
role of cash incomes in rural areas is not clear-cut and must be examined further.
This is important, as most efforts for technological change in smallholder
agriculture inevitably count on increasing commercialization of the rural econ-
omy. At present, much of the literature deals with the adverse consequences for
food consumption and nutrition when farming households turn increasingly to
producing crops for sale. However, inadequate attention has been paid to the
role of the increasing cash economy in changes within households that influ-
ence resource allocation and decisionmaking that impact on nutritional status.
Changes in the availability and prices of goods and services that may also be
taking place can affect household resource allocation, and this needs to be
considered as well. Finally, more needs to be known about the influence of
household assets, including land and access to technology and labor, on house-
hold production choices and on the consequent availability of food and cash
incomes. Analysis of these questions concerning changes in nutrition will
improve the capacity of agricultural development programs to better incorpo-
rate nutritional objectives.
Third, the importance for nutritional status of such individual charac-
teristics as age, sex, and work are least clear of all. Most studies cannot take the
vulnerability of all household members into consideration. It has generally
been assumed that those with the most critical biological needs-young chil-
dren, pregnant women, and lactating mothers-receive a lower proportion of

Defining the Food and Nutrition Problem

their nutritional requirements than working members. Thus longer-term human
biological growth potential is assumed to be sacrificed for shorter-term survival
concerns. If such patterns actually exist, the real costs and options for house-
holds and their policy implications are not clear.
While the general causes of nutritional problems are well known, the
significance of a wide range of different factors varies in different situations
(Kennedy and Pinstrup-Andersen 1983; Schofield 1979; Longhurst and Payne
1979). In most instances, however, we are interested only in policy measures-
such as pricing policies, technological change, and promotion of specific
crops-that may be expected to have a large impact on the food chain leading to
individual nutritional status. IFPRI, for example, is currently engaged in specific
studies of this nature in Zambia, Kenya, the Gambia, and Rwanda.
A careful examination of the nutritional problem in Africa is clearly
merited on welfare grounds. However, it also appears justified by the available
evidence, which indicates that, in many households, food availability and
nutritional status may prove to be an impediment to development programs
requiring more intensive labor input at certain times of the year. This is particu-
larly relevant, since labor, especially household labor, is a major constraint in
the extensive agricultural production systems in Africa. The significance of
seasonal fluctuation in food availability and its role in accentuating year-to-year
variability in production should also be considered. A fuller understanding of
these processes may provide the basis for more effectively removing production
constraints for smallholder agriculture in sub-Saharan Africa. Furthermore,
since nutritional and development objectives coincide, attempts should be
made to trace effects of various policy measures likely to produce shifts in the
levels and composition of food consumption and nutrition of different popula-
tion groups.

chapter 4

Commentary on Defining the Problem

Alexander P. Vamoer

Kumar's summary of the nutritional problems in sub-Saharan Africa indicates
that African countries have generally not perceived malnutrition as pervasive.
This problem should be viewed in its historical context. Before independence,
most eastern and southern African countries, with the probable exception of
Uganda, had very little data on the prevalence of malnutrition. Since then, a
number of surveys have established the magnitude of the problem. In countries
where such data is now available, it is concentrated largely in aggregates such
as national statistics on food production, supply, or consumption. It must be
disaggregated to provide information on specific risk groups in the population
likely to be affected by development policies and programs.
Since independence, many African countries have gone through rapid
changes in government administrative systems. National development plans
and priorities had to be formulated, with many interests competing for very
limited resources. Productive economic activities got a larger share than the
direct human services. Malnutrition was given low priority and was considered
a health problem to be solved within the context of health services. It has taken
governments several years to realize the complexity of nutritional problems and
to give it the attention it deserves.
Malnutrition and undernutrition is much more of a problem in rural than in
urban areas. This is partly due to the difficulty of providing rural infrastructure
and services such as roads, schools, and health services. However, with the
removal of food subsidies plus rural-urban migration and high unemployment,
the number of acute cases of malnutrition in cities is increasing. Malnutrition
and the unsanitary conditions in the shanty compounds has resulted in higher
mortality rates for children in cities, particularly those children under five years
of age.

Defining the Food and Nutrition Problem

Fluctuation of food supplies in most rural areas in most cases is due to the
failure of government policies to address rural poverty, particularly the needs of
peasant farmers. In most African countries, allocation of development re-
sources, including social facilities, are biased in favor of developing, sustain-
ing, and maintaining urban infrastructures. Low productivity of rural house-
holds has been compounded by lack of supporting services. For example, until
recently, extension services in Zambia existed only for commercial farmers,
though smallholders need them most. Lack of roads makes it difficult to trans-
port food from areas of surplus to areas of deficit and as a result hampers food
marketing. Lack of adequate storage facilities and the absence of village-level
technologies to process and preserve perishable foods further aggravate nutri-
tional levels of rural people.
Low productivity in rural areas is associated with the high prevalence of
mostly preventable diseases. The major problem contributing to low resistance
to infections is chronic food shortage. In some rural areas of Zambia, calorie
intake is as low as 1,300 per capital per day. Available data also show that acute
food shortages coincide with the planting season, when more energy is needed
for cultivation and weeding.
Development of irrigation, especially in the light of recent droughts in
eastern and southern Africa and declining food production trends, is urgent,
especially in countries of unimodal rainfall. In developing irrigation, produc-
tion of staple foods to alleviate hunger should be emphasized rather than foods
such as wheat, which only serve the urban elite.
Increases in rural food production also have been constrained by the high
cost of agricultural inputs. Facilities are needed to provide small-scale farmers,
who have little collateral, with some form of credit under terms and conditions
they can afford. Alternatively, some form of subsidies should be considered for
rural farmers, as mentioned by Norman in his foreword to this book.
In addition to the implications of high consumer subsidies for on-farm
storage of food grains mentioned by Kumar, two other factors have contributed
to the low level of on-farm retention of food grains. First, the annual pricing
system of agricultural products (in Zambia, for example) has adversely affected
small peasant farmers, who tend to sell most of their harvest as soon as producer
prices have been set by the government. Second, the tendency to sell most of
the food grain and buy back processed food items like maize meal is closely
linked to the time factor in rural women's work load. Village-owned hand
grinders or hammer mills would enable rural communities to mill their maize
locally and release women's time for other activities.
The general impression given by Paulino's chapter is that of declining
food supplies and increasing population. The chapter discusses the changes
from the early 1960s to 1980 and does an excellent job of warning the countries
concerned to take stock of their positions and to remedy the situation. However,

