Front Cover
 Title Page
 Table of Contents
 Part I. Limits to measurement
 Food systems and needs
 Defining malnutrition
 Energy and protein requirement...
 Food system indicators
 Part II. The causes of malnuti...
 Multiple causes in malnutritio...
 Functional classes and targeted...
 Part III. Food and nutrition policy...
 Agrarian change and poverty
 Markets and food availability
 Nutrition interventions
 Professional roles and researc...
 Author index
 Subject index

Title: Agricultural development and nutrition
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00087167/00001
 Material Information
Title: Agricultural development and nutrition
Series Title: Agricultural development and nutrition
Physical Description: 255 p. : ill. ; 23 cm.
Language: English
Creator: Pacey, Arnold
Payne, Philip
Publisher: Hutchinson ;
Westview Press
Place of Publication: London
Boulder CO
Publication Date: 1985
Subject: Food supply -- Developing countries   ( lcsh )
Nutrition -- Developing countries   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Spatial Coverage: India
Bibliography: Includes bibliography and indexes.
Statement of Responsibility: edited by Arnold Pacey and Philip Payne.
General Note: Published by arrangement with the Food and Agriculture Organization of the United Nations and the United Nations Children's Fund.
 Record Information
Bibliographic ID: UF00087167
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 12935194
lccn - 85050544
isbn - 081330265X (US)

Table of Contents
    Front Cover
        Page 1
        Page 2
    Title Page
        Page 3
        Page 4
    Table of Contents
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
    Part I. Limits to measurement
        Page 15
        Page 16
    Food systems and needs
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
    Defining malnutrition
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
    Energy and protein requirements
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
    Food system indicators
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
    Part II. The causes of malnutition
        Page 95
        Page 96
    Multiple causes in malnutrition
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
    Functional classes and targeted policies
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
    Part III. Food and nutrition policy and agriculture
        Page 143
        Page 144
    Agrarian change and poverty
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
        Page 159
        Page 160
        Page 161
        Page 162
    Markets and food availability
        Page 163
        Page 164
        Page 165
        Page 166
        Page 167
        Page 168
        Page 169
        Page 170
        Page 171
        Page 172
        Page 173
        Page 174
        Page 175
        Page 176
        Page 177
        Page 178
        Page 179
        Page 180
        Page 181
        Page 182
        Page 183
    Nutrition interventions
        Page 184
        Page 185
        Page 186
        Page 187
        Page 188
        Page 189
        Page 190
        Page 191
        Page 192
        Page 193
        Page 194
        Page 195
        Page 196
        Page 197
        Page 198
        Page 199
    Professional roles and research
        Page 200
        Page 201
        Page 202
        Page 203
        Page 204
        Page 205
        Page 206
        Page 207
        Page 208
        Page 209
        Page 210
        Page 211
        Page 212
        Page 213
        Page 214
        Page 215
        Page 216
        Page 217
        Page 218
        Page 219
        Page 220
        Page 221
        Page 222
        Page 223
        Page 224
        Page 225
        Page 226
        Page 227
        Page 228
        Page 229
        Page 230
        Page 231
        Page 232
        Page 233
        Page 234
        Page 235
        Page 236
        Page 237
        Page 238
        Page 239
        Page 240
        Page 241
        Page 242
        Page 243
        Page 244
        Page 245
        Page 246
        Page 247
        Page 248
    Author index
        Page 249
    Subject index
        Page 250
        Page 251
        Page 252
        Page 253
        Page 254
        Page 255
Full Text

Agricultural Development
and Nutrition



Development and


Edited by Arnold Pacey
and Philip Payne

London Melbourne Sydney Auckland Johannesburg
Boulder, Colorado
by arrangement with
the Food and Agriculture Organization of the United Nations and
the United Nations Children's Fund

Hutchinson and Co. (Publishers) Ltd
Aft imprint of the Hutchinson Publishing Group
17-21 Conna.\ Street. London W1P 6JD
Hulchinson Publishing Group I Aulralia I PtI Lid
lh-22 Church Street. Ha thorn. Melbourne. Victoria 3122
Hutchin-on Group (NZ) Ltd
32-34 \ iew Road. PO Bo\ -i -086, Glenfield, Auckland 10
Hutchinson Group I SA 1l P\ I Ltd
PO Bo\337. Berg% leit i"12. South Alrica
First published 1985
Published in 1985 in the United States of America by
Westview Press, Inc.
5500 Central Avenue
Boulder, Colorado 80301
Frederick A. Praeger. President and Publisher
1985 b\ FAO and UNICEF
The designaltons emploJed and the presentation o material in this publication
do noi impl, the c\pression ol an1 opinion % hatsoe er on the part ot the Food
and Agriculture Organization ot the Lniicd Nationsor the United Nations
Children's Fund concerning the legal tatu'.ot an counir\. territory, city or
area or o itt authorities, or concerning the delimitation of it, frontiersor
The designate ions 'des eloped" and 'developing' economies are intended for
sItatistial convenience and do not necessarily e\pre's a ludgementabout the
stage reached b\ a particular counter or area in the de elopment proc: '.
All rights reer'ed No part o this publication mj he reproduced. stored n a
rerine\ al -tem. or transmitted in an\ lorm or b\ an mean. electronic.
mechanical. photoicop ing or .other Iste. without the prior permission ol the
copi rgh t l.ner .Applicaion, btr ,uch permission. nith a statement ot the
purpose and e\leni ,t the reproduction. shouldd be addreed to the Director,
Public.iions DI inion. Food and Agriculture Organizalnon ot he United
Naiions\ it, delle Terme di Caracalla .00100 Rome, Italy.
Set in Linotpe Time' b, Sa\on Lid. Derby, England.
Printd and bound in Great Brtain bV
Anchor Brendon Lid. Tipirec. E-e\
Brilish Library Cataloguing in Publication Data
Agricultural de% elopment and nutrition
1. Food crops-Tropics
I. Pacey, Arnold II. Paine. Philip
338.1'9'0913 SB176.T7
Library of Congress Calalog Card Number 85-50544
ISBN (UK) 0 09161330 2 cased
(UKtIii.i) 1613310 paper
(US) 0-8133-0265-X cased


Foreword 9
Preface 11

Part One Limits to Measurement 15

1 Food systems and needs 17
Changes have taken place in views of malnutrition
in the last two decades, and these are outlined.
Two examples of new approaches are given the
current debate about human energy
requirements, and ecological studies of farmers'
energy balance. Malnutrition does not arise from
isolated single causes, but from dysfunction in
the 'food system'. Examples are given of the
interaction of health, work and family structure,
in the system
2 Defining malnutrition 37
Even the identification of 'malnutrition' as a
state, is not simple. Physical definitions of
malnutrition can be set lower or higher according
to their intended purpose. These purposes may
range from 'selecting the most malnourished
individuals for treatment' to 'selecting groups to

6 Agricultural Development and Nutrition

receive certain benefits'. The policy implications
of varying classification methods are discussed.
3 Energy and protein requirements 51
At this point the question 'how much food do
people need?' is discussed. Man's needs for
energy and nutrients have often been calculated,
with the use of various assumptions. These are
examined, with the question 'What is the nature
of successful and unsuccessful adaptation to
variable and low food intakes?'
4 Food system indicators 73
Chapter 1 introduced the general idea of a 'food
system', and Chapters 2 and 3 outlined current
thinking on how utilisation of food, and human
outcomes in that system may be evaluated. Data
available for describing other parts of the system
are now reviewed. Indicators of individuals'
nutritional state were discussed in Chapter 2;
here they are put alongside other indicators, and
the use of standard values and cut-off points is

Part Two The Causes of Malnutrition 95

5 Multiple causes in malnutrition 97
Little needs to be added to this chapter's title,
by way of summary. Having outlined the 'food
system', it is logical to consider some of the
interacting dysfunctions which lead to
malnutrition. Illness and poverty are important
6 Functional classes and targeted policies 120
It is one thing to outline the causes of
malnutrition, and another to direct specific
policies towards the malnourished. This is called
targeting. One way to link explanation and

Contents 7

action is to use an analytical framework which is
itself action-oriented. 'Functional classification'
is such a framework. It provides descriptions of
population groups which made administrative
sense, and which enable agriculture, nutrition
and health policies to be better targeted on
situations which give rise to malnutrition.
Examples of the implication for some
interventions are discussed.

Part Three Food and Nutrition Policy and Agriculture 143

7 Agrarian change and poverty 145
'Nutrition in agriculture' is not just about
increasing food production. In this chapter we
discuss the contradiction between increasing
production and declining consumption and the
concept of people's entitlement to food. We
review the effect of agricultural development
strategy on the social status, labour
opportunities, and hence the nutrition of
different agrarian classes.
8 Markets andfood availability 163
Markets are an important component of the
'food system', governing some people's access to
food and the returns to others' labour. Market
systems may behave in many different ways but
they normally work to the advantage of those
classes who are least likely to become
malnourished. Policies on food price supply and
processing technology tend to favour the urban
rich rather than the rural poor. We give two
important cases where nutritional justifications
for these policies are argued after the event
rather than before. These 'justifications' are
scientifically invalid and socially and
culturally irrelevant.

8 Agricultural Development and Nutrition

9 Nutrition interventions 184
Chapters 7 and 8 have shown that in practice,
government policies in agriculture and health
have considerable impact on nutrition impact
which is often not recognized. On the other
hand, several kinds of government-sponsored
action are regarded characteristically as
'nutrition interventions'. Yet under some
circumstances they may have little impact on
malnutrition, as is indicated in this chapter.
10 Professional roles and research 200
What have agricultural and nutritional workers
to contribute to one another' understanding of
malnutrition, and to programmes which may
reduce its prevalence? They can exchange
insights derived from their characteristic
approach to problem analysis; this can help to
avoid narrow oversimplified 'solutions'.
Research can be interactive- and not only among
different classes of researcher but among
researchers and 'subjects'. It is important, if we
aim to study and manipulate the 'food system', to
do so as advocates for those who are
disadvantaged within it.

Bibliography 220



During the decade which followed the World Food Conference
held in 1974 in Rome, much debate, research and evaluation has
focused on the impact of development strategies on the preva-
lence of malnutrition. Within the UN system the Food and
Agriculture Organization (FAO) and the Children's Fund
(UNICEF) have been actively involved in promoting new
approaches in this field. FAO has been at the forefront of recent
efforts to improve the nutritional impact of agriculture and rural
development. UNICEF's well-known work guides and supports
many types of development that relate to child welfare and
services, including the problems of access to food and feeding of
There is a unanimity of opinion in all circles that the
agriculture sector lies at the centre of any solution to the world's
food problems. Yet the sobering thought that has become all too
evident in recent years is that food production alone is not the
answer. Rapid progress in production and exports of food
commodities are found to co-exist with malnutrition even in the
same areas of certain countries. Therefore while increased food
production is an essential precondition, improved distribution is
also necessary to overcome inadequate food consumption and
malnutrition. The work that remains to be done is the application
of suitable approaches through which agriculture and rural
development can prevent the waste of human resources and
potential which is caused by malnutrition.

10 Agricultural Development and Nutrition

This text, prepared by the Nutrition Policy Unit of the London
School of Hygiene and Tropical Medicine, is an expos of the
many food consumption-related problems which need to be
considered alongside agricultural production issues in develop-
ment. It examines those social, environmental, economic and
political factors that determine the degree to which people have
access to food and can assimilate its nutrients. After reviewing
the present scientific knowledge of energy, nutrient require-
ments and human growth, the authors concentrate on the many
obstacles rural families face in trying to satisfy their basic food
They argue for a balanced agricultural and rural development
which would alleviate not only the technical constraints to
increased production but also those many social, economic and
political impediments that prevent people from having access to
the foods that are produced.
This compilation of materials originally prepared for a
UNICEF/FAO/Indian Council of Agricultural Research Work-
shop, is therefore a most important contribution to the thinking
that has evolved recently about an interdisciplinary approach to
food and nutrition problems. It will stimulate agriculturalists,
rural development experts, food and nutrition planners and
nutritionists all over the world in devising development
strategies that are as concerned about impact as they are about
FAO and UNICEF are pleased to have supported the
preparation of this manuscript and, while not necessarily endors-
ing its every argument, are gratified to be associated with a book
which provides a modern and responsible vision of the basic
themes and questions which they consider must guide develop-
ment. The text is intended to stimulate interdisciplinary thought,
discussion, research and action at the village, district, national
and international level. By questioning the weaknesses of
present approaches it cannot help but challenge all of us to search
for more effective answers to food and nutrition problems in
future development.