Commentary on Defining the Problem

it does not address the factors that have caused the decline in food production in
the face of increasing incomes, nor does it suggest solutions.
Although the eastern and southern Africa subregions are largely agri-
cultural, their farm population has dropped about 9 percent over the 1961-80
period. This is hardly surprising, since many of the white commercial farmers
who dominated the agricultural sector left following independence, accounting
in part for the decline in food production.
The reduction in cereal food crops in some eastern and southern African
countries can be partly attributed to policies that encourage certain crops that
are easily exportable, such as maize, at the expense of local staples, such as
millet and sorghum. Until this year, Zambia had no guaranteed support price
for sorghum, millet, or cassava, for example, and there was no incentive for
farmers to grow these crops.
Paulino's chapter mentions that growth in food production was due to the
expansion in harvest area. As it turns out, this is much more expensive than
increasing output per unit area, unless new methods are employed that permit
lower labor input per unit of land. In Zambia, the policy is for improvement in
output per unit area by increased use of inputs such as hybrid seeds and
fertilizer. The chapter also draws attention to inputs such as irrigation, fertil-
izer, and use of tractors, which have the potential to enhance growth in food
production. These inputs will be important in increasing output per unit area.
It is commonly assumed that malnutrition is primarily a supply problem
and that the solution is to grow more food. Increasing the supply of food
through improved agricultural prices, new technologies, or more efficient mar-
keting systems is probably a necessary condition for improving nutritional
standards in most countries. However, it is not enough. It is now widely
recognized that increased national food supplies alone will not eliminate mal-
nutrition and that any policy that does not insure an equitable distribution of
these supplies will be inadequate.
Many of the malnourished are small-scale rural farmers, and the majority
may be landless agricultural laborers or low-paid or unemployed rural nonfarm
workers. These groups benefit only indirectly from smallholder production
programs. Malnutrition has repeatedly been linked to low purchasing power of
the poor. Any rural development schemes that fail to include income-generat-
ing activities will fail to eliminate malnutrition of the needy poor. Infrastruc-
tural changes need to be geared specifically to their needs.
For these measures to succeed, it is necessary that the macroeconomic
factors be conducive to their implementation. Furthermore, it is imperative that
nutritional and food policies be integrated. In many countries, the former is
regarded as a health issue and the latter as an agricultural issue. It should be
borne in mind that the objective of policies for accelerated food production is
improvement in the nutritional status of our populations. It is therefore crucial

56 Defining the Food and Nutrition Problem

that nutritional considerations be systematically incorporated into agricultural
policies and programs at the planning level.
The challenge facing African countries in meeting food and nutritional
requirements is a serious one. It has internal and external components, and
quick action is urgently needed. Although public pronouncements of many
African governments give high priority to agricultural and rural development,
actual budgets have often been inconsistent with such pronouncements. Deter-
mined commitment is needed.

Part II

Potential and Practice
in Food Production
Technology Development

chapter 5

The West African Semiarid Tropics

Peter J. Matlon

The West African semiarid tropics (WASAT) are defined as those areas where
precipitation exceeds potential evapotranspiration from two to seven months
annually. This corresponds to mean annual rainfall limits of approximately 250
to 1,300 millimeters. The area encompasses all of Senegal, the Gambia, Bur-
kina Faso, and Cape Verde; major southern portions of Mauritania, Mali, and
Niger; and the northern portions of Ivory Coast, Ghana, Togo, Benin, and
Nigeria. Except for Senegal, all are classified by the World Bank (1981b) as
among the poorest third of the world's developing countries, with mean per
capital incomes of $320 or less. These countries are experiencing rapid popula-
tion growth and remain primarily rural.
Cereals occupy nearly 70 percent of total cultivated area in the WASAT, and
absorb 50 to 80 percent of total farm-level resources. Millet and sorghum
account for roughly four-fifths of cereals production, but yields are less than
half those for cereals in Africa outside of the WASAT. During the 1970s, cereal
output grew about 1 percent annually, due primarily to area expansion.
Compared to the whole of Africa, pressure on the land in the WASAT
appears to be low, with only 3 percent of the total land area under cultivation in
any given year. For continental Africa, this figure is 6 percent. Moreover, the
average density of agricultural population in the Sahel is 156 inhabitants per
square kilometer of cultivated land, compared to 183 for all of Africa (FAO
The amount of land potentially available for cultivation, however, is con-
siderably less than the above figures might suggest. Substantial portions are

The author wishes to thank Wilhem Stoop of the Royal Tropical Institute who reviewed and
commented on drafts of this chapter.

Potential and Practice in Food Production Technology Development

unsuitable for agriculture due to desert conditions, large expanses of rock
outcropping, and so forth. The distribution of the rural population is also not
always positively associated with agricultural potential.' Moreover, due to
microrelief, soil quality can vary widely within even small areas. Thus even in
zones of only moderate population density (30 to 50 persons per square kilo-
meter), the better soils often are already under cultivation, and the expansion of
cropped area means declining production potential with existing technologies.
Finally, soils in the WASAT tend to be unstable, with productivity rapidly
declining under continuous cultivation. The traditional bush-fallow system can
require a five to one or higher ratio of fallow to cultivated land to maintain soil
Within the WASAT, east-west isohyets delimit a set of agroclimatic zones
with distinct agricultural systems and development potential. Most generally,
one distinguishes three zones: the Sahel, the Sudan, and Guinea. The first two
form the primary focus of this chapter.
The Sahel includes all farmed areas of Cape Verde, Niger, and Mauritania,
one-third to one-half of the cultivated areas of Senegal, Mali, Burkina Faso,
and Chad, and smaller portions of northern Nigeria (Norman et al. 1981). This
zone receives 250 to 750 millimeters of rainfall and has a rainy season of 60 to
120 days. The shortness of the growing season, low and variable rainfall
(especially during planting and grain-filling periods), and low soil fertility are
the principal constraints to agricultural production.
The Sudan zone includes the Gambia, the southern portions of Senegal,
Mali, Burkina Faso, and Chad, most of northern Nigeria and northern Cam-
eroon, and portions of northern Ivory Coast, Benin, Togo, and Ghana. Rainfall
is less variable, ranging from 750 to 1,300 millimeters, and the growing season
of up to 150 days occurs from May to October. The Sudan zone has a wider
range of crops, higher yields, and a greater development potential than the


Several climatic characteristics other than low rainfall limit the region's agri-
cultural potential. The WASAT has a significantly shorter crop-growing season
than other semiarid-tropics (SAT) with similar rainfall. This limits the transfer
of certain technologies, such as preplanting plowing for example, from semi-
arid India (Oram 1977). Although high temperatures and solar radiation during
the rainy season are conducive to rapid plant growth, they also cause high
evaporation, which reduces available soil water. Unlike the Indian SAT, evap-

1. This is often due to tribal or colonial history (for example, the Mossi Plateau in Burkina
Faso) or to disease (for example, onchocerciasis in many of the major river valleys).

The West African Semiarid Tropics

orative demands are highest in May and September, the planting and grain-
filling periods, increasing the risk of early and late season water stress. Further-
more, rainfall intensities are two to four times greater than in temperate cli-
mates. The result is high risk of topsoil erosion and loss of up to 60 percent of
rainfall through runoff.
Even in years of "normal" total rainfall, distribution tends to be erratic,
with drought periods of two weeks or longer common, particularly in the Sahel.
Variation in annual totals is also high. The annual coefficient of variation of
rainfall is 20 to 30 percent in the Sudan and 30 to 50 percent in the Sahel.
Rainfall data also reflect the tendency for abnormal years of rainfall to occur in
more or less uninterrupted periods for as many as 15 years (Nicholson 1982).
Consequently, systems of production, storage, and exchange cannot be based
exclusively on expectations of an average climatic situation. Rather, the food
system must be sufficiently flexible to reduce the welfare impact of a series of
bad years, while permitting reasonable exploitation of resources during good
Charreau (1977) has defined two broad soil groups in the WASAT, which
correspond roughly to the rainfall belts described above. Between the 200- and
500-millimeter isohyets, modal brown and reddish-brown soils (camborthids)
predominate. Further south, between the 500- to 900-millimeter isohyets, the
most common soils are red to grey, ferruginous, leached soils (ustalfs). High-
potential alluvial soils and black clay soils (vertisols), which occupy large areas
of the Asian SAT, are far less common in the WASAT and tend to occuronly in the
more southern belts in isolated patches. Soil texture varies from loamy sands in
the northern Sahel to sandy loams in the southern Sudan areas. Except for the
limited vertisol pockets, clay content is uniformly low, less than 20 percent, and
the soils are structurally inert and have poor water-holding capacity. Compared
to the red soils of the Indian SAT, the clay content of typical WASAT red soils is
approximately one-half (Stoop et al. 1981). Soil depth is generally shallow,
thus further limiting the soil's water storage, except in the extreme north, where
eolian deposits result in deep sands merging into dune conditions.
Both soil groups have low to very low natural fertility. Due to low clay and
organic matter content (generally less than 1 percent), cation exchange capaci-
ties tend to be less than five milliequivalents per 100 grams of soil. As a result,
the soils are highly fragile. Although nitrogen and phosphorus are commonly
the most limiting nutrients, other deficiencies (potassium, trace elements) and
acidification are readily induced by intensified continuous cropping (Pichot et
al. 1981).
In addition to low natural fertility, the major physical properties of WASAT
soils that limit crop production potential include: (1) very low structural poros-
ity and consequently high bulk density, which reduces root penetration and
water circulation; (2) a tendency for compacting and hardening during the dry
season, which results in high early season runoff and which severely restricts

Potential and Practice in Food Production Technology Development

preseason and postseason cultivation; (3) generally poor infiltration, except on
eolian sandy soils, due to rapid surface crusting of soils even after cultivation;
(4) low values of available water compared to typical Asian SAT soils (Virmani
et al. 1978); and (5) increasing susceptibility to erosion with continuous
cultivation. Combined with the climatic conditions described above, these
characteristics result in a general fragility, which can cause rapid deterioration
of soil productivity under some forms of intensification.
Within the broad soil groups, microvariations importantly determine pro-
duction potentials of particular fields. Farmers have adapted to these variations
with highly flexible management practices. For example, millet is mainly
grown on the shallow, gravelly soils generally located on the plateau and upland
portions of the catena. These soils tend to be drought, rapidly exhausted, and
subject to high risk of erosion. Deeper soils (sandy or silt loams in the Sudanese
zone) often occur in depressions and become more plentiful toward the mid-
slope and lowlands. Less drought-tolerant crops, including sorghum, maize,
and rice, are sown in these areas. One result of microvariability in soils is the
highly fragmented patterns of traditional cropping systems and the often un-
even adoption of components of improved technology.


Cereal cropping patterns in the WASAT gradually change from short-cycle,
drought-tolerant varieties in the Sahel to longer-cycle, more drought-sensitive
crops in the Sudan. The major crops and their areas of concentration are as
Millet is grown throughout the WASAT, from the Sahel to the southern
Guinea savanna, but is the dominant cereal between the isohyets of 250 to 650
millimeters. Sorghum production is generally concentrated in areas of 650-1
thousand millimeters, where it is the principal food grain. In traditional systems
in the Sahel and the central Sudan, maize cultivation is limited to heavily
fertilized soils adjacent to compounds and on the margin of swamps, where it
rarely exceeds 3 percent of total cereal area. Only in the southern Sudan and in
northern Guinea does maize replace sorghum as a major field crop. Rice is a
minor crop throughout the WASAT, with cultivation limited to unimproved or
moderately improved swampland. Cowpeas, the major grain legume in the
WASAT, is usually grown as an intercrop at low density (1 thousand to 8
thousand plants per hectare) with millet or sorghum.
The regional importance of export crops is influenced more than food
staples by institutional factors (credit, input, and marketing services) and by
price factors, both of which vary widely between countries. Although
groundnuts remains a major export crop in Senegal, diseases and unfavorable

The West African Semiarid Tropics

prices for groundnuts have substantially reduced planting since the early 1970s,
especially in northern Nigeria. Cotton is a major cash crop in large areas of the
Sudan and the northern Guinea savanna. Particularly within the francophone
countries, cotton production under relatively high management conditions has
been successfully promoted by national parastatals.
Farm units throughout the WASAT tend to diversify the crop mix to reduce
aggregate production variability, to produce most of the foods required in local
diets, to satisfy part of family cash needs, and to exploit soil microvariability
(Norman 1974; Abalu 1976). The range of mixtures is greatest in the southern
zones (Kabore et al. 1983). Despite the traditional importance of intercropping,
agricultural research has only recently begun to explore improved intercropping


Most of the agricultural production in the WASAT is from small, family-based
farms. The operators range from relatively small nuclear families to large,
extended or compound families. Unlike in the Asian SAT, a landless class of
workers has not yet emerged. The cultivated area per household generally
averages between two and six hectares. The distribution of land among house-
holds tends to be highly equitable, with some exceptions in areas of high
population density (Matlon 1981). Nevertheless, there is evidence that conven-
tional forms of usufruct are increasingly yielding to rights of possession and
transmission of ownership through inheritance, and that land control is becom-
ing more concentrated.
Labor input per unit area, primarily from family sources, ranges from 350
to 1,200 hours per hectare, varying as a function of crop, population density,
and length of cropping season. Periods of peak labor input generally corre-
spond with planting and weeding, when timeliness can critically affect potential
yields (Matlon and Newman 1978).
In traditional production systems, capital consists primarily of hand tools,
seed, and small quantities of fertilizer. Nonlabor cash expenditures in manual
systems without chemical inputs are low, generally less than $5 per hectare. In
areas where hired labor is common, however, cash expenditures can be as high
as $20 to $60 per hectare (Matlon 1977). In spite of efforts to extend animal
traction since the early part of the century, probably less than 10 percent of
farmers in the WASAT employ animals as a power source.
Nearly all farm households maintain fowl and small ruminants and, in
some cases, cattle as easily liquidated forms of savings, insurance, and income.
However, potential technical complementarities between livestock raising and