Starting in 1971, the Government of India, with the assistance of
FAO, UNICEF and UNDP, embarked on a long-term program-
me to extend and reinforce the subject of human nutrition in
higher education. The programme envisaged the teaching of
food and nutrition subjects in agricultural, veterinary and home
science colleges, and these institutions have been assisted in
building up food and nutrition departments to undertake train-
ing and research in subjects related to nutrition.
Since 1980, under the title 'Education in Food and Nutrition in
Agricultural Universities' (EFNAG) the programme has de-
veloped both in direction and content.
The general objective remains the same: to improve the
nutrition of rural families and the means of achieving this is
through three main aims:
1. to introduce a wider understanding of nutritional issues in the
agricultural sector in order to incorporate nutritional considera-
tions in agricultural programmes and hence bring about better
2. to promote an understanding of the factors contributing to
malnutrition, especially those related to poverty, with special
reference to infant malnutrition;
3. to develop a more comprehensive view of training in agricul-
ture and allied subjects that includes elements of food and
nutrition education and basic services for children: this broader
view should encompass the human aspects of agriculture, such as

12 Agrictdiiral De'uelolnneni, anduI Nutrition

development communication, extension methods, programme
planning and food planning.
Of particular concern has been the need to strengthen the
capacities of agricultural universities to include a nutrition
dimension into the training of post-graduate students, and to lay
the foundations of an expanding programme of applied and
operational research in areas which link the production of food
w ith the nutrition of rural people.
As an expression of that concern, the Indian Council of
Agricultural Res.earch. with the sponsorship of FAO and
UNICEF, organised a workshop at Hariana Agricultural Uni-
versity in April 1982. The Nutrition Policy Unit of the London
School of Higiene and Tropical Medicine wa as in ited to propose
a programme of topic areas; to provide background working
papers for the review sessions; and to conduct and co-oldinate
the proceedings generally.
The objective of the workshop was to review in depth those
aspects of the science of food and nutrition which are rele\ ant to
agriculture and to rural de elopment, and similar\ those aspects
of agricultural change which ha\e a direct or indirect impact on
the nutritional condition of human populations. With this re\ iew
as a background, the workshop \was then to determine priorities
for the inclusion of nutrition topics into programmes of post-
graduate training in agriculture and home science, and hence to
pro% ide guidelines for future curriculum and research program-
me development. This book has been based for the main part on
material presented to the \workshop. and on a record of the
discussions \which ensued.
It is conventional to preface books of this kind with broad
statements about the critical importance of improving the
nutritional conditions of human populations, and of accepting
that such improvement should be an explicit objective of
development. It is logical to assume that the pace and direction
of agricultural change will be a critical factor in achie% ing this
objective: more people consuming a better diet must have
implications for food production, and more effectl\e deploy-
ment of the means of production should be reflected in more
people able to afford to eat adequatel.

Preface 13

However, having indicated this fundamental relationship, it
has usually been assumed that the objective of improving the
nutrition of human populations at the same time as improving
the practice of agriculture will be achieved automatically through
bringing about an interchange between the two academic
disciplines of nutrition and of agriculture. But what exactly
should be the nature of that exchange? Should it be factual
knowledge? Will malnutrition be reduced if agricultural students
in future know that there are 10 amino acids essential for health,
or that beans contain more protein than rice, or that vitamin A
deficiency can be prevented by eating green vegetables? Or
should the exchanges be of concepts, of the nature of the
processes, social and economic, as well as physiological, which
result in malnourished individuals?
We believe that it is particularly important at this time to take a
critical look at how the two disciplines, agriculture and nutrition,
should be expected to interact. We hope that this will be
especially fruitful in the context of post-graduate activities, since
this is traditionally the area in which subjects develop and new
concepts are formed. Not the least reason for the timeliness is
that despite the widely acknowledged importance of adequate
nutrition as a basic component of human welfare, and despite the
manifest success of agriculture in supporting the growing popula-
tions of the world, malnutrition persists. We have both more
food and more destitute and hungry people. How can the
connection be made between our growing technical capacity to
generate wealth from the soil on the one hand and our
understanding of the biological and social needs of man for food
on the other? One thing is certain, there are no easy answers, and
the reader of this book will search in vain for prescriptions,
progress will come only through the realisation that the real
problems and solutions lie in the nature of social and political
institutions and the human relationships that underlie them. For
some people this realisation is painful, and leads either to a
rejection of science as irrelevant, or to a retreat into the more
comfortable distractions of academic research. We shall be
happy if this book helps to strengthen the view that curiosity
about the way the world works, and the urge to extend

14 Agricultural Development and Nutrition
knowledge about what things are possible in the world through
analysis and criticism can be applied to the solution of the
problems that give rise to hunger.
Peter Cutler
Elizabeth Dowler
Barbara Harriss
Philip Payne
Erica Wheeler

Those whose names appear above are the members of the
Nutrition Policy Unit of the Department of Human Nutrition at
the London School of Hygiene and Tropical Medicine. All of
them contributed material to the book. It could not have been
written, though, without the help of many other people, in
particular N.S. Jodha, Claire Kelly, David Nabarro, Adam
Pain, Paul Richards, John Rivers, Young Ok Seo and Anne
Thomson, who either work with us, or generously allowed us to
use their work. We are also grateful to Madeleine Green and
Barbara Kenmir who typed the manuscript.
Above all the book is the result of the integrating perspectives
and unstinting labour of Arnold Pacey.
Our grateful thanks go to Margaret Khalakdvia and her staff at
Haryana Agricultural University, and to Franciso Coloane of
UNICEF for their outstanding hospitality, constant encourage-
ment and careful arrangements which ensured the smooth
running of the workshop.
The costs of editing and preparation of the manuscript were
borne by FAO and UNICEF. However, neither of these
organizations are responsible for any of the views or opinions
expressed herein.

Philip Payne


Limits to Measurement

'The people are crying out for bread and we are going
to give them statistics.'
John Boyd-Orr in 1945 on proposed terms of reference for
FAO (see Orr 1966, pp. 20, 162)

1 Food systems and needs*

Changed perspectives
During the last three decades, the application of nutritional
science to the problems of hunger and malnutrition has passed
through a phase of great confidence and hope, follow\ ed by one of
increasing uncertainty and doubt. Twenty years ago, a book such
as this would have discussed well-defined nutritional interven-
tions, aimed at achieving some equally well-defined nutritional
objectives. Very little space would have been devoted to
questioning the validity of those objectives, or the effectiveness
of the programmes themselves. In so far as progress towards
reducing malnutrition was acknowledged to be slight, this would
have been interpreted as showing the need for more extended
programmes, hence for more resources, and especially for more
trained people to be deployed.
The connection between nutrition and agriculture would have
been presumed to rest on a number of premises: that hunger
would be eliminated if there were an increase in the overall
production of food; that malnutrition is often caused b\ dclicien-
cies of specific nutrients (e.g. protein) and that this can best be
countered by emphasis on the production of certain kinds of
foods, or by fortification or enrichment of staples; that poor

*This chapter is b:icjd in.iinl on nm.icrl.I pilp.iJ h1 Plulip Payne; see Rivers
and Payne (1982) and Payne (1982).

18 Limits to Measurement
nutrition is often the result simplN of ignorance: and hence there
is a general need to educate people on how to make proper use of
the resources available to them. Infectious disease, though
recognized to interact with malnutrition, would be regarded as
essentially a separate problem to be dealt with by specifically
designed programmes.
There has been a fairly radical change of viewpoint over the
last two decades. This has been broadly for three reasons.
Firstly, there has been a change in concept: malnutrition,
previously regarded as something caused by single physical
factors, is now accepted as having multiple causes, many of
which are closely linked to the conditions of inequality of
resources, or poverty, and of social discrimination. Changes in
the system of food production which leave these conditions
unchanged will also leave malnutrition unchanged or may even
aggravate it. Indeed, we have witnessed some countries becom-
ing net exporters of food, while sections of their populations
remain inadequately fed, or even experience famine. In addition
to this, among young children especially, malnutrition is so
intimately related to infectious disease that it makes no practical
sense to pursue programmes aimed at improving food consump-
tion without also tackling at least some of the environmental
causes of such disease. This may lead the nutritionist to
encourage improvements in water supply, sanitation, housing
and domestic fuel.
The second reason for a changed perspective is that there has
been progress in our understanding of the physiological and
biochemical proces.se underlying malnutrition. We now know
that man has the capacity to adapt to a fairly wide range of
dietary situations, and that only when that adaptive capacity is
stretched beyond its limits does the body fail to maintain its
functional capacity\ ." and malnutrition ensues.
Thirdly, there is a growing awareness that many of the more

*Functional capacity here is taken to mean all aspects of the behaviour of an
individual in response to the environment, such as physical and mental activity,
response to stress, disease etc. This should not be confused with the term
'functional classification' which is used in later chapters.

Food systems and needs 19
conventional types of nutrition programme have not achieved
what was hoped of them. Interventions such as the promotion of
high-protein crops, food delivery systems aimed at young
children, and projects for educating people about the nutritional
value of foods, have often been totally ineffective, or effective
only as short-term palliatives. Frequently they have also dis-
tracted attention from the need to attack more fundamental
It is in this context of doubts and questioning about past ideas
that we need to review the relationship between knowledge of
nutrition and planning for agriculture. We believe that from a
basis of analysis and criticism, we shall be able to begin a
synthesis. But the ideas that will carry us forward will be different
from those of the past. They will include ideas about food
systems, about the epidemiology of malnutrition, and about
'livelihoods'. They will also include suggestions about the use of
indicators, and ways of understanding the multiple causes of
illness and deprivation. These and other concepts give us a new
point of departure, so that we should not expect simply to
substitute a new set of nutrition programmes, more potent and
more efficient than the old a set of 'right' answers to substitute
for the 'wrong' ones. There are no simple answers, and whatever
it is that a nutritional approach to human problems can provide,
it is not likely to give us any short-cut solutions or technical fixes.
The problems of malnutrition will be overcome as and when we
overcome those of poverty, deprivation and disease. But we are
all concerned with these, and therefore we could well begin by
asking: what is so special about the nutritional viewpoint? Why
should we be especially sensitive about nutritional needs rather
than economic or social needs?
One clear lesson from the past is that when a particular
problem at first sight requires nutritional expertise for its
solution, this analysis is not always borne out by closer critical
analysis. The skill and integrity of a profession rest at least as
much on its willingness to show when it cannot and should not
play a key role, as on its ability to demonstrate the effectiveness
of its methods when they are relevant. However, there is a
general case for concerning ourselves with nutrition and with

20 Limits to Measurement
food as a fundamental aspect of agricultural de elopment, and
this rests on two propositions.
The first identifies the nutritional status of a person as both the
outcome of the process of acquiring, consuming and utilising
food, and one of the critical inputs to that process: the food a
woman or a man eats decides the amount of effort she or he can
afford to invest in order to secure food in the future. If ie can
measure nutritional status, therefore, we have a unique index\ of
the impact upon individuals of the \\ hole system of production,
utilisation and e\chance.
The other proposition is related to this: it is that. of all the
symbols and objects of social exchange. food is ;rguabl\ the imo't
basic. Co-operation in the acquisition of food. and it, sharing
among the members of a family marked the beginning of the
social evolution of mankind. The extent of an individual's
integration with society can be measured by the adequacy and
security of his or her ability\ to produce, control, purchase,
borrow or othemi\ise acquire food (i.e. his 'food entitlement'
(Sen, 1981)).*
This book is focused on a range of topics which impinge upon
different %rage, and process operating within the food "- s tem';
that is to say. the ,\stem comprising the production, distribution.
consumption and biological utilisation of food. Figure 1 1 shows,
in a much simplified way, a few of the ke\ relationships and
processes which will be dikcussled It i, not intended as a complete
definition of the 'system', Iut more a starting framework w within
which we may wish to elaborate certain areas. Then, by
understanding how people become malnourished in terms of
processes within the system, we ina\ be better able to make
statements about who such people are. and to describe their
relationship \ ith production and the basis of their entitlement to
tood. Malnutrition. in this context, is a symptom or signal that
cei tain processes are regularly occurring in the lives of people
which, if disregarded. will result in the continued generation of
sickness and ph\-iological impairment.

*FIll reli icnce, quoted in the text are contained in thc Billliogr iph\ beginning
on p.220.

Food systems and needs 21

Figure 1.1 The food system as it is envisaged in this book, in its
simplest form

Analysis of the nutritional problems within a society may
therefore have implications for agriculture at many different
levels. At one level, it may consider the impact of agricultural
change on the ability of people to earn an entitlement to food. To
produce more, but at a cost many people cannot afford, may be
self-defeating. An orientation of technology to production,
which neglects problems of consumption, needs to be corrected.
At other levels, the analysis points to the need for an integrated
approach by planners to the development of land resources for
food, for fuel and for cash crops. It underlines the need to
improve the domestic environment as well as to plan resource
inputs to agriculture. It may also indicate the need to avoid
certain directions of change because of the danger of aggravating

22 Limits to Measurement
the risk of malnutrition, or may at least demonstrate the cost of
that effect in assessing policy choices.
These, then, are the themes of this book. Our starting point is
actually near the bottom of the diagram, which is where we find
the more usual focus of food and nutritional science. The
traditional emphasis continues in much teaching and research,
that is to say on the properties and uses of foods rather than on
the nutritional problems of populations. Part of the purpose of
this book is to present a reverse perspective to put the problems
of people first and see what questions they prompt about
nutrition. To reverse the traditional perspective of nutrition in
this way is not, of course, to say that everything done previously
was mistaken. What we would argue, though, is that if one
constantly approaches a subject from the same point of view,
important insights may be missed, and received views may not be
adequately questioned. For example, to tackle the question of
how much food human beings require we first ask what kind of
life those individuals lead, and what physiological functions,
therefore, does their food intake have to support?
If the science of nutrition has any central concept, it is surely
this notion of nutrient requirements. The initial impetus for the
subject came from nineteenth-century attempts to define those
requirements, and in many quarters, that approach is still
important. The process of revising estimates of energy and
protein requirements continues unabated, and the problem of
fixing them remains apparently unresolved. What has changed
since the early days is that originally pronouncements about
nutrient requirements were made by individual scientists and
were based on their own research; now, similar pronouncements
emanate from committees of experts convened by governments
or international agencies and to a large extent reflect the experts'
selection of scientific evidence and of theory (Rivers and Payne,
Despite the authority and influence associated with some of
these figures, the concept of a 'nutrient requirement' remains
somewhat intangible. Discrepancies between estimates, and the
revisions that are made from time to time, do not seem to reflect
differences in technique or theory, but may have more to do with

Food systems and needs 23
political pressures or changing social valuations of the accepta-
bility of particular intake levels. For example, successive esti-
mates of food energy requirements made by the US National
Academy of Sciences for a moderately active man with a body
weight of 70kg have been as follows (NAS 1943, 1958, etc.):
12.5 MJ (3000 kcal) in 1943;
13.4 MJ (3200 kcal) in 1958;
11.7 MJ (2800 kcal) in 1968;
11.3 MJ (2700 kcal) in 1974.
The 16 per cent fall in estimated requirements since 1958 is not
the result of improvements in the process of estimation, nor of
any factor such as change in levels of activity, but must be
attributed more vaguely to climates of opinion and perhaps
concern about obesity.
Nonetheless, the questions 'how much food does a man/
woman/child need?' and 'what are the reasons for and consequ-
ences of failing to meet that need?' remain central to agricultural
development. We set out below some of the issues involved in
answering them and the problems raised thereby. We use a series
of short case studies (from different parts of the world) to
illustrate them.

Food energy and agrarian ecology
Lack of progress in attempts to define human energy and
nutrient needs illustrates one reason why it is useful to think in
terms of the level of nutrition needed to maintain the functional
capacity of the body. If we take the traditional view of a nutrient
requirement as the minimal amount of nutrient needed to
maintain a given physiological state such as 'health', then we are
faced with the problem of specifying that state. And although
health and well-being are certainly to be valued as social goals,
the utility of health as a reference criterion is limited by our
inability to define a state of ideal health which an adequate
nutrient intake should sustain.
The social nature of the definition of health has been

24 Limits io Meo tirenent

frequently discussed else\\here (Illich, t197; McKeown, l'7't),
its use a a criterion tor filing energy requirement is particulaihl
complicated b% the fact that it seems impossible to identify a
health\ population. As one official hod\ states, most of their
recommendations on food enerY\ are based on measurements of
what actual populations eat. assuming that the..e populations are
healthh. Ho\e\ er. they point out that 'l;Man\ groups of people
... li\inw in the industrialized countries are obese. On the other
hand some groups living in developing countries are small in
stature, light and thin. yet may not be ph!isicll\ less, health
because of their different body size' (FAO/WHO, 1973).
Another approach is to attempt the definition of energy
requirement in terms of the maintenance ot a specified ph\ ,iolo-
gical state. disregarding notions such as "health' and this leads to
the apparently more precise concept of 'nutrient balance', a
decepti\el\ simple notion which has been \widel\ applied in
animal nutrition (Blaxter, 1967).
Ob\ ious,, in the non-pregnant adult animal, nutrient intake
and expenditure must match it the body content of the nutrient is
not to rise or be depleted. The problem is that balance can be
achieved over a ranee of intakes thlioulh adaptations of various
kinds. Thu, the question 'which level of nutrient balance is
prefer red" 's is inescapable, and can only be answered by referring
back to 'health'.
The balance method is thus no more objective th;n the health
criterion. The decision about \what le\el of equilibrium or for
children, what rate of grow th, and hence what degree of positive
balance, will be regarded as a norm remains entire sIubjecti\e
unless, we are prepared to specify a particular set of desired
functions and a set ol undesirable s hmptom, we wish to avoid.
The decision whichh is usually made (albeit not always e\plicitl.\
is to say that we prefer the levels and growth rates which are
t pical of \\etern developed countries, though we are discover-
ing that these are not \ itholut diadantar e for health.
It is in order to a\ oid such ambiguities and difficulties that thin
book uses functionall' criteria of the adequacy of food intakes.
From this point of view, malnutrition is defined as a state in
which the ph\hical function of an individual is impaired to the

Food systems and needs 25
Table 1.1 Energy intakes of New Guinea adults

Village Body weight Energy intake per day

kg MJ kcal
Kaul (coastal location) males 56 8.12 1940
females 47 5.94 1420
Lafa (highland location) males 57 10.54 2520
females 51 8.79 2100
Source: Norgan et al. 1974.

point where she or he can no longer maintain adequate perform-
ance in such processes as growth, pregnancy, lactation, physical
work, or resisting and recovering from disease. The notion of an
adequate level of performance is itself not a simple one.
However, it avoids the rather greater difficulties of talking about
adequate levels of health if, in the first instance, we use it to mean
the achievement of a sustainable mode of existence. Thus, for
example, to avoid malnutrition, members of a farming family
must be able to do physical work on their land and crops,
sufficient not only to secure their immediate food needs, but to
sustain productivity over long periods, and to survive through
bad years as well as good. They must in addition be able to adjust
total production from their resources to keep pace with the
demographic changes (numbers of dependants and number of
productive adults) in the family itself.
In taking this approach, we regard dietary energy as the most
likely limiting factor. This does not imply that vitamin and other
deficiencies do not occur, or are of minor consequence. Rather
we are suggesting that the avoidance of energy constraints is a
necessary even if not always a sufficient condition for avoiding
In the context of agriculture, the problem of energy deficiency
is quite a subtle one, partly because of the wide range of different
farming ecologies that can be successfully exploited, and partly
because we have to look at equilibrium situations where the
working capacity of adult family members is both the outcome of

26 Limits to Measurement

a particular level of food consumption, and the labour input to
the production system which generates the food. Viable equilib-
rium can therefore be maintained over a range of levels of intake
and output. As an example, Table 1.1 shows the average food
energy consumption levels of men and women in two villages in
New Guinea.
The people of both villages had similar and generally good
states of health, without any signs of malnutrition. However,
members of the coastal community needed to devote only small
amounts of labour to their gardens, and despite a plentiful source
of fish, rarely bothered to go and catch them. Most of their day
was spent in leisure and socialising. The highland community, by
contrast, lived and worked on steep mountain slopes, and spent
several hours a day in fairly heavy labour. The two situations are
both successful from the point of view of health and survival.
Food energy needs are different, however, and if malnutrition
did occur at some future time because of limited land access, or
because of some climatic event, it would do so at different levels
of food consumption.
Consider now rice cultivation. It is estimated that dryland rice
requires approximately 130 man-days of labour input per hectare
in one environment (Bayliss Smith 1981), corresponding to a
work input of about 390 MJ per hectare. It will yield on average
about 15,000 MJ net edible yield of rice about 1000 kg so the
ratio of energy output to input is 38:1. This is not an unusual
figure, since according to Leach (1975) energy ratios for pre-
industrial crops are generally in the range from 13 to 38. When
food energy production in a pre-industrial society is at or below
the lower end of that range, it is likely that not enough food is
being produced to support non-agricultural work such as water-
carrying or house-building, and to support non-working mem-
bers (old people and children), so malnutrition in the sense of
functional failure becomes increasingly probable.

Changes in energy needs over time
At this level of analysis, the picture presented is of a simple,

Food systems and needs 27
static equilibrium, with energy flows averaged over time, and no
account taken of the effects of either regular seasonal factors, or
irregular unpredictable hazards. Seasonal factors may impose
two kinds of restriction, in the first place related to time intensity.
Certain tasks must be carried out in specific seasonal time
periods and this may impose a lower limit on the number of
working adults or their equivalents per hectare necessary to
secure a given type of crop. Secondly, work intensity may impose
limitations bearing in mind that the physical work required for
certain tasks may be much greater than average and may only be
sustainable by individuals with a high level of physical fitness, i.e.
good nutritional status, and freedom from disease or injury.
Further, agricultural seasons are not always the same; there
are good years and bad years. Survival in bad years means either
the ability to achieve a surplus during good years, with some
means to carry this over as stored food, as cash, or as assets which
may be collateral on loans, or some other opportunities for
converting labour resources to food or cash by alternative
An additional, longer-term result of the passage of time is that
the size and demographic structure of the family will change:
total food needs increase steadily as successive children are born,
but later the children grow up and contribute to the family's
labour power. Together, then, the influence of seasonal cycles
and family development is such that we need a dynamic view of
the whole food system, of the functions which take place within
it, and of the environmental and temporal changes which form its
A village in Burkina Faso, West Africa, where millet is the
principal crop, illustrates one kind of sharply seasonal contrast. In
this area, the dry season is a long one; it is followed by a relatively
short period when the soils are sufficiently softened by the rain to
be workable, and when crops must be sown. During this period,
therefore, long hours of labour are necessary. The time intensity
and work intensity of the effort people must make are both high.
Human energy expenditures are, as a result, comparable to those
found only in coal mining in more industrialized economies. By
contrast, during the dry season there is little work to be done and

2S Limits to Measurement

Table 1.2 Energy expenditures of farmers in an Upper Volta vilage
Ent eiy expended per lda
MJ kcal

ry women 9.7 2320
Dryseason men 10.1 2410
Wet season women 12.1 2890
et season men 14.4 3460

Sources: Bleiberg et al. (1980); Brun et al. 1981.

energy expenditures are low. The men in particular have a very
sedentary lifestyle and their energy expenditure drops sharply
(Table 1.2).
This example sho ws that "hen "e look at energy expenditure
and food supply on a single occasion, or indeed, averaged out
over a year, we can tell very little about the adequacy of diet, or
the likelihood of malnutrition. For these West African villagers,
malnutrition might consist of inability to meet peak labour
demands because of either insufficient tood a\ ailable to balance
expenditure at that time, or because of bod\ stores insufficiently
replenished since the previous working season
In all probability, people using this kind of production system
face problems of both time and work intensity, and their
responses are partly beha\ ioural and social. part\ physiological:
the periods of peak work output, which may be quite short in
time, are sustained partly at the expense of body energy stores.
Thus people lose %weight, and then regain it during a subsequent
period when \ork output is dramatically reduced, but food
intake is maintained at the previous level, or even increased.
Figure 1.2 sho"w these effects as they have been observed in
Gambia. This country is also in \\'et Africa, and the seasonal
ecology is similar (though less extreme) to that of Burkina Faso
(Fox 1953). Periods of peak energy expenditure are associated
with harvesting, and with soil preparation and planting.
Ob\ iously, there are physiological limits to the extent of this
cyclic process: if too much weight is lost, perhaps because of a

Food systems and needs 29

65 kg-

60kg /

130 / /

55 kg-
120- 1

Women I

50 kg- 110- /
/ I
/ I

4 5 kg 1 0 0 J i l l 1 1 I l l l l i 11 1 'i l l l i il i l l i'I
July Oct. Mar. June Sept. Mar. June Sept. Dec. Mar.

1947 1948 1949 1950

Figure 1.2 Fluctuations in adult body weight by season in a Gambian
Source: Fox 1953

30 Limits to Measurement

Table 1.3 Analysis of labour time for different operations of
cultivation of HYV and TVpaddy in North Arcot District, Tamil Nadu,

Agricultural Percentage of Additional person- Percentage of
operation person-days spent days for HYV over female labour
in the operation TV per hectare used


Ploughing 18 18 5.9 0 0
Manuring 2 3 1.2 22 26
application 0.5 0.4 0.2 0 3
Pulling seedlings 4 4 2.2 17 21
Transplanting 15 15 6.2 100 98
Weeding 12 12 3.7 100 97
application 0.2 0.1 0.5 0 2
Harvesting and
threshing' 32 28 24.8 57 60
Irrigation2 13 15 1.5 0 1
Others 3 4 3.0 29 55

TOTAL 100 100 41.2 47 47

Labour energy
input per ha 830 710 120
Net edible rice per
ha as food 33,000 24,000
energy output (GJ)
Energy ratio 40:1 34:1

1 The labour required for harvesting and threshing could not be
separated in many instances.
2 The work required for irrigation involves switching on pump sets,
waiting for electricity, or operating the kavalai.
3 Labour input is estimated roughly as 3MJ per person-day.
Source: Based on Table 14.2 in Chinnappa and Silva 1977.