Potential and Practice in Food Production Technology Development

cropping are rarely achieved. Under the traditional, extensive land-use sys-
tems, livestock forage away from the farm, with the result that most manure is
lost. Systems of entrustment, whereby Fulani herdsmen manage the livestock
of cultivators, are well adapted to this ecological setting (Delgado 1977). With
mounting population pressure, however, two opposing forces point to the need
for profound structural changes in current crop-livestock systems. With a shift
to less frequent fallows, low organic matter levels are becoming a serious
constraint to current production and to long-term soil stability. Substantially
larger amounts of manure are needed to stabilize these systems than are cur-
rently being applied. And second, growing land pressure is concurrently reduc-
ing the area of wild forages and, in turn, livestock-carrying capacity.
It is generally held that WASAT farmers are allocatively efficient (which
means insignificant production gains would result from a reallocation of avail-
able resources given present technology) and responsive to economic incen-
tives. Although conventional wisdom states that farmers are averse to risk and
generally conservative with respect to innovation, in fact little rigorous em-
pirical research on farmers' risk attitudes and behavior has been done in the
WASAT (Norman et al. 1981). Moreover, the wide-spread adoption of ground-
nuts and cotton and the active, "traditional" experimentation with new cereal
varieties (ICRISAT 1983) demonstrate the openness of farmers to change when
adequate incentives and means are available.
WASAT farming households typically pursue other income-earning ac-
tivities, such as crafts, salaried labor, and crop and noncrop trading. In the
highly monetized rural economies of northern Nigeria, for example, these
activities can contribute up to one-third of total farm household income and
more than one-half of net cash revenues (Matlon 1981). The highly variable
returns to labor in nonagricultural employment are location specific and trade
specific. In northern Nigeria, average returns in more than 20 nonagricultural
occupations are roughly 50 percent greater than the farm wage rate, which is
approximately equal to the marginal value product of farm labor. Returns to
labor in the most capital-intensive occupations are generally two to three times
the farm wage rate (Matlon 1977). Migration to urban areas and to coastal
countries for off-season or multiyear employment is also important and poses
an attractive alternative to farming to the younger generation.
Support services to small farmers are generally poorly developed through-
out the WASAT. Understaffing, multiple responsibilities for agents, lack of
transportation, inadequate supplies of inputs, and insufficient training charac-
terize the national extension services of most WASAT countries. Ratios of farm-
ers to extension agents are as high as 3 thousand to 1 (USDA 1981). Two
important exceptions to these more general patterns are parastatals, which are
responsible for extension services for cash crops, for input supply, and for
marketing programs; and externally financed, integrated, agricultural develop-
ment projects, as in northern Nigeria.

The West African Semiarid Tropics


Fundamental structural changes are now occurring in WASAT farming systems.
With population growth, cultivation is being systematically expanded to fields
more distant from habitation and to marginal soils, which were formerly
avoided due to lower fertility or greater drought proneness. The traditional
bush-fallow method of maintaining soil quality is being abandoned, with a
consequent decline in fertility, permanent elimination of vegetative cover, and
increasing erosion. Alternative methods of maintaining the topsoil and the
fertility of marginal soils under intensive cultivation are required. Similarly,
local cereal varieties originally selected for cultivation on the most favorable
soil types are increasingly less well adapted to the less fertile, more shallow,
drought-prone soils now being brought under cultivation.
Increased economic opportunities both within and outside agriculture have
resulted in a trend away from large family units and toward more individual
decisionmaking. Simultaneous changes in land tenure systems are resulting in
greater personal control over landholdings-but also fragmentation of hold-
ings. While these changes increase incentives to produce for personal gain and
to invest in land improvements, they also reduce the effectiveness of tradi-
tional, communal support mechanisms, which protected families against se-
vere welfare loss in case of crop failure. The dissolution of large units also
reduces the feasibility of new technologies, which require large investments or
group management to achieve economies of scale-for example, watershed-
based land management and, perhaps, animal traction.
Finally, farmers are increasingly weighing investments of capital and
labor in agriculture against opportunity costs in the nonfarm sector. On one
hand, the growing importance of nonagricultural employment offers viable
alternatives to farming and reduces incentives to develop stable, intensive
cultivation systems for the long term. On the other hand, more capital from
nonfarm employment is becoming available for reinvestment in farming. The
issue is whether currently available technologies present sufficiently profitable
agricultural investment opportunities.


The major areas of possible technological change are irrigation, other forms of
land and water management, mechanization, genetic improvement, and the use
of chemical inputs. We consider these in turn.

Areas under irrigation in eight Sahelian countries have increased at an annual
rate of 3 to 5 percent during the last two decades. Despite this growth, irrigated

Potential and Practice in Food Production Technology Development

areas currently account for only 3 percent of total farmed areas. Moreover, due
to the lack of complete water control and general absence of double cropping,
yields in both modern and traditional perimeters are low and highly variable.
Annual yields of rice, the main crop under irrigation, range between one and
three tons per hectare, for example-substantially below the potential of up to
eight tons achieved under double cropping in parts of Niger (CILSs 1980).
The technical potential to expand irrigation substantially surpasses current
levels. A study by the Food and Agriculture Organization reports that the Sahel
has approximately 14 million hectares of irrigable soils, or roughly 20 percent
of total arable land (FAO 1975). However, because seasonal flow patterns in the
Sahelian river systems are extremely variable, substantial investment in dam
facilities would be required for more complete water control and more intensive
double cropping. And because of the generally flat topography of the region,
opportunities for large dam construction are rare and costly. Similarly, although
groundwater reserves are estimated to be substantial, it is generally conceded
that costs of exploiting them rule out their use on a major scale for crop
production, given foreseeable input and output price relationships (CILSs 1980).
In what appears to be an overly optimistic assessment, which assumes full
exploitation of surface water, CILSS (1980) projects the irrigated agricultural
potential of the Sahel sites that could be developed within 25 years to be 2.3
million hectares, approximately 10 times existing levels. A USDA report (1981)
projects a similar technical potential based primarily on large-scale projects.2
The economic potential of large-scale irrigation projects is less encourag-
ing. First, per hectare investment costs are substantial, varying between $5
thousand and $20 thousand (CILSS 1980; World Bank 1981a). Second, yields
are not sufficiently superior to rainfed agriculture, due to poor water control,
absence of double cropping, inappropriate agronomic packages, and lack of
complementary inputs. Third, poor management and high recurrent costs
result in inadequate maintenance of equipment and structures. Fourth, little
consideration has been given to farmers' incentives and to the problems of
incorporating traditional farmers into intensive and externally directed irri-
gation systems. Fifth, external costs, including loss of grazing and farming
land, fishing opportunities, and movements of displaced population, can be
In a review of the limited available literature on the economics of irrigation
in the Sahel, Sparling (1981) concludes the small-perimeter, labor-intensive
irrigation projects tend to be substantially more profitable, privately and so-
cially, than large-scale projects. The potential area affected by such projects,
however, is extremely small.

2. The USDA study identifies at least 11 projects in the Sahel that could increase total irrigated
area to over 1.9 million hectares.