Food systems and needs 31
poor previous season, then peak work output may be impaired,
reducing the potential for next year's crop still further. Despite
recent research into the nutrition and physiological work, most
experimental measurements have sought to measure the steady-
state behaviour of individuals and animals. Thus relatively little
is known about responses to fluctuations in intake or work
output, or about the body's energy and nutrient storage mechan-
isms. Yet the latter may have a very much larger role than
previously realized for key vitamins and minerals as well as for
energy (Longhurst and Payne, 1979). Lacking knowledge of
these, we cannot assess the risks of malnutrition for people living
in highly seasonal environments.
Data on seasonal energy relationships in the context of rice
cropping is available from the North Arcot district of Tamil Nadu
state (India). Table 1.3 shows a breakdown of the time spent on
various stages in the agricultural cycle, both for traditional (TV)
and high yielding (HYV) varieties of rice (Chinnappa and Silva,
1977). The change to HYVs in this region increased yields to the
point where a family of five people could theoretically meet its
basic energy needs by cultivating only 0.5 ha. This estimate is
based on a food energy output of 33,000 MJ per hectare
annually. The ratio of energy output to input was then 40:1,
which is slightly higher than before HYV rice was introduced.
It is clear, then, that a family which adopted HYVs did not
require any additional labour to meet its own food requirements.
But since yields were greater, these requirements could be me"t
from less land. The remainder of the land formerly used c ld
then produce a surplus for sale, but only if more labour was
employed. In practice, most families adopting HYVs were those
with the larger holdings, and they secured the extra labour by
hiring workers from among the smaller cultivators and landless
families. That particularly affected the employment of women
for tasks occupying short, intensive seasons such as transplant-
ing, weeding and harvesting (Table 1.3), which is significant in
relation to a seasonal peak occurring locally in the birth rate
(Chambers et al., 1981). At the time when they need to be
working in the fields, many mothers have very young infants
whom they should be breast-feeding. This can mean that feeds

32 Limits to Measurement

are hurried, or badly spaced, and babies do not get enough
(Devkota, 1981; Rickleton, 1981).
Not only do work patterns and birth rates commonly vary
markedly with season, but so do infectious diseases. When
seasonal peaks of infection coincide with times of food shortage
and hard work, the consequences of malnutrition may be
compounded. The greatest regular seasonal impact on rural
society is often due to infections and parasites (Bradley, 1981),
and these have strong interactions with nutrition.
Clearly, disease can be both a cause and a result of poverty. It
affects the efficiency of work, and that is one of several reasons
why the previous discussion of the minimum energy inputs
needed to feed a family from a rice crop is a considerable
simplification. In practice, the family's survival would depend on
exceeding these levels, so that there is a margin to allow for times
when individuals are not working because of illness, as well as to
allow for grain lost in store (5-10 per cent) or retained for seed.
Furthermore, although the rice could supply all the energy and
protein needs of the family, it would not fully supply vitamin and
mineral needs. The family would thus have to invest more labour
either in farm production of vegetables and other foods, or in
paid employment, or in an increased production of rice for sale
and exchange.
Very often, the levels of labour input referred to earlier would
need to be increased by up to 50 per cent to cover the needs for
exchange purposes, and to provide an insurance against poor
harvests and crop levels. This insurance investment would
depend upon the level of ecological variability of the region. In
an extremely variable climate such as that of Burkina Faso, the
traditional practice before 1920 was to aim for a level of stocks
after harvest equivalent to 2-3 years' grain consumption
(George, 1980). However, the ability to make any insurance
clearly depends on the labour and land resources available. In
South Asia the marginal farm family with land restricted to less
than their subsistence production needs will be in a precarious
situation with regard to bad years and the landless even more so.

Food systems and needs 33

Family development
Seasonal changes in food supply must be studied in relation to
the longer-term cycle of the family's demographic expansion, as
children are born and grow up. During the first year or two of
marriage, both parents are usually able to contribute in labour or
other economic activity, particularly if the timing of pregnancy is
such that the women can still work during the time-intensive part
of the production cycle. Earning or production capacity will then
be high in relation to the number of family members to be
supported, and the economic dependency ratio is low.
Things may become more difficult as the number of children
increases and if and when they are at school. Later, as the
children take on more and more work, the economic burden
carried by each working members of the family eases. Food
energy needs in a family and the number of workers necessary to
support that family are thus constantly changing over time.
Young children require relatively large amounts of dietary
energy. A one-year-old child is one-fifth the weight of an adult,
but his/her energy intake will be about half the adult level.
Somehow the young child has to take in a relatively large amount
of food. Much depends on the frequency with which a child is fed
and on the amounts of milk she/he is given. Rutishauser (1974),
working in Uganda, found that the factors associated with a good
energy intake in children under two years old were:
1. breast-feeding continued while other foods were introduced
2. three or more non-breast feeds per day
3. 'good appetite', i.e. absence of infection.
It is of critical importance that the mother or whoever cares for a
child should have sufficient time in the working day to help that
child to feed, and in some cases, to prepare special items for
him/her, because even though the family diet may have a good
nutrient density, certain nutrient-rich items may be too spicy or
hard to handle. The high growth rates and high relative energy
needs of young children require frequent concentrated feeds, yet
the lifestyle of many poor families makes this extremely difficult
to achieve, especially at seasons when work is highly time-
intensive (Wheeler, 1982).

34 Limits to Measurement

We have already noted that adult agricultural workers some-
times lose weight during peak labour seasons and then regain it at
times of year when there is more food available and less work
This fluctuation may be regarded as a satisfactory adaptation to a
changing environment. However, if pregnant and lactating
women suffer energy deficits during periods of hard physical
work, their offspring will be smaller and more vulnerable (Paul et
al., 1979). Evidence that Asian women's energy status does
fluctuate seasonally is provided by the observations of Chow-
dhury et al. (1981) in Bangladesh. There, the women of landless
families tend to lose weight during the August-October period,
when the highest labour demands for rice cultivation are just
over, and the price of rice is highest. Should these women be
pregnant or lactating at that season, the risk of malnutrition t9
both mother and child is increased.
Illness may temporarily alter nutrient requirements. Gastro-
intestinal diseases often lead to malabsorption of food, and many
illnesses lead to tissue breakdown or blood loss to a varying
degree. During the acute phase of an illness, nutrient utilisation

Table 1.4 Food energy needs and labour available at three stages in the
development of a family

Food energy Adult male Ratio of energy
needs labour need to labour
MJ/year equivalent available

Woman and man,
1 child under 2 years 9,500 1.8 5.3
Woman and man,
1 child under 2,
1 child between 2 & 3,
1 child between 4 & 5 14,500 1.8 8.0
Woman and man,
1 child under 5,
1 child between 8 & 10*
1 child between 10 & 12*
1 child between 12 & 15** 21,500 3.6 6.0

Notes: assumed equivalent to 0.5 of an economically active man;
** assumed equivalent to 0.8 of an economically active man.

Food systems and needs 35
falls so there is need for some increment during recovery.
Clearly, this is a more serious matter for children than for adults;
they need nutrients and energy for growth, and also their body
nutrient reserves are small. This may be seen particularly in
children with acute fevers or gastro-enteritis, who can become
wasted and also dehydrated very fast, unless there is a deter-
mined effort to keep their fluid and food intake as high as
possible (Mata, 1977; Briscoe, 1979).
For a great variety of reasons, then, human nutritional needs
vary markedly throughout life. There is a steady fall in necessary
intake in proportion to body weight until adulthood, after which
the appropriate level of intake depends greatly on work and
activity, and on physiological stresses, of which the most
important for women is pregnancy.
Human requirements for food are not constant, but vary with
the development of family groups as much as for individuals.
Table 1.4 presents typical figures for a particular family at three
stages in its development, and shows how there may be a critical
period when the children are all young. At this stage, energy
needs in relation to the labour available to meet them peak
sharply. During this period, there is a maximum likelihood of
malnutrition occurring, particularly at the most intensive stages
of the seasonal cycle. Thus the focus of many nutrition program-
mes on the problems of young children makes sense not only
because of the vulnerability of the children themselves, but also
because of the vulnerability of the family group as a whole while
the children are young.
To understand the prospects of the family group in more
detail, however, we need to look beyond the children and
enquire whether the family owns and cultivates its own land, and
whether it is able to hire farm workers at especially busy times of
year, since that is an important way of dealing with peak energy
demands. Families with little or no land of their own will often be
in more critical situations. Most difficult of all are the seasonal
problems of families with small plots of their own, but who are
also partly dependent on paid work. For them, the best
opportunities for farm employment are likely to coincide with
the heaviest demands of their own crops.

36 Limits to Meastrement
Any investigation of the problems faced by such people should
also take account of agricultural developments which may
modify the magnitude and timing of seasonal effects, crop
varieties and irrigation.
All our examples so far have been of rural agricultural
households, but the principles could be applied to other groups
such as urban unskilled labouring families, plantation workers,
pastoralists, fishing families, and so on. Investigation of social or
economic differences, of the effects of technical change, or
seasonal constraints, and the problems of family development,
entails classifying families by land ownership and tenure, by
dependence on markets, and by major sources of income. This
would be an example of what is referred to later (Chapter 6) as a
'functional classification' of the population being studied. Ex-
amining the prevalence rates for malnutrition season by season
for each of the groups identified, enables us to study malnutrition
on a comparative basis. Because of the importance of this
approach, we shall repeatedly return to such questions as land
tenure and cropping patterns. First, though, it is necessary to
consider certain other nutritional concepts, and these are the
subject for the next two chapters.

2 Defining malnutrition*

Comparative views
Malnutrition, according to the previous chapter, should be
understood in terms of failures of bodily functions. The condi-
tions under which it occurs must be appreciated as specific to
particular localities, with characteristic agricultural ecologies
and work patterns. People adapt to a great variety of dietary and
work regimes, and this sometimes involves modifications in body
size, levels of physical activity, and metabolic changes. Detecting
malnutrition is a matter of detecting degrees of such change
which carry unacceptable penalties in terms of hunger, illness,
dysfunction and risk of dysfunction. We shall return later to the
implications of the word 'unacceptable', and merely note that
these ideas may together be characterized as an 'individual
adaptability view' of malnutrition.
It is this view upon which the present book is based, and to
avoid misunderstanding, we need to compare it with other
concepts, of which the most important is the 'fixed genetic
potential view'. The basic premise here is that there is an optimal
or preferred state of health, fixed for each individual, and
determined by his or her genetic potential for growth, resistance
to disease, longevity, and so on. It is assumed that everyone

*This chapter is based on material prepared by Elizabeth Dowler and Philip
Payne. See Dowler et al., 1982.

38 Limits to Measurement

could and should achieve their full genetic potential and that
malnutrition starts as soon as there is any departure from the
preferred state. We cannot measure human genetic potential,
however, so it is further assumed that the standards of body size
and food intake observed in 'well-fed' and 'healthy' populations
approximate to this optimum.
It may be useful to compare these contrasting views with one
which gained acceptance two decades ago. Jelliffe (1966, p.8)
regarded malnutrition as, 'a pathological state resulting from a
relative or absolute deficiency or excess of one or more essential
nutrients, this state being clinically manifested or detected only
by biochemical, anthropometric or physiological tests'.
Two points made by this definition should be especially noted.
Firstly, it implies that deficiency or excess of nutrients is the
major cause of the pathological state referred to, whereas we
would now wish to emphasise the role of many other contribu-
tory causes (Chapter 4). Secondly, Jelliffe states that malnutri-
tion can be detected only by biochemical, anthropometric or
physiological tests. It cannot be deduced from an individual's
level of food intake, nor can it be estimated from the average
intake of a population.
In this second respect, the individual adaptability view and
Jelliffe's definition are in close agreement, and stand in opposi-
tion to the assumption often made that malnutrition can be
inferred whenever intakes of nutrients fall below certain fixed
levels. Many statistics which purport to represent the number of
malnourished people in the world (or in specific countries)
actually represent the number with low food intakes. We cited
one example in the previous chapter of two societies of apparent-
ly healthy populations whose energy intakes were markedly
different. We also mentioned the importance of food in cultural
and social exchange, and suggested that an individual's ability to
obtain food may be a good indication of his integration with
society. Thus the number of people whose food intakes fall
below a certain level may be an important statistic for measuring
poverty or social differentiation (Chapter 6); it may often include
some malnourished people and some who are seasonally hungry.
But whatever its social significance, such a figure does not
measure the prevalence of malnutrition in its scientific sense.