The West African Semiarid Tropics

Land and Water Management3
The most common method of maintaining adequate soil moisture and achieving
soil conservation in traditional systems has been to alternate cropping with long
bush-fallow cycles. Due to increasing population pressure, however, these
systems are yielding to shorter grass fallows and in some areas to continuous
cultivation. Alternative methods available or under research in the WASAT are
soil tillage, including various methods of soil preparation and ridging; mulch-
ing; contour bunding; and watershed-based management. The scale of these
techniques ranges from plot and field to watershed. Given the fragmentation of
field ownership, larger-scale approaches, while technically more comprehen-
sive, raise important ownership and group management problems.
Construction of ridges that are subsequently tied (at planting, at first
weeding, or at flowering) has significantly increased water infiltration and
enhanced yields under research station conditions in East Africa (Ruthenberg
1980). Experimental results in West Africa also indicate considerable technical
potential for tied ridging. Yield response is greatest under conditions of soil
moisture stress and where soil fertility is not limiting. Average yield increments
of 1 thousand kilograms per hectare for maize (Rodriguez 1982), 930 kilograms
per hectare for sorghum, and 570 kilograms per hectare for millet have been
observed on research stations in Burkina Faso, where medium to high doses of
NPK fertilizer have been applied (ICRISAT 1982).
Despite this technical potential, tied ridging is not practiced by WASAT
farmers. Major questions concerning this technique are labor costs, the poten-
tial response under low-fertility soils typical of farmers' conditions, and the
yield gap between the research station and farmers' fields, even with fertilizer
complements. Limited on-farm tests in Burkina Faso suggest that yields under
farmers' conditions are 60 to 90 percent below station results at high-fertility
and low-fertility levels, respectively (Rodriguez 1982; FSU 1982). Returns to
incremental labor at these levels are noncompetitive with labor use in flatbed
Mulching has also received considerable research attention. The potential
advantages of applying crop residues or free-cut straw as a soil cover are to
increase infiltration, reduce erosion, control weeds, improve structure, and
reduce soil temperature. Experimental results of mulching trials have been
contradictory, however, probably because of variation in soil types, topogra-
phy, and seasonal rainfall patterns.4
In traditional farming systems, mulch is generally applied by WASAT
farmers in limited quantities on small areas and in special circumstances, such

3. This section draws heavily on Charreau (1977).
4. Charreau and Nicou (1971) in a general review of research results, for example, found no
clear superiority of yields under mulching, whereas Perrier observed yield increases of 200 percent
for both local and improved sorghum varieties in central Burkina Faso (ICRISAT 1982).

Potential and Practice in Food Production Technology Development

as on termite mounds or particularly shallow portions of fields where water
infiltration or retention is limited. A major constraint to expanded use is that
increasing amounts of available straw are being fed to livestock or used for
construction, for mat making, or for fuel. The rising demand for straw as fuel
and forage, in particular, suggests limited possibilities for expanding mulching
on a significant scale.
The traditional use of bunds to control erosion and improve water infiltra-
tion, while observed in some areas of the WASAT, is rare and is practiced on a
relatively small scale. In zones of high population pressure, for example,
farmers occasionally construct small rock dikes across water courses to reduce
runoff velocity and prevent gully erosion. Similarly, low dirt dikes are some-
times built around small, heavily manured household plots to avoid loss of
fertilizer. Although early large-scale projects in the WASAT that employed
unstable dirt contour bunds as a principal intervention were not successful,5
evaluations of recent large-scale dirt-based bunding projects in Burkina Faso
suggest considerable potential. In addition to the long-term benefits of reduced
topsoil loss, recent farmers' tests in Burkina Faso have measured highly signifi-
cant 20 to 80 percent yield increases in the short term (ICRISAT 1983). Visible
yield increments on these magnitudes are probably essential to motivate farm-
ers to maintain the fragile dirt bunds.
More stable, rock-based, water-harvesting bund systems have also been
developed and extended on a relatively small scale in the most densely popu-
lated and environmentally degraded portions of the Mossi Plateau in Burkina
Faso (Wright 1983). Although the potential of this method for increasing yields
on currently cultivated fields has not yet been determined, it has been successful
in bringing highly eroded, abandoned fields back into production and is proba-
bly one of the most promising technologies now available.
Comprehensive development of entire elementary watersheds, using alter-
nating broad beds and furrows to reduce soil erosion and to increase the efficient
use of watershed precipitation, has been successfully tested by ICRISAT in the
Indian SAT. Although no research has been undertaken to determine the poten-
tial for transferring this technology to West Africa, differences in the physical
characteristics of the two areas suggest that the potential is considerably more
limited in the WASAT and that direct transfer of existing methods probably is not
possible (Charreau 1977). The area of elementary watersheds in India is gener-
ally between 5 and 15 hectares, while that in the WASAT tend to exceed 100
hectares. As a result, investment requirements not only would be substantially
larger in Africa, but a larger number of households would be involved, entail-

5. The bunds constructed in the Yatenga region of Burkina Faso during 1960-62-perhaps
the largest of these early projects-have generally been judged a complete failure. Project planners
bypassed the farmers, who, observing no immediate yield increase, failed to maintain the bunds.
As a result, potential longer-term benefits in soil quality improvement were lost as the bunds

The West African Semiarid Tropics

ing important management and tenure problems. The general flatness of the
WASAT terrain also means that water storage facilities would be less efficient
and would cover substantially larger areas of what are often the best and most
intensively cultivated soils. Finally, economic analyses of the Indian systems
have shown them to be profitable only on deep vertisols, which are rare in the
WASAT, but not on medium or shallow vertisols or on alfisols, which are the
most common soils in the WASAT (Binswanger et al. 1980).


It is currently estimated that little more than 15 percent of cropped area in the
WASAT is cultivated by animal traction. Nevertheless, efforts to introduce draft-
animal technology to WASAT farmers, under way since the early 1900s, are
increasing. Since 1970, more than 50 projects involving a draft animal have
been funded in francophone West Africa, alone (Sargent et al. 1981). Large
investments in tractorization have also been made in several WASAT countries,
particularly during the 1950s and 1960s, but with little success.
In the face of these continuing investments, assessments of the technical
and economic impacts of mechanization in the region show mixed results. In
contrast to major yield and labor-saving effects observed on research stations
(Charreau and Nicou 1971; Kline et al. 1969; ICRISAT 1978), the limited avail-
able farm-level studies show that yield effects are generally insignificant and
that area effects range between only 10 and 40 percent in the Gambia (Mettrick
1978), southeastern Mali (Whitney 1981), and Burkina Faso (Barrett et al.
1982; McIntire 1981).
A recent review of more than 100 animal traction projects in francophone
Africa reveals common institutional and technical constraints, which limit
fuller attainment of the potential benefits and thus block more rapid adoption
(Sargent et al. 1981). First, adoption of traction systems tends to be lowest in
areas where cash cropping is minor. It is likely, however, that this is better
explained by institutional rather than technical factors. In areas such as the Sine
Saloum in Senegal groundnutss) and in southern Mali and southwestern Bur-
kina Faso (cotton), where mechanization is well advanced, vertically integrated
marketing institutions for cash crops provide short-term credit (for such subsi-
dized complementary inputs as fertilizer), medium-term credit (for equipment
and animals, intensive extension service, and veterinary support), and assured
markets. Where such support systems are lacking, traction adoption rates
remain insignificant.
Second, the range of mechanized operations performed by farmers is
generally less than that required to achieve full benefits from an integrated
traction system. Land preparation equipment-scarifiers and shallow plows in
the sandy soils of the northern belt and deeper plows for the heavier soils further
south-are often the only cultivation equipment adopted. Fewer than 25 per-