Defining malnutrition 39
The manifestations of malnutrition vary. In what follows,
unless otherwise stated, we are referring to protein-energy
In a population with relatively low food intakes, only a
minority of children may show clinical symptoms of malnutrition
such as oedema, fatty liver, hair and skin changes, or severe
muscle wasting, but others may very well be affected especially
with regard to their vulnerability to infection. Other clinical signs
include diminished subcutaneous fat, and low body weight
relative to height. Beaton and Bengoa (1976) and Alleyne et al.
(1977) give useful descriptions of these signs and symptoms, and
of the recommended methods for weighing children and taking
measurements of height (or length of infants), and of arm
But the question then arises: by what standard should we
decide whether body weights are low? If this is one of the
populations cited in the previous chapter where it is 'normal' for
people to lose weight markedly during certain seasons, are we in
a position to make allowance for that? If the average height of
the people is unusually small, do we interpret this as an
abnormality and call it 'stunting', or might it be an 'adaptation' to
environmental conditions? If the people are New Guinea
tribesmen whose children are relatively free from illness but
whose environment has unusual features, 'adaptation' would
seem the appropriate description. But with Indian village
children whose growth is checked by episodes of serious illness
from which recovery is slowed by low food intake, 'undernutri-
tion' would seem a more relevant description. Between these
two extremes there are many uncertain categories. Moreover,
we still do not know very much about how to distinguish
populations going through 'normal' seasonal cycles of weight loss
and recovery from those whose body reserves had been dimi-
nished by a downward spiral of impoverishment. Thus questions
about how we interpret small body size or low body weight are
particularly important.
What we really need to know about are the relationships
between body size and such functions as immune competence,
mental function, physical work capacity, or the likelihood of
survival. All these functions are interrelated, but the links

40 Limits to Measurement

between them are not well understood. As Chapter 4 will
demonstrate, little is known about the relationship of anthro-
pometric data to mortality risk and morbidity, but this suggests
that we shall probably never be able to locate fixed points on,
say, a weight-for-age scale which will sharply divide individuals
at risk from those we regard as normal. To the extent that such
cut-off points can be established at all, they will probably only be
valid for specified population groups living in particular locali-
ties. So, for example, it seems impossible to classify mortality
risks for a group of children in Zaire using anthropometric
criteria which worked well for children in Bangladesh (Kasango
Project Team 1983).

Screening and anthropometry
The practical importance of establishing and defining criteria for
labelling people as 'malnourished' can be illustrated by discus-
sing the problems of selecting a subgroup of the 'malnourished'
for treatment or intervention.
Despite the difficulties of classification, in any practical
situation where action is contemplated, there will be a need to
have some procedure for comparing risks and for allocating
resources. We might in some circumstances consider that risk of
death is the most important criterion and confine ourselves to
this. We shall find, however, that even in such a relatively simple
situation, classifying individuals as 'at risk' will depend not only
on the risk relationship but also on the resources available and
the logistics involved.
For example, Chen et al. (1980) have reported the experience
of a sample of 2019 children in Bangladesh. Their weights,
heights and arm circumference measurements were recorded
when they were one year old. After two years, during which only
treatment for diarrhoea was available, 112 of them had died.
We can employ the data about these children to show what the
effect would be using two different variables, weight-for-age and
weight-for-height, as a basis for allocating some more effective
treatment. To do this, we imagine two kinds of situation, in both

Defining malnutrition 41

of which it has been decided to use anthropometry for screening
the group that is, as a means of selecting individuals who will be
given food rations or provided with some kind of treatment
aimed at rehabilitation. We shall assume that resources are
limited, and that the treatment, if given, is equally effective in
reducing the risk to a level close to the minimum, however
severely the recipient is affected. The two situations are:

1. The population to be screened is of fixed and known size, such
as an entire refugee camp; our problem is simply to decide who
gets treated and who does not.
2. The 2019 children are just a sample of a very much larger
group, and there is a continuous queue of applicants for
Table 2.1 shows the numbers of children in various categories

Table 2.1 Numbers of children out of a population of2019, who
would have been identified and treated for malnutrition, using two
different indicators'

Number Number of Number of Deaths pre-
identified as preventable false vented per
in need of deaths2 positives treatment
treatment treated3

<60% of reference 427 48 379 0.112
<75% of reference 1473 92 1381 0.062
<70% of reference 75 11 64 0.147
<80% of reference 641 39 602 0.061
<90% of reference 1620 92 1528 0.057
Treatment of the
total population 2019 112 1907 0.055

1 Data taken from Tables 1 and 2 in Chen et al. 1980, which consists of
a study of mortality rates in 2019 Bangladeshi children during a two-
year period following a survey of nutritional status.
This is the number of children who actually died in the group studied
by Chen et al.
3 Number of children who survived.

42 Limits to Measurement

of weight-for-age and weight-for-height whose deaths in our
theoretical case would be prevented if they were selected for
treatment. It will be noticed that instead of quoting the actual
weights of the children, we quote them as percentages of a
reference weight or standard in this case the 'Harvard'
standard. This represents the average weight of a large sample of
apparently healthy children at the same age as the individuals
being measured. Recording data as 'percentage of reference' for
each individual's age allows older and younger groups to be
compared directly, or allows the data for all the children to be
amalgamated. (We consider this practice in more detail in
Chapter 4.)
In this example, those children who are below a stated
percentage level are selected for treatment and the second
column in the table shows how many would be involved. Thus if
60 per cent of reference weight-for-age were taken as the critical
point and if all children under that were treated, this would entail
a total of 427 treatments. Of these, 48 would be given to children
otherwise likely to die, so there would be a ratio of one 'genuine'
to 8.9 'false' positives being treated.* Obviously, if the critical
point on the reference scale were raised from 60 towards 100 per
cent, the numbers of treatments that would have to be given out
would rise and eventually cover the whole population. The ratio
of 'false' to 'genuine' subjects treated would also increase.
Supposing the group of 2019 children comprised a complete
community and the critical level depended on the number of
treatments available. If there were resources for, say, 500
treatments, a point slightly above 60 per cent would give the best
allocation, reaching over 50 children who would otherwise die;
1000 treatments would reach 80, and so on.
If weight-for-height were to be used, the situation would be
somewhat different. At low critical points, such as 70 per cent
reference weight-for-height, the selection is much more specific
than by using weight-for-age. It excludes many false positives,

*The notion of a false positive here refers to mortality and not to mortality plus
morbidity, where the concept might have to be modified.

Defining malnutrition 43
and gives lower ratios of 'false' to 'genuine' subjects treated;
however this specificity is at the expense of also excluding many
'genuine' subjects as well only 11 deaths would be prevented.
If we again consider what happens when 500 treatments are
available, we find that they would have to be allocated on the
basis of about 75 per cent of reference weight-for-height, but that
only 30 of them would then go to children who would otherwise
die. With 1000 treatments, we would need a cut-off point at
rather less than 85 per cent, and 60 lives would be saved, and so
on. This, of course, assumes that the kind of risk assessed by
these two variables, weight-for-age and weight-for-height is the
same: that high risk cases identified by both would respond
equally well to the same treatment.
The other situation, where we envisage a continuous queue of
children being brought for treatment, is more complex. If the
rate at which individuals can be treated (the rate of admission for
rehabilitation, or the rate at which food supplements can be
delivered) is fixed, we might want the critical point for screening
and selection to be that which identified individuals for treatment
at a matching rate. On the other hand, we would want as low a
ratio of treatments to preventable deaths as possible. The second
requirement would be satisfied by using 70 per cent of reference
weight-for-height as the cut-off point. But then the first require-
ment might only be realized if the measurement were carried out
faster, for example by employing more people to do it, thereby
selecting more children for treatment per unit of time.
We have elaborated this example to illustrate that however
artificially we constrain the system, we cannot expect to find a
unique answer to the question of how to classify the mal-
nourished. It would of course be convenient if a particular
percentage of reference weight-for-age could be taken as an
unambiguous cut-off point or boundary between 'severe' and
'moderate' malnutrition, or between 'malnutrition' and 'normal-
ity'. Some systems for classifying anthropometric data imply
distinctions of this sort, but the example shows that wherever we
draw a dividing line, some individuals who are not at risk will be
selected for treatment while others who are at risk will not. Thus
in practice, the precise percentage of reference weight we take as

44 Limits to Measurement

a dividing line or cut-off point is chosen according to our
resources, circumstances and purpose.
It is clear how the resources available, and how the circum-
stances of a refugee camp or, alternatively, an endless queue of
children, may alter procedures. As to purpose, we may compare
this example of screening to select children for treatment, with
several other purposes for which body weights and heights may
be recorded.
For example, when the illness of a single child is being
diagnosed, his weight might well be noted, but it is clear from the
figures in Table 2.1 that weight-for-age would not signify much
by itself. If the child has been attending a clinic regularly, and his
weight has been plotted on a growth chart each time, any
faltering of growth will be readily recognized. But the clinic staff
will also examine the child, looking for evidence of infection as
well as for classical signs of malnutrition such as oedema (though
these signs are present only in severe cases). Enquiries might also
be made about the child's home circumstances, to assess whether
food has been in short supply, whether there are particular
hygiene problems, or whether the mother is well enough, or has
sufficient time, to prepare food for the child. In other words, to
form a clear view of an individual's condition, one requires a
combination of different kinds of evidence, and as always,
diagnosis of an individual condition and needs is by a combina-
tion of clinical experience and measurement.
By contrast, where there are large numbers of people to cope
with, screening procedures such as we have envisaged being used
in a refugee camp must usually be employed. Then, individuals
have to be selected for treatment on the basis of only one or two
routinely applied tests. In these circumstances, we know that
many of those selected will not need treatment, and some urgent
cases will be missed. Thus the problem is to design the test so that
it selects as well as possible within its limitations.
Yet another purpose which is sometimes served by recording
children's body weights and heights is monitoring and surveil-
lance. The aim here is to identify social groups at nutritional risk,
or to check whether there is a trend for nutritional status in a
population to improve or worsen over time. In surveys of this

Defining malnutrition 45
kind, we again look at the number of individuals whose weights
or heights fall below a stated percentage of some reference level.
However, we are not now worried by 'false positives' included in
the numbers we count, provided that nutritional position of the
two is clearly and fairly shown. Thus we may decide to record the
proportion of children who are below, say, 85 per cent reference
weight-for-age, comparing farmers' children with children in the
families of labourers, plantation workers, artisans, and so on.
The choice of the 85 per cent reference level would not
necessarily reflect any nutritional benchmark, nor would it be
determined by the resources available to support some kind of
action. It would be chosen because in some social groups, many
children are below this level, while in others, most children are
above it. If we counted the number of children whose weight
relative to age was below 110 per cent of reference, contrasts
would disappear because nearly everybody would be in this
These examples of different ways in which anthropometric
data could be interpreted and used reflect the assumption
associated with the 'adaptability view' that an individual's
nutritional condition cannot be directly measured. Data on
weight, height, arm circumference and so on are therefore
indirect or proxy descriptions whose relationship with 'risk of
dysfunction' is a complex one. In particular, measurements
above which risk is uniformly low, but below which the adaptive
capacity of the body has been exceeded.
The 'genetic potential view', by contrast, stresses maximisa-
tion of potential for growth, which means that measurements of
height and weight are identified much more closely with the
individual's nutritional condition. So after allowing for the
proportion of the population whose small body sizes are
genetically determined, any deficiency in height and weight is
regarded as evidence of malnutrition. This leads to more people
being counted as malnourished than under most of the other
procedures discussed in this chapter.

46 Limits to Measurement
Some policy implications
These differences in the definition and classification of malnutri-
tion are carried over, not surprisingly, into policy recommenda-
tions. One approach, appropriate to the conditions in many
low-income countries where death rates are high and resources
are scarce, is to concentrate on those conditions in which lives are
threatened most immediately and directly.
A much more comprehensive approach is sometimes encoun-
tered, however, especially among those who take the 'genetic
potential' view. One early exponent was John Boyd-Orr, whose
concept of adequate nutrition was 'a state of well-being such that
no improvement can be effected by a change in diet' (Boyd-Orr,
1936, p.12). Interpreting 'well-being' in terms of a fairly ample
and varied food intake, Boyd-Orr made surveys of how much
food people in Britain could afford to buy during the economic
depression of the early 1930s and estimated that about half the
population was below this adequate intake more than 20
million people (Boyd-Orr, 1936, p.49). When he became the first
Director-General of FAO, Boyd-Orr initiated its first World
Food Survey, which found that over 50 per cent of the world's
population lived in 'calorie-deficient countries' (FAO, 1946).
This was regularly interpreted as meaning that half, and by 1950,
two-thirds of mankind were malnourished (Poleman, 1981).
A recent instance of estimates derived from the view that
every individual should achieve his or her full genetic potential
for growth was a statement attributed to Dr C. Gopalan which
attracted attention in the western press (Guardian, 1982; Eco-
nomist, 1983). Of the 23 million babies born in India each year,
he said, 4 million die in childhood and 16 million experience
malnutrition to a greater or lesser extent, leaving only 3 million
to grow completely unscathed into healthy adults.
These viewpoints have validity as general social commentaries
about the persistence of levels of nutrition which ought to be
regarded as unacceptable. But there are dangers in using such
all-embracing views of malnutrition and the high estimates of the
numbers of malnourished people to which they lead as a
sufficient basis for identifying policy. The kinds of policies
proposed are likely to be general and non-discriminatory in

Defining malnutrition 47
character and their effectiveness will be judged on the basis of
their impact on the average levels of nutrition of the population.
The total benefit may be large, but it will be widely spread, so
that even those whose needs are slight will experience some. In
addition, of course greatly increased total food production is
usually seen as a precondition. The difficulty with this kind of
policy, however, is that most benefit is usually felt in the 'middle
ranks' of society, and the bottom 5 or 10 per cent may be little
affected, and are unlikely to recover the priority that their
greater degree of nutritional deprivation should entitle them to.
In addition, there is a very real likelihood that the problem of
malnutrition comes to be seen as so large and so demanding of
resources for increased consumption and welfare as to be
impossible of solution within any reasonable time frame without
sacrificing other development objectives. To plan welfare mea-
sures effectively requires either a more discriminating view of
which groups within the population are most in need or the
acceptance of the very high costs of interventions which do not
A different policy approach is suggested by saying that instead
of seeking 'the greatest happiness for the greatest number', one
should demand, more modestly, 'the least amount of avoidable
suffering for all'. Still other policies would result from suggesting
that where some suffering cannot be avoided, as with hunger in
times of food shortage, that suffering should be distributed 'as
equally as possible'. Karl Popper (1945), whose point this is, goes
on to argue that it aids clarity 'if we formulate our demands
negatively, i.e. if we demand the elimination of suffering rather
than the promotion of happiness' or health.
The change in emphasis in political action which the minimisa-
tion of and equalisation of unavoidable suffering would imply is
important. This is because the nature and intensity of nutritional
deprivation is highly differentiated as between different groups
of people and between various causes of dietary inadequacy.
Action either to minimise or to equalise suffering might there-
fore start by identifying the worst cases. This might then lead to
the implementation of measures that are specifically designed to
change their circumstances. The point can be illustrated in

48 Limits to Measurement

Table 2.2 Malnutrition in children aged under 5 measured in Rwanda
and Kenya

Severe malnutrition Moderate malnutrition
(below 60% of (between 60% and 75%
reference weight-for- reference weight-for-age,
age, or 70% weight- or between 70% and
for-height) 80% weight-for-height)

Rwanda, % of children
under 5 9.8 44.9
Kenya, % of children
under 5 1.0 25.0

Source: FAO 1977.

principle, though not in historical fact, by data from anthro-
pometric surveys quoted by FAO (1977), and particularly by
figures from two African countries (Table 2.2).
In Rwanda, it would clearly be important to plan policy
interventions around the 10 per cent of children who are severely
malnourished. In Kenya, policy could be based on a broader
definition of malnutrition we could take 75 per cent weight-for-
age as a cut-off point without having a problem which is
unmanageable in a given political economy but we would still
need to investigate the 1 per cent of severely malnourished
children to check whether they represent concentrations of
poverty for which resources for special measures might be
As conditions in a country improve, and if pockets of severe
deprivation can be eliminated, cut-off points for defining mal-
nutrition can also be raised. In Britain in the 1930s, Boyd-Orr
was right to use a high standard because many people in the
nation were already well nourished, and as he pointed out,
resources were available which could raise everybody to that
level. But to apply the same standards in modern India is to
present planners with a task too large for the resources at their
disposal; more seriously, it hides the nutritionally urgent prob-
lems in the immensity of the less nutritionally urgent.

Defining malnutrition 49
Scientific models
In discussing the interpretation of anthropometric data, Trow-
bridge (1979) recalls the story of the blind men who encounter an
elephant and have to identify it from the shapes they can feel.
Each man touches a different part of the elephant and comes to a
different conclusion about the nature of the beast. Similarly,
weight-for-age, weight-for-height and arm circumference each
touch upon 'different aspects of the vague entity called malnutri-
tion'; none gives us a view of the whole condition.
In the next chapter, we explore the adaptability view of
malnutrition further by examining the scientific model of food
requirements with which it is associated. The elephant analogy is
again worth bearing in mind because it is a feature of all scientific
models that they represent the phenomena being studied in a
selective and deliberately simplified way. Only small parts of any
elephant are adequately portrayed. In the physical sciences, the
simplifications are obvious when investigators talk about 'ideal
gases' or 'weightless bodies', or when they ignore friction. Useful
conclusions can be drawn from such models, but when these
conclusions are applied to the actual world, care and realism are
The scientific view of malnutrition with which we are con-
cerned simplifies and selects from the overall problem area by
taking 'failure of function' as its basic criterion. Partly this is
because physiological functions such as growth, or resistance to
disease can be measured much more satisfactorily than, say,
hunger: but also because we are making a judgement about the
seriousness to individuals of the consequences of physiological
break-down as compared to hunger, or minor forms of ill health.
We must not forget that those value judgements have been
made, and are implicit in every application of this (or any other)
Some comparable issues arise in connection with the so-called
'human capital' model of welfare policy discussed in Chapter 4.
This represents a way of thinking about education and health
services as 'investments' in human productive power, and as
capable of being evaluated like other investments in cost-benefit
terms. Such ideas may be of academic interest to economists, but

50 Limits to Measurement
could lead to very perverse decisions if applied without caution
directly to policy. The same goes for biological concepts which
regard the human body as a self-regulating machine. Either view
might imply, for example, that welfare should be provided at a
level which supports the maximum ratio of labour output to cost,
regardless of whether the people who form the labour 'resource'
experience hunger and distress. The models do not necessarily
lead to such conclusions. They can be interpreted this way only if
we forget that they are highly simplified views and do not
represent the whole human condition.
It must be stressed, therefore, that the adaptability view of
malnutrition and related scientific models are not intended to
discount as unimportant the hunger and ill-health apparently
associated with inadequate food intake. What they may do,
however, is to encourage us to recognize circumstances where
providing more and better food does not solve every problem. In
many instances, for example, it will make more sense to improve
water supplies or housing, to develop immunisation programmes
or find ways of enabling mothers to devote more time to child
care. More fundamentally, since the underlying reason why
people have to put up with inadequate food and bad health is
poverty, to over-state the case concerning malnutrition may
distract governments from the politically uncomfortable task of
redistributing wealth.

3 Energy and protein

Food energy estimates
Two kinds of statement are commonly made about the amount of
food which people need. Firstly, figures are quoted for quantities
of nutrients which could safely be recommended for practically
all individuals, even though they may be living under a wide
variety of situations. When energy is referred to, these figures
are based on observed average intakes or expenditures in
apparently healthy populations. Figures for other nutrients are
for intakes which are adjusted according to the statistical
distribution of individual measurements; they are set at two
standard deviations above the average. These 'recommended
intakes', or 'safe dietary allowances' are thus based on over-
providing for the vast majority of people in order to ensure that
everybody gets enough. Such a policy would certainly enable
each individual to achieve his or her genetic potential for growth,
and might be a sensible basis for planning diets for institutions.
However, such figures are of little help in assessing the likelihood
of malnutrition (Longhurst and Payne, 1979).
Secondly, however, estimates are sometimes quoted for
minimum physiological requirements, that is, levels of intake
below which there is an increasing probability that some

*This chapter is based mainly on material prepared by Philip Payne. See Payne,
1976, 1982.

52 Limits to Measurement

specified symptom of deficiency will appear. The symptoms
concerned are the failures of body function previously men-
tioned, such as growth in children, reproduction and work in
adults, and failure to resist infection. Use of these criteria to
define minimum requirements is, of course, consistent with the
adaptability view of malnutrition discussed in the previous
chapter, except that when figures are quoted, the extent to which
requirements may vary with adaptation to different environ-
ments is not usually stated.
Most research has been done on western populations in which,
for example, the pressure to follow a seasonal work pattern or to
continue to work during pregnancy or lactation is relatively
slight. Thus while we know something about levels of energy
intake which on average are adequate for these populations, we
know very little about populations which depend heavily on the
seasonal use of stored energy and nutrients, or about populations
adapted to a higher level of efficiency through small body size or
in other ways.
Overall energy expenditures in the body depend on levels of
activity, growth, body composition, diet and metabolic charac-
teristics. All these are subject to adaptation, but the underlying
rate of metabolism necessary for the maintenance of the func-
tions of a resting individual is thought to be more consistent and
predictable than most other factors. This is measured as the basal
metabolic rate (BMR), and is found to vary with body weight in a
fairly predictable way (Kleiber, 1961).
Because of this assumed regularity, and because BMR is
clearly of fundamental importance, it was used by an FAO/
WHO committee in 1971 as the starting point for their estimates
of minimal energy requirements (FAO/WHO, 1973). However,
the values of basal metabolic rate accepted by the committee
referred to studies made on American subjects by Talbot (1938),
and the measurements relate to resting but wakeful individuals
who were comfortable and warm after an overnight fast of 12-15
hours. Metabolic rates below these accepted figures for the
resting rate may in fact be recorded during sleep (Benedict,
1915); they may also be recorded in subjects who have adapted to
a low food intake over a period of time or who are undisputably

Energy and protein requirements 53

Table 3.1 The effect of undernutrition on BMR; mean values reported
from various studies

Reference and number Days of Mean daily intake Percent Percent
of subjects fast during fast change change
MJ kcal in BMR in BMR
per per unit
person body

Benedict (1919)
studying 24 subjects 120 5.85 1400 19 12
Beattie and Herbert
(1947), 11 subjects >100 <7.3 <1750 26 14
Keys et al. (1950),
with 33 subjects 200 6.55 1570 39 19
Grande (1964)
1 with 13 subjects 20 4.2 1000 17 9
2 with 12 subjects 14 4.2 1000 21 14

Observations of people undertaking partial fasts over long
periods suggest that basal metabolic rates in underfed people
may be anything between 10 and 30 per cent below the figures
regarded as normal (Table 3.1; compare Kleitman, 1926). These
observations would provide a rational explanation for the
frequent reports that populations exist in developing countries
which habitually consume extremely low food intakes without
catastrophic results (Miller and Rivers, 1972; Norgan et al.,
It has been assumed that we can arrive at a figure for the food
energy intake required to maintain body weight by working from
BMR as a basis: hence the importance of measuring it. There are
problems about doing so, however, of which the most serious is
that we do not know the efficiency with which energy in food is
converted and used by the body. Direct measurements of this
efficiency in man are scarce. The FAO/WHO expert committee
approached the problem by an ingenious comparative approach.
As we have noted, BMR has a definite relationship with body
weight that holds for most mammals as well as man. The
FAO/WHO committee collated the evidence on food energy
intakes at zero energy or zero nitrogen balance in experiments on

54 Limits to Measurement

animals and man; they showed that this too was related to body
weight, and that it was consistently about 1.5 times greater than
BMR. Thus they concluded that the minimal food energy cost of
weight maintenance is approximately 1.5 x BMR.
This result could be used to estimate a minimum energy intake
required to sustain life, and indeed was so used by FAO in its
Fourth World Food Survey. FAO needed to take account of
possible adaptations in body maintenance requirements and
physiological efficiency which seem to allow some people to
survive on less than 1.5 x BMR. Faced with this problem, FAO
took the lower limit of normal variation of BMR (80 per cent of
the mean value) and regarded it as a lower limit of adaptation.
Hence they adopted a formula for the minimal energy cost of
maintenance as 1.5 x 0.8 x BMR (= 1.2 BMR). FAO then
accepted this value as a 'critical limit' for maintenance require-
ments which could be used in defining undernutrition.
The conventional method of analysing food energy require-
ments, according to Wood and Capstick (1928), partitions the
energy expenditure of an animal into three separate com-
ponents: maintenance, growth, and activity. These three
components are assumed not to interact, and maintenance, which
we have already considered, often represents the biggest part of
the total energy demand. With regard to the energy cost of
growth, the FAO/WHO committee estimated a requirement of
21KJ (5 kcal) above maintenance for every gramme of tissue
gained. (This is in fact a high figure compared with some
recorded values (e.g. Ashworth, 1969), but given the errors that
must be involved in other parts of the food requirements
estimate, accuracy with regard to this component matters little.)
The third component of energy expenditure is activity. Here,
of course, the output of people doing manual work can be
measured, and a typical figure of 0.735 MJ (175 kcal) per day for
adults employed in agriculture is sometimes used as a basis for
estimates (Bayliss-Smith, 1981). Bearing in mind the seasonality
of agriculture, we may expect average work rates to be less than
this. Peak demands may partly be met by drawing on body
energy stores, so how much extra food is required to support his
labour will depend on the efficiency of body storage as well as on

Energy and protein requirements 55

the conversion efficiency of the muscles. Where people have
adapted to a low food intake, that could partly be by increased
efficiency, but probably also depends on a reduction in unneces-
sary exertion. There is evidence for both types of adaptation.
Keys et al. (1950) reported that voluntary activity decreased
during experimental semi-starvation, but also said that the
energy cost of specified tasks was the same per unit of body
weight throughout. This differs from Seckler's (1981) finding that
overall work efficiency is inversely related to body weight over a
limited range. But Keys agrees that the total energy cost of work
tends to fall because body weight is reduced. Some physiologists
have claimed that 'work capacity' is also impaired by undernutri-
tion, but this is difficult to substantiate; in practice, the maximum
rate of work a person can sustain is rarely important in limiting
his or her effectiveness as a producer. Actual work rates are
more often limited by the nature of the work itself, and by the
social factors determining its organisation.
The outcome of all this is to leave maintenance as the
dominant aspect of food energy requirements, especially for
people surviving on very low intakes. Thus FAO equated their
estimate of minimum maintenance requirement with the mini-
mum energy intake required to sustain life. So we end up with
two different kinds of estimate for human energy requirements,
one an observed average and the other an estimated minimum,
the latter being intended to represent a 'critical limit' of
adaptation to low food intakes.
To illustrate the difference between these two kinds of
requirement, we may consider the energy needs of a 'reference
man', aged 25 to 40 and weighing 65 kg. The figures are as
1. recommended intake (FAO/WHO, 1973), equivalent to a
predicted average intake for a well-fed population: 12.5 MJ/day
(3000 kcal/day);
2. minimum physiological needs, corresponding to a critical
limit of adaptation taken as 1.2 BMR (FAO, 1977): 7.5 MJ/day
(1800 kcal/day).