Potential and Practice in Food Production Technology Development

cent of Sahelian farmers with traction equipment weed mechanically. This is a
major constraint to area expansion in zones where plowing has been mecha-
nized. As a result of the limited range of operations performed, animals tend to
be grossly underutilized unless transportation is also mechanized.6
Third, the short rainy season tends to create severe labor conflicts between
plowing and timely planting. This helps explain the lack of an adequate rental
market for plowing and first-weeding equipment in all but the most southern
zones, where the cropping and preparatory rainfall periods are longer. With
rental equipment unavailable, each farm unit must amortize fixed costs for
animals and equipment over its own cultivated area. Similarly, farmers rarely
plow at the end of the season due to conflict with harvesting activities and due to
the rapid drying and hardening of the topsoil after the rains end. As a result, full
incorporation of crop residues is not achieved.
Finally, several studies suggest that, due to an extended learning period
both for farmers and their traction animals, at least 6 to 8 years are required to
achieve full farm-level benefits (Barrett et al. 1982; Jaegar 1984). During this
period, credit for purchase of equipment and animals must nevertheless be
repaid. Consequently, net incremental benefits tend to be negative during the
first several years, creating a serious cash-flow problem for recent adopters. In
addition, the risk to farmers of animal loss from sickness or death and of
production shortfalls from climatic variability add a powerful disincentive.
These results, placed within the context of the region's demographic and
ecological trends discussed earlier, have important implications for future ani-
mal traction programs and research. It must be recognized that the extension of
area alone through animal mechanization is not a viable long-run option and
must give way to intensive systems. Because farmers adopt land-extensive
systems in part as a means of risk reduction, new land-saving approaches using
animal traction as a key element must provide not only short-term profitability
but reduced production variability and long-run land conservation benefits as
Thus complementary biological components should be considered part of
an animal traction package. Cereal varieties more responsive than local vari-
eties to plowing and that permit later planting-thereby removing the timing
conflict with plowing-would increase returns to traction operations while
simultaneously reducing weather-related risks. Similarly, greater integration of
crop and livestock activities would allow farmers to more effectively manage
and recycle biomass production for the maintenance of soil structure and long-
term fertility. Since increased cash crop production may be necessary to assure
the financial viability of animal traction, agronomic research should also focus
on development of cropping systems incorporating cash crops into cereal-based

6. Barrett et al. (1982), for example, observe that in eastern Burkina Faso 45 percent of
farmers with traction equipment used their animals for less than 50 hours per year in field work.

The West African Semiarid Tropics

systems. Finally, efficient support services are needed, including farmer and
animal training, credit, input and output marketing, veterinary care, timely
equipment repair, and credit programs.

Crop Improvement
Breeding and varietal selection programs aimed at improving productivity of
food grains have existed in the WASAT for several decades. The greatest ad-
vances have occurred in upland rice and maize, where a number of varieties
developed by Institut de Recherches Agronomiques Tropicales and Interna-
tional Institute of Tropical Agriculture have met with some success. However,
both are relatively minor crops in the semi-arid areas. Far less success has been
achieved for sorghum and millet. Despite frequently encouraging on-station
results, yield gaps of 40 to 60 percent are consistently observed when improved
materials are tried on farms. Also, new varieties are often found to have
undesirable consumption or storage qualities. The result is that, within the
entire region, only insignificant areas are now sown to improved sorghum and
millet varieties.
Experience suggests that the relatively slow progress in sorghum and
millet improvement is due to several factors: difficulty in adapting Asian vari-
eties to African conditions; inappropriate crop improvement objectives in view
of the region's soils, infrastructural development, and farm-level capital; over-
reliance on varieties and hybrids grown solely under research station condi-
tions; and lack of feedback from the farm to the research station.
It has been observed that Indian millet is highly susceptible to African
races of mildew, smut, and ergot. Moreover, physiological factors tend to
accelerate the growth of the Indian varieties, causing them to be spindly and
partially sterile (Scheuring 1980). Similarly, the high-yielding sorghum hybrid
CSH-5, which had substantial success in India, experienced unacceptable prob-
lems of charcoal rot and lodging in station trials and farmers' tests in several
West African countries.
With respect to objectives, priority in crop improvement programs in the
WASAT has traditionally been given to identifying cultivars that yield well under
high-input management. Although this general approach achieved substantial
production gains for wheat and rice in South Asia during the 1960s and 1970s,
critical differences in conditions have blocked similar progress in the WASAT.
First, high-yielding varieties generally require increased plant density and the
use of chemical fertilizers to obtain production potential. However, technical
response rates to plant population and fertilizer are substantially lower and the
risk is higher on soils with low water-holding capacity or when water control is
absent, as is usually the case in the WASAT. Even in India, rates of adoption of
high-yielding sorghum varieties and use of chemical fertilizer tend to be great-
est in areas of more assured rainfall or greater irrigation density (Jha et al.

Potential and Practice in Food Production Technology Development

1981). Second, extension support and the infrastructure for supply of improved
biochemical inputs is considerably less well developed in the WASAT than in
most Asian countries. Third, since land pressure is substantially lower in the
WASAT than in the Asian SAT, there is less immediate economic incentive to
intensify land use through the use of complementary cash inputs.
In short, the risk to farmers of adopting varieties that require good soil
moisture, high soil fertility, thorough soil preparation, and other aspects of
good management in order to outyield local varieties, may be unacceptable.
Such varieties are unlikely to be adopted on a wide scale in the near future,
though some may be suited to intensive development projects or to limited
At issue is the priority given to management-dependent high yields, to the
practical exclusion of other possible breeding goals. Moderate yield increases
and substantially greater stability could be achieved through breeding for re-
sistance to the most common pests and diseases, for greater resistance to
drought, and for improved seedling vigor. Development of varieties with a
wider range of agronomic characteristics, such as reduced crop cycle or modi-
fied plant structure, could also increase farmers' management options by open-
ing new intercrop or relay cropping possibilities, by permitting late planting
without yield loss, and by permitting cropping on the most drought-prone soils,
where moisture limitations reduce the growing period. It must be recognized,
however, that although such a strategy would achieve greater production sta-
bility it would likely have marginal impact in terms of aggregate production.
For the longer term, development of stable and more input-responsive
varieties or hybrids is clearly necessary for major breakthroughs in cereal
production. The shift will take time, however, and require major investments in
both agronomic research and improved input supply infrastructure.
It must be emphasized that achieving success in crop improvement
urgently requires greater interdisciplinary research at the farmers' level, in-
volving specialists in physiology, plant protection, agronomy, food science,
economics, and breeding. Greater a priori understanding is needed of the
physical and social environment into which new varieties are being fit (Stoop
et al. 1981; Oram 1977). Factors causing the yield gap between the research
station and the farmers' fields need to be identified and on-station objectives
and methods modified accordingly. Constraints and points of flexibility in
current production systems need to be identified so that management practices
can be adapted to new materials. Varietal selection should take into account
the desirable characteristics for storage, processing, and consumption, as well
as for production. This approach requires greater work with farmers at several
stages of the breeding effort, rather than at the final stage of preextension
screening as is conventionally done. A continuing, interactive relationship
with farmers to define appropriate breeding objectives and to test concepts and