56 Limits to Measurement
The corresponding figures for a 55 kg woman of the same age
who is neither pregnant nor lactating are 9.2 and 6.3 MJ/day
(2200 and 1500 kcal/day).
Neither figure, of course, can be regarded as a correct
statement of the requirements of any specific group of real
human beings. The higher figure is now generally thought to
o% er-estimate requirements in most populations. As to the lower
figure, this refers to people who are likely to be experiencing
hunger, often painfully and for long periods, but whose essential
body functions are not impaired. We have noted some of the
arbitrary assumptions in its estimation, and have pointed out that
to quote a single figure ignores the variations to be expected in
different environments. Lipton (1982) stresses that a lower limit
of energy\ intake will not only vary between individuals, gene
pools, age-groups, and so on, but that it also depends on the
choices people make (or are forced to make) about maintaining
work output when food is short. We may well end up by
questioning whether it can ever make sense to state requirements
in this way, rather than specifying ranges, or stating them relative
to local conditions.
It is worth recalling the three apparently healthy populations
with sharply contrasting lifestyles which were referred to in
Tables 1.1 and 1.2. Body weights in these populations were
mostly less than for the reference man and woman quoted above,
but the range of average intakes for the groups represented is
from 8.12 to 14.4 MJ/day for men, and from 5.94 to 12.1 MJ/day
for women, varying with season.

Adaptations and responses
The foregoing discussion left many questions unanswered about
the energy required for work in populations surviving on very
low intakes. Two factors need to be considered: long-term
adaptations, particularly in body size, and short-term responses
to temporary (or seasonal) food shortages. The latter include
physiological responses whereby the conversion efficiency of
energy in the body can rise when intakes fall. This may happen,

Energy and protein requirements 57
for example, if food energy is used directly as glucose instead of
being converted to fat and released at some later time, and other
variations in the pathways for energy use may take place. But
behavioral responses to reduced intake are important also, and
Lipton (1982, p.36) pointedly asks: 'Suppose a body can, when
daily intake falls 10 per cent, maintain work levels and cut
requirements via higher conversion efficiency. To what extent
will different groups of people choose not to reduce work...so
forcing their bodies to "raise conversion efficiency"?' Nothing is
known about such matters.
With regard to long-term adaptation, one necessary adjust-
ment in conventional thinking concerns the status of individuals
who are small and under-weight by international standards, but
who show no signs of functional impairment. The nutritional
status of such people has often been described as 'mild malnutri-
tion' when populations are screened by height or by weight in
relation to age, but it might be better to regard them as 'small but
healthy' where no contrary evidence exists. The latter interpreta-
tion may seem especially relevant in the light of what has already
been said about adaptations to a wide range of seasonal and work
environments. Before people are classed as malnourished,
therefore, and before their food intakes are deemed insufficient,
we ought to examine the environments from which such people
have come, and consider the question of adaptation.
For example, traditional diets in Japan have produced many
small but undoubtedly fit people. With current trends to more
westernised eating habits, many individuals are growing taller
and heavier, but there is concern that they may be more
vulnerable to the degenerative diseases of later life, and that
stamina for work may be less. Certainly there are no objective
data to support the idea that western diets and western anthro-
pometric standards represent an optimum.
Under conditions of food scarcity, in particular, there may be
distinct advantages in small body size, since nutrient require-
ments are related to body weight. However, Gopalan (1983) and
others have advanced a strong argument against this position.
Since large people have more muscle tissue than small people,
their work capacity is greater and therefore they can earn more in

58 Limits to Measurement
manual labour than small people. A commonly cited study in
support of this position is that of Satyanarayana et al. (1977)
showing that among 47 workers engaged in assembling fuses into
bundles on a piecework wage basis, the heavier workers
consistently produced more than the lighter workers even in this
moderate level of work activity.
Seckler (1981) argues that we should not confuse total
production with efficient use of energy in production. Although
the heavier workers produced more, they may have done so less
efficiently. Two contrary points should be noted, however.
Firstly, this job was paid as piecework, which may have led to
untypical results. Secondly, it was light industrial work for which
the energy requirements of people when they were on the job
would be dominated by body maintenance needs rather than
muscular energy expenditures. In these latter conditions, it is
certainly likely that small people would be more efficient.
In heavy work, there is no evidence of any particular body size
favouring high efficiency; muscular energy conversion efficiency
is now thought to be the most important rate limiting step and
this is not likely to be affected by body size. In studies of road
workers made by Tandon et al. (1975), the bigger, heavier men
certainly tended to produce a greater work output. But in a
comparison of those with the highest outputs and those with the
lowest, food energy intakes were found to be almost the same.
When local workers (average body weight 45.5 kg) were
compared with group of more practised road workers from
outside the area (average weight 50.4 kg), it was found that the
latter did more work with a lower food energy intake. Their
efficiency in this respect was about 50 per cent greater, apparent-
ly due mainly to their greater experience and dexterity.
With regard to heavy manual work, therefore, small people do
not seem to have any advantage, but it does seem possible that
populations of 'small but healthy' people represent other kinds
of adaptation with respect to work and survival; especially
because for them, maintenance requirements are less. What we
need to notice, however, is that small body size is an adaptation
which is made during growth, sometimes at the cost of consider-
able deprivation. By contrast, we have also talked about

Energy and protein requirements 59
short-term modifications of behaviour and metabolism related to
seasonal conditions, describing these as 'responses' to low food
intakes. They are sustained for a few months at most, and where
they entail the depletion of body stores of fat or nutrients, they
are clearly not sustainable in the long term.
This distinction between sustained adaptation and short-term
response is important if 'critical limits of adaptation' are specified
for use in assessing the extent of undernutrition in a population.
In most circumstances, we will be enquiring whether people's
capacity for short-term response is overstretched, and we will be
assuming that any long-term adaptations are fixed. But such
adaptations must differ markedly from one region and one social
group to another, and the limits of response within that
long-term framework must presumably also differ. Thus even if
the critical limit of 1.2 BMR is a reasonable best estimate of the
lowest level to which existing social groups will be found to have
adapted, it should not be regarded as a limit below which no
further adaptation is possible.
Some of these ideas challenge the conventional wisdom about
the variability of human food requirements. They conflict with
the belief that each individual in a population has his or her own
specific requirement level for energy or protein, related to
genetic potential for growth and health, and in some way fixed
for that individual. People are acknowledged to differ slightly,
but in a way that is assumed to conform with a normal Gaussian
distribution. Standard deviation is then a clear and unambiguous
expression of the extent of variations.
By contrast, the arguments put forward in this chapter imply
that individuals each have a range of possible responses
appropriate to a range of food intakes, and are not characterized
by any fixed requirement. Indeed, Sukhatme has argued that the
very notion of a fixed requirement is in conflict with most of the
evidence collected over the last forty years. The data are much
more satisfactorily understood in terms of a self-regulating,
homeostatic model (Sukhatme and Margen, 1978).
This is a complex idea, but a simple analogy may be useful in
showing what it involves. Suppose a leaking bucket is being filled
at a pump. It is possible to adjust the flow into the bucket so that

60 Limits to Measurement
it exactly matches the rate of leakage. Then equilibrium is
maintained, with a steady water level. We can alter the rate of
inflow so that equilibrium is reached with the bucket almost full,
or with the bucket nearly empty. In the latter case, the rate of
leakage is less because of lower pressure, and perhaps because
some leaks are above the water line; thus a lower rate of inflow is
necessary. Clearly, there is a range of flows over which many
different equilibrium states are possible. There is also a critical
lower limit of inflow below which the bucket empties, and an
upper limit, beyond which it overflows. To some extent, at least,
the responses of the human body to variations in food intake are
One implication is that when we observe a range of different
food intakes in an apparently healthy population, this not only
reflects variability between individuals, but more fundamentally,
it reflects a range of possible equilibrium states for each of those
people. Thus standard deviation takes on a different signi-
ficance, since it is now seen to represent the potential for
response to changes, rather than random statistical differences.
A food intake that is below average by a matter of two standard
deviations conventionally represents an intake which is lower
than the requirements of all but 2.5 per cent of the population;
according to Sukhatme however it could just as reasonably be
taken as an estimate of the lower limit of possible adapted
responses in that population.
Applied to food energy intakes, this estimate gives a rather
low figure for minimum physiological requirements, but not as
low as that obtained from FAO's (1977) assumption that the
lower limit of adaptation for many populations might be around
1.2 BMR. For example, working on two standard deviations
below the average, Sukhatme (1961) calculated the minimum for
a 55 kg man as 8.8 MJ or 2100 kcal, which is 22 per cent below the
FAO/WHO recommended intake, whereas using 1.2 x BMR
gives a value of 40 per cent below.
This example of extremes reinforces the warnings about
scientific models that were sounded at the end of Chapter 2, since
in this instance, one's choice of model affects conclusions very
substantially. In fact little has been done in a systematic way to

Energy and protein requirements 61
test the various models. A major difficulty is that knowledge of
adult human requirements is largely based on experiments in
which the objective has been to maintain equilibrium. The
argument applies to protein as well as to energy intakes, and
experiments have been done on apparently healthy people in
which levels of intake are found which are just sufficient to
maintain the existing level of body nitrogen. However, the
assumption then made that nitrogen balance indicates adequate
intake is not valid. It is, rather, just an indicator that the subject's
equilibrium state is not changing; the significance for health of
the changes in equilibrium state that do commonly occur is not
Even if the adaptability model comes to be accepted as being
closer to reality than that currently advocated by FAO and WHO
committees, it will still not provide us with a completely general
basis for prediction because of other factors which influence
variability. Measurements made at a single point in time on
individuals of the same age and sex will include at least the
following sources of variance:
1. differences in adaptation, including those reflecting the
primary expression of genetic factors, and those arising from the
effects of diet and environment during growth;
2. differences due to the extent and nature of continuous
auto-regulatory responses;
3. a 'noise' component due to random, unrelated fluctuation in
the environment, or diet, or due to measurement errors.
If we were to conduct experiments in order to establish the safe
limits of adaptive response to different intakes of some nutrients,
we would need to distinguish between these sources of variabil-
ity. Assuming that in general we would find both upper and lower
limits to the range, with some risk of dysfunction attached to high
levels of intake as well as low levels, the situation would be as
shown in the diagram (Figure 3.1). Instead of a single figure for
the recommended intake or safe level for the particular nutrient,
we would have a 'safe range', with upper and lower limits
bounded by points for which the probability of dysfunction rises
above an acceptable level.

62 Limits to Measurement
'Adaptive' models of body weight control
The concept of an adaptive, self-regulating system implies some
type of negative feedback response by which the body is able in
part to cancel out changes in inputs and outputs, and maintain
the internal state near its original condition. There will be several
such feedback loops, and in the particular case of energy, one
even extends outside the body, so that an individual may adapt
by modifying his interaction with the environment, and hence
change his requirements for energy to bring them into line with

2 '\ /
. \ /

Intake --

Range of
variation of
limits of
adaptive response

Figure 3.1 Schematic view of the range of adaptive responses likely to
be encountered with different levels of nutrient intake

Energy and protein requirements 63

intake. Thus the total strategy of adaptive response for energy
comprises individual and social changes in behaviour as well as
changes in body weight, body composition and metabolic
regulation. It may well be that individuals differ in the relative
balance of these components, and that for some, changes in
behaviour will be a more important adaptive mechanism than for
On a physiological level, it is clear that individuals differ very
markedly with regard to energy storage within the body. For the
majority, energy is chiefly stored in the form of fat, but for a few
people, significant quantities of energy are also stored as
protein. We can represent these differences by defining a ratio
which represents the proportion of energy stored as protein in
the body. The same ratio predicts the energy mobilized from
protein when energy intake is less than expenditure. The
'P-value' (Payne and Dugdale, 1977) expresses this component
of energy as a ratio of the total energy stored in the form of fat
and lean tissues in the body (or the total energy mobilised when
intakes fall). Figure 3.2 shows the proportions of energy
mobilised from protein in semi-starving men as observed by Keys
et al. (1950), and it is evident that these men showed a wide
variation in the way they responded to a negative energy
balance. On average, 15 per cent of energy was lost from lean
tissue stores and 85 per cent from fat. But the average actually
spans a tenfold range: some individuals lost only 3 per cent of
energy from protein, others as much as 30 per cent.
We do not know the extent to which the propensity to store
and utilise protein as opposed to fat is a result of genetic
differences, and how much is a result of patterns established
during early growth. We do know that under some circum-
stances, a tendency to use fat as an energy store might be a good
survival strategy, whereas in other circumstances it could lead to
obesity. The nature of responses to increased energy intakes may
be similarly varied. The data of Miller and Mumford (1967)
showed considerable variations between individuals in respect of
degree of weight gain during overfeeding. This may partly be
because the metabolic efficiency of energy storage varies from
one person to another and partly because some people are less

64 Limits to Measurement


o \

04- -

00oo 0-- 0 co

d d d d do d o
0 0 tO 0 0 O 0 0
0 0 ,.- .- CM N CO (0

P value

p= energy stored as, or mobilised from protein
total energy stored in the body, or mobilised/

Figure 3.2 Differences between individual men with respect to the
storage of energy as fat or as protein
Notes: P-value would be zero for an individual whose body store of
energy consisted solely of fat, and 1.0 for an individual storing energy
solely as protein.
Source: Dugdale and Pa. ne 19"", based on data from the experiments of Keys
et al. 1950.

Energy and protein requirements 65

efficient when they over eat. It also seems likely that individuals
with a propensity for fat storage will gain more weight before
increased metabolism balances out the higher intake than will
individuals who store metabolically active lean tissue.

Protein requirements
As we know so little about the complex strategy of response and
adaptation, and have no reliable indicators to predict how a
given individual will respond to changes in intake, either by body
weight, composition, or behaviour, the definition of energy
requirements is an uncertain business. In respect of protein, we
are hardly better off. Obsession with nitrogen balance as a
'criterion' of adequacy has led to a neglect of the study of
adaptive changes and mechanisms. Nonetheless, it is well
established that such mechanisms exist. Waterlow (1981) has
reviewed the evidence for adaptive responses in nitrogen meta-
bolism, and concluded that the body possesses powerful mechan-
isms for economising on nitrogen.
For example, Durkin et al. (1981) showed that in most cases,
responses to reduced levels of nitrogen intake included the loss
of body weight, Nicol and Phillips (1976) showed that Nigerian
farmers, habituated to diets relatively low in protein, utilised
that protein more efficiently than American subjects, and could
maintain balance on about 30 per cent less. It appears that the
Nigerians were adapted to low intake.
Considering that little is known about such adaptations, it is
not surprising that efforts to define human protein requirements
have had a chequered history. 'Kwashiorkor' was described by
Williams (1933) as an illness, 'in which some amino acid or
protein deficiency cannot be excluded'. The children who
developed the disease did so after weaning, when they were fed
on starchy porridge rather low in protein, energy and other
nutrients, and they were cured by milk. In time, the idea that this
syndrome was specifically caused by protein deficiency became
generally established.

66 Limits to Measurement

Subsequent work, however, challenged the definition of
kwashiorkor as protein deficiency, and has led to a revision of
ideas about how much protein young children need. Lower
values for their protein requirements have slowly gained accept-
ance, but energy needs are still estimated to be at around the
same level (Table 3.2). No set of symptoms or signs is now
generally agreed to be evidence of a specific deficiency of protein
in man. This is, of course, not to say that there is no malnutrition
in children, nor is it to deny that there is such a thing as
kwashiorkor. It is simply that neither clinical studies nor
attempts to reproduce the condition in animals provide sufficient
justification for identifying the well-known symptoms oedemaa,
hair and skin changes, fatty liver, blood changes) as specific for
protein or amino acid deficiency (Waterlow and Payne, 1975).
The results of food consumption measurements made on
individual children in a variety of societies seem to confirm this
conclusion (Waterlow and Rutishauser, 1974). Where intakes of
protein are low relative to estimates of requirements, so also are
intakes of energy, and often all other essential nutrients as well.
But these deficits of nutrients are often such as would be made

Table 3.2 Estimates of protein and energy requirements at I year of
age, corresponding to 'recommended intakes'

Year Per kg body weight Ratio of
Protein' Energy protein energy Source
to total energy
g MJ kcal %

1948 3.3 0.42 100 13.2 NAS (1948)2
1957 2.0 0.42 100 8.0 FAO (1957)
1963 2.5 0.42 100 10.0 NAS (1963)2
1965 1.1 0.42 100 4.4 FAO/WHO (1965)
1968 1.8 0.42 100 7.2 NAS (1968)2
1969 1.3 0.47 110 4.7 DHSS (1969)3
1973 1.25 0.44 105 4.8 FAO/WHO (1973)
1974 1.35 0.42 100 5.4 NAS (1974)3

' As protein which is 100% utilised.
2 US source; protein estimates consistently higher than in other sources.
3 UK source.

Energy and protein requirements 67
good automatically if energy needs were satisifed by taking more
of the same food. Thus the most significant factor causing child
malnutrition would seem to be simply inadequate food intake.
Because of their rapid growth rates, infants need a somewhat
higher proportion of protein in their diets than is required by
older children and young adults. But we need to remember that
protein plays a highly complex role in physiology. It is at once a
nutrient in its own right, and a source of energy. When protein
levels in diets are high, a large proportion of the amino acids
derived from protein is deaminated and oxidised as fuel. This
also occurs in situations of overall energy deficit. Thus a child
given an adequate protein diet, but at intake levels below his or
her energy needs, will not derive full benefit for growth from the
protein, but will use it as an energy source.
A further complication is that proteins do not all have the same
growth-promoting value. Depending on the balance of their
component amino acids, some have higher 'quality' than others.
Protein quality indexes (such as 'net protein utilisation', or
'chemical score') are estimates of the percentage of the total
protein which can be expected to be utilised for growth and
maintenance in a young animal or a child. On this basis, eggs and
milk have efficiencies close to 100 per cent. A food or diet with a
quality of 70 per cent will contribute more of its amino acids to
growth than one of 40 per cent. However, the amino acids which
are not used for protein metabolism will not be wasted: rather,
they will be diverted for use as an energy source.
Because protein can provide the body with energy in this way,
at roughly 17 KJ or 4 kcal from each gramme converted, it is
useful and usual to express the protein value of diets in relation to
their energy content. The relevant ratio, expressed as a percen-
tage, is given by one of the following formulae, depending on the
units employed:

protein-energy ratio = grammess protein in diet x 0.017 x 100)
total megajoules in diet

= grammess protein in diet x 4 x 100)
total kilocalories in diet

68 Limits to Measurement

A further refinement is possible by correcting the energy value of
the proteins by a factor which takes quality into account. The result
is a ratio of energy derived from fully utilisable protein: the 'net
dietary protein-energy ratio' (NDpE%) (Payne, 1972).
Animals are often used to assess the nutritional values of different
foods, and for example to compare different strains and varieties of
cereal grains as protein sources. Thus the superior growth of rats
and pigs on high lysine varieties of maize and rice have often been
cited as evidence of their advantages for humans. However, in
interpreting such feeding trials it is important to note the major
differences that exist in the nutrient needs of different species.
Most young mammals grow very much faster than human infants
when growth is measured as \eight gain per unit body weight, and
need a larger proportion of protein in the diets to support this
growth. Figure 3.3 illustrates the point diagramatically. If feeding
experiments were carried out with an adult man, a one-year-old
child, a new-born baby, and a young pig, using a diet containing 10
per cent of a protein, one of whose amino acids could be varied
from zero up to 100 per cent (with perfect balance of all other amino

Pig achieves
Positive growth
Baby achieves
maximum growth

One-year-old achieves
maximum growth
Adult man maintains
body weight

Weight Adult man
loss loses weight

0 20 40 60 80 100
Protein quality (egg or milk taken as 100)

Figure 3.3 Feeding experiments based on a diet containing 10 per
cent protein of varying quality

Energy and protein requirements 69
acids), then the man would lose weight with a zero ratio, and would
show improvement as the level was raised up to about 40 per cent of
maximum, and then no further response. Similarly, the one-year-
old would continue to benefit from increases up to about 50 per
cent, the new-born baby up to 70-80 per cent, but the pig up to 100
per cent.
It should be stressed that this diagram represents the protein
requirements for maximum growth in human babies, which is not
the same thing as the optimum protein requirement. Children,
unlike pigs, are not deliberately fattened for market, so it should not
be assumed that the fastest rate of growth is necessarily the best;
indeed, we will shortly quote reasons for thinking that it is not. What
the diagram makes clear, however, is that the growth of children
does not always suffer if protein quality is less than 100. In any
practical situation, what has to be assessed is whether the quantity of
protein eaten, in conjunction with the balance of amino acids, i.e. its
quality, is such that the requirements of the consumer for the most
limiting amino acid are met. The balance of evidence from feeding
trials seems to show that although cereals are limited in protein
quality by their content of lysine, threonine and tryptophan, and
although this is as true for children as it is for pigs, provided that
sufficient cereal is consumed to meet energy and total protein
needs, then conventional varieties of most cereals contain sufficient
of these amino acids to meet the requirements even of preschool
As with energy, so with protein, therefore, the conclusion to
emerge from this chapter is that human food requirements are less
certainly known than was once thought. This ought not to be
regarded as totally discouraging, however, but rather as a chal-
lenge to regard nutrition with a more comprehensive understanding of
physiological function and environmental constraint. Part of that
challenge is to combine these insights with a study of the variability
between individuals' capacity to adapt to environmental changes
without unacceptable loss of function. Up to now, many nutrition-
ists have tended to overlook this problem area, partly because of the
difficulty of studying it experimentally, but partly also because the
models conventionally used to describe the interaction of require-
ments and nutrient intake implicitly assume that under normal

70 Limits to Measurement

conditions, adaptation is a relatively minor part of total variability.
If adaptation does occur, the assumption has usually been that the
situation must be abnormal, and failing proof to the contrary, the
adapted state is regarded as likely to confer some disadvantage.

When a nutritional requirement level is stated, a prediction is being
made on the basis of past knowledge about the probability of
specific functional disabilities affecting a particular group or
individual. Ideally, the requirement level should be presented in a
form which indicates this fact. However, it has to be admitted that
few satisfactory statements of this kind are possible, e.g. that 'there
is a probability of 0.1 0.01 that fifteen-year-old non-pregnant
females will develop night blindness if they sustain an intake of less
than 400 RE (retinal equivalents) of vitamin A per day' (needless
to say, we do not know if this is in fact true). Ideally, too, the
probability should be established within confidence limits, and
the dysfunction to which it refers should be specified and
measurable. Other kinds of dysfunction related to vitamin A
status may be known to exist, and their probabilities could also
be stated for that sample level of intake. Since many dysfunctions
have multiple causes, some statement about the context (envi-
ronment infection, etc.) might also be necessary. Where other
departures from health may only be suspected, no true require-
ment figure can or should be stated with respect to these.
Something which is specifically excluded from this definition is
the notion of an 'optimum' state of nutritional health, achievement
of which might be the criterion for a requirement level. This is an
idealistic notion which seems to be sustained by two kinds of
attitude. One is a view which reacts against minimum prescriptions,
and would prefer to see emphasis placed on the complete fulfilment
of human genetic potential. This feeling demands respect, and if
resources were available and if the concept of genetic potential
could be clarified policies based on such ideas would deserve

Energy and protein requirements 71
The other attitude is that optiinum states are somehow inherent
in 'natural' systems, and that the right nutritional conditions are
those that allow these states to be expressed. This is commonly
linked to the idea that natural selection will have maximised
'fitness', and that there should exist a state of nutrition in the
individual which will be a reflection of that fitness. However, we
have no reason to believe that there is some preferred state of
metabolism or physiology in the individual which simultaneously
brings to a maximum all qualities that we happen to consider
desirable, and so, at the other end of life, is longevity. But there is
no evolutionary selection for longevity only for survival to sexual
maturity (Kirkwood, 1977). Nor is there any evidence that children
who grow rapidly enjoy long life. Indeed, to the extent that
experiments with animals can throw light on the matter, Ross et al.
(1976) found that when young rats were allowed to select their own
diets, there was considerable individual variation in food energy
consumption and growth rate, and that high rates were negatively
correlated with lifespan. In this instance, at least, it is clear that
energy requirements for young animals could be based either on the
criterion of achieving the full genetic potential for early growth, or
on subsequent longevity, but not both.
Whether or not this evidence can be paralleled in humans, it
serves by analogy to show why any views of 'desirable' or 'optimal'
food intakes for human individuals or groups can only be value
judgements. If our aim is to be as objective as possible, we have to
be content with describing how the probability of failure to sustain
function varies with intake. Deciding what levels of probability or
risk are unacceptable, brings us back to value statements again.
What is quite clear is that where energy and protein requirements
are concerned, doing this leads to critical limits considerably below
what many people would judge to be desirable.
Yet science is not objective and value-free. The preceding
paragraph is itself a value-judgement above the limits of the
scientist's responsibilities. Other aspects of the value-system inhe-
rent in science are reflected in the simplifications which scientists
build into their models, and are to be observed also in uncritical
reverence for measurements and statistics. This is often exploited by
policy-makers