The West African Semiarid Tropics

materials should reduce the time necessary to develop well-adapted, improved

Chemical Fertilizer

Despite an annual rate of growth of approximately 15 percent since the mid
1960s, use of chemical fertilizer in the WASAT remains below any other area in
the developing world. Excluding Senegal, an average of less than one kilogram
per hectare of NPK is applied to food crops in the Sahel,7 and only about 10
percent of total cultivated area receives any chemical fertilizer. Major factors
explaining low use rates are (1) costs of foreign exchange; (2) high transport
costs to and within land-locked countries;8 (3) low and variable response rates
to local cereal varieties, particularly in areas with less than 700 millimeters of
annual rainfall; (4) poorly developed extension and distribution systems; and
(5) inadequate farm-level liquidity.
In several countries, parastatal agencies responsible for the production and
marketing of cash crops have provided high-quality seed of more responsive
cash-crop varieties and subsidized fertilizer on credit to small producers. As a
result, rates of application are substantially higher on such crops as groundnuts
and cotton than on cereals.9 An additional result of the cash crop emphasis is
that often the only fertilizer available for cereals are formulas developed for the
cash crops. In Burkina Faso, for example, extension service recommendations
for sorghum and millet are based on a complex fertilizer for cotton, despite
evidence that this formula can actually reduce cereal yields after several years
of continuous application (Pichot et al. 1981).10 In fact, little research has been
done to determine optimal formula and doses for different cereals by soil type
and agroclimatic zone, or to determine the long-term soil effects of continuous
fertilizer use.
Economic analyses of the response to available chemical fertilizers in the
WASAT are limited in number and often biased by on-station conditions that
positively interact with fertilizer (deep plowing, complete weed control, high

7. This compares with 1.4 for West Africa as a whole; 3.6 for East Africa; 29 for North Africa
and the Middle East; 23 for Asia; and 33 for South America. These figures are for 1975 and are from
IFDC (1977) and Oram (1981); also see chapter 2 in this volume.
8. Representative costs for Burkina Faso in 1982 are 39 CFA per kilometer ton by truck and 15
CFA by rail (Bonnal 1983).
9. It is estimated that in 1978 approximately 30 kilograms of fertilizer per hectare was applied
to groundnuts in Senegal; and in 1981 nearly 100 kilograms per hectare was applied to cotton in
Burkina Faso (Societe Africaine d'Edition 1983; SOFITEX 1982).
10. In trials conducted over 18 years in Burkina Faso, IRAT observed that sorghum yields
steadily declined following seven years of chemical fertilizer application due to soil acidification,
potassium deficiencies, and aluminum toxicity. Only large applications of animal manure with
chemical fertilizer were found to counteract this effect (Pichot et al. 1981).

Potential and Practice in Food Production Technology Development

applications of organic matter, and so forth). Nevertheless, these studies dem-
onstrate the technical response and maximum financial returns of recom-
mended doses across a range of crops. Research during 1978-82 at two stations
in Burkina Faso, for example, (IRAT 1983), indicate that the yield increment per
kilogram of NPK nutrient was highest for maize (a ratio of 13.5 to 1), followed
by sorghum (10.3), soybeans (8.6), cotton (4.3), millet (3.1), and groundnuts
(2.3). At nominal prices and using only the direct costs of fertilizer, the ranking
of crops according to financial returns was almost identical. Maize and
sorghum were highest, at rates of return of 450 and 330 percent. Millet was
lowest, with a return of only 37 percent, well below the FAO rule-of-thumb of
100 percent returns necessary for adoption under small-farm conditions. After
eliminating the effect of a 49 percent fertilizer subsidy, however, only maize,
sorghum, and soybeans showed positive returns under experiment station con-
ditions (Bonnal 1983).
Responses are substantially lower for farmers' demonstrations, with finan-
cial returns one-half to two-thirds those on experiment stations. In five years of
FAO farmers' demonstrations in Burkina Faso, for example, financial returns
(with a greater than 40 percent subsidy in effect) to recommended fertilizer
levels applied to sorghum varied between only 70 and 150 percent (FAO 1982a).
Response gaps are even wider when fertilized crops are managed entirely
by farmers. Two-year farmers' tests of compound fertilizer (14:23:15) con-
ducted in three agroclimatic zones of Burkina Faso, for example, concluded
that, at the economic cost of fertilizer and when applied to local sorghum and
millet varieties, returns justifying adoption occur only at low doses-50 per-
cent of currently recommended rates-and only in the southern and central
Sudan. Average negative returns were observed both years in the Sahel. Farm-
ers' tests also showed the high risks of fertilizer use in arid conditions. During a
year of below-average rainfall, incremental yields for local sorghum did not
cover even the subsidized cost of fertilizer when applied at recommended rates
on approximately 40 and 70 percent of farmers' fields in the southern and
central Sudan, respectively. It is important to note, however, that fertilizer was
both more profitable and less risky for the improved varieties included in these
The production and distribution of rock phosphorus from large local de-
posits in several WASAT countries is also receiving increased attention. Al-
though trials confirm residual yield effects of a basal dose of granulated rock
phosphate, it is a generally less economical source of phosphorus than imported
soluble phosphates (Bonnal 1983). Additional constraints at the farm level are
the difficulties in applying and incorporating finely granulated phosphates and
the multiyear delay in realizing the full yield benefits. Recent results with
partially acidulated forms of rock phosphate show promise in overcoming some
of these problems.
The above evidence suggests that increased use of chemical fertilizers now

The West African Semiarid Tropics

available to farmers in the WASAT does not offer an economically viable tech-
nology for sustained agricultural growth, especially in the drier areas. Low
response of local varieties, high risk, and negative long-term impact on soil
quality, combined with infrastructural and foreign exchange costs, underlie this
Several key issues must be addressed by both researchers and pol-
icymakers if chemical fertilizer is to play a more important role. First, specific
cereal-based fertilizer formulas need to be developed, which provide the nu-
trients required for different cereals under various soil and climatic conditions.
Second, greater applied research is required at the farm level to better determine
optimum dose levels, taking risk as well as profitability into consideration.
Third, major investments are required in production and distribution of fertil-
izer and in the development of complementary inputs-particularly improved
varieties and animal traction-if the potential of fertilizer is to be exploited.
Fourth, greater basic and applied research must be directed to managing the
long-term effects of fertilizer use on WASAT soils. Preliminary evidence indi-
cates that combining livestock raising with crop growing, which recycles
biomass through animal manure, may be essential to sustained chemical fertil-
izer use. Fifth, current price and subsidy policies must be reconsidered in view
of the economics of available fertilizers. Fertilizer subsidies at present levels
would appear to be justified only at implicit cereal prices well above domestic
market and import price levels.


Farming systems in the WASAT reflect a long process of adaptation to low and
variable rainfall, generally poor and fragile soils, and readily available land.
The extensive land-use systems, which have evolved under these conditions,
are marked by low productivity per unit area and high yield variability. Soil
quality has traditionally been maintained by long bush-fallow rotations, requir-
ing at least a five to one ratio of fallow to cultivated land. In rapidly expanding
areas of the WASAT, however, growing populations are upsetting this ecological
balance by cultivating more marginal soils and by continuous cultivation.
Increased cash needs are also inducing farmers in some areas to put greater
resources into cash crop production, often employing technologies that acceler-
ate a decline in soil quality. The immediate result is nearly stagnant growth in
yields of food crops and a general fall in aggregate farm output per rural
habitant. A more pervasive long-term effect in areas of greatest population
density is the steady decline of the natural resource base.
Because these processes have developed unevenly in different locations,
cropping intensities vary considerably within the WASAT. Production potentials
as well as short- and medium-term technological needs also differ along the

Potential and Practice in Food Production Technology Development

north-south axis and within climatic zones. In areas of low population density,
there is still scope in the short run for labor-augmenting technologies to permit
more efficient area expansion. In areas of highest land pressure, immediate
priority must be given to technologies that will arrest declining land quality.
Between these extremes, a mix of approaches may be appropriate, but each
approach should be viewed as a stage in the evolution toward more stable long-
run intensification.
For areas of lowest population pressure, marginally profitable animal
traction systems may permit some expansion of cultivated area per worker.
However, these require a substantial and continuing investment in support
infrastructure and in the provision of complementary inputs. Moreover, a con-
sistent approach toward the evolution to more intensive, ecologically sustaina-
ble, animal-based systems is generally lacking. In the southern and central
Sudan, several technologies are now available that can achieve substantial yield
increases in the short run under research station conditions. Yield increments of
20 to 40 percent are typical for moderate fertilizer doses, or for plowing, or for
improved land management. Yield responses of 100 percent in on-station trials
are not unusual with all these improvements. Even greater increments can be
attained by adding more input-responsive crop varieties. However, only a small
proportion of farmers who apply these innovations approach the performance
levels of experimental stations. Average yield gaps of 40 to 60 percent are
normal, resulting in high risks of financial loss and low adoption rates.
As important in the long run is the absence of proven farmer-adapted
systems to maintain soil quality under the high-input management necessary to
achieve significant yield increases. Economic means to generate, recover, and
recycle biomass at levels adequate to maintain the soil's organic matter are
particularly lacking.
Reversing this situation will require not only continued investment in
agricultural research but also important changes in conception and approach.
First, the objectives of both research and development programs should reflect
greater balance between immediate production gains and resource base conser-
vation. This implies a broader set of research criteria and a multiyear time frame
for the evaluation of new technologies. In the policy domain, this implies that
subsidies may be justified for stable antierosion systems and for farm-level
inputs with important long-term benefits. Profound structural changes are es-
pecially required to move toward more efficient, mixed farming systems. Be-
cause such systems imply radical changes in production objectives, in re-
source use patterns, and in interethnic group relations, both research and
policy interventions addressing this issue must have a long-term evolutionary
Second, the complementarities among components of improved technolo-
gies, particularly those affecting soil moisture, fertility, and varietal change,
argue for a package approach in technology development and extension. This is

The West African Semiarid Tropics

subject to two important qualifications. First, given limited farm-level invest-
ment capacity and the risks linked to poorly developed delivery systems, each
component of such packages should be profitable when used in isolation and, as
far as possible, should employ resources available at the farm level. Second,
because of differences in the quantity and quality of resources among produc-
tion units, packages should be developed that fit the needs of distinct farm
Third, research and development programs should be based on a finer
definition of regional recommendation domains. A delimitation of zones that
more accurately reflect rainfall, soils, population density, and major farming
systems would permit greater specificity in research objectives and greater
efficiency in applying results. It should be recognized, however, that, due to
differences in potential among zones, application of standard profitability crite-
ria to guide investment in production technologies will favor the relatively more
humid southern zones, thus widening interregional income disparities.
Fourth, there is an urgent need for more emphasis on research off the
research stations. The factors explaining yield shortfalls between the research
station and farmers' fields need to be identified and fed back to modify on-
station objectives and methods. Greater participation by farmers in the develop-
ment and testing of technologies also is necessary to insure earlier, more
efficient farm-level adaptation.

chapter 6

Eastern and Southern Africa

Michael Collinson

The drier monsoon tropical climate of most of eastern and southern Africa has
the world's highest solar energy intake (150 to 180 Kcal per square centimeter
per year). However, its potential for crop growth (48 to 57 tons per hectare dry
matter per year, assuming 3 percent conversion of light energy), is never
realized due to water limitations (Cooper 1970). In addition, radiation, tem-
perature, and soils modify rainfall potentials of local areas.


Little of the land area of eastern and southern Africa reliably receives more than
1,250 millimeters of rain per annum; a significant proportion receives too little
for permanent agriculture.' Zambia and Ethiopia usually fare somewhat better.
Much of the region receives 750 to 1 thousand millimeters of rainfall annually,
but because of high levels of radiation its usefulness is concentrated in a short
period. Nearly 88 percent of the area of East Africa has a single-season pre-
cipitation pattern. The rest has a bimodal pattern, with typically four wet
months and eight dry months (Morgan 1969). The monomodal rainy season of
five to seven months is particularly characteristic of eastern and central Africa.
It dominates Zambia, Malawi, and Zimbabwe, and is prominent in Tanzania
and Uganda. Peak rainfall typically occurs in December and January. Some
areas of East Africa with bimodal distributions, especially the highlands, re-
ceive too little rainfall for reliable crop production because of the short duration
of the two seasons. Machakos, Kenya (table 6.1), is a classic example. Most of

1. In this chapter, regions are defined according to the note to table 2.1; East Africa is defined
in the conventional historical sense to include Kenya, Uganda, and Tanzania.

Eastern and Southern Africa

the 700-millimeter average rainfall is split about equally between the March to
May season and the November to January season. The seasons are short and
marginal for crop production. Furthermore, variability in the timing of the start
and finish of the rainy period at Machakos and variability of rainfall during the
season create uncertainty about the choice of crop species, varieties, and per-
formance. This is characteristic of much of the region.
Zaka, Zimbabwe, provides another example of the effect of low rainfall in
semiarid zones that are becoming increasingly populated. This area has a 20
percent probability that any 15-day period in January and February will receive
less than 45 millimeters of rain. This brings the risk of crop damage on the
sandy veld soils, particularly to those crops vegetating heavily in the middle of
their growing season. Farmers in Zaka also face the uncertainties associated
with variability in the start and finish of the rains. Although the growing season
averages five and a half months, only 37 percent of seasons are 165 days or
more, 41 percent are 125 to 165 days, and 22 percent are less than 120 days
(Bromely and Lang 1980).
As population densities build up in the less well-watered areas of eastern
and central Africa, the exploitation and management of short and variable
rainfall regimes pose an increasing challenge to agricultural research.
The high radiation levels of the region reduce the effectiveness of rainfall
by increasing transpiration and therefore crop water requirements. For exam-
ple, sugar requires 2 thousand millimeters per annum of moisture to realize its
full potential. This compares with 750 millimeters in temperate regions of
Europe. Open water evaporation (Eo) varies from 1,400 millimeters per annum
in the highlands to 2,600 millimeters per annum in the desert. Much of the
region has E, between 1,800 and 2,200 millimeters.
Temperature, which decreases with altitude and increases with latitude,
also modifies plant growth potential. Lower temperatures imply longer matu-
rity periods. Maize in Kenya matures in 100 days at sea level but requires up to
300 days at 2,500 meters above sea level. Frost damages certain crops above

Table 6.1 Monthly long-term rainfall in Machakos, Kenya (millimeters)

Rainfall October November December January February March

Mean 29 160 93 55 42 84
Low 0 27 12 0 0 0
High 183 489 219 202 177 169

Rainfall April May June July August September

Mean 145 68 9 6 4 14
Low 20 11 0 0 0 0
High 346 151 35 36 20 134

Source: Stewart and Hash 1981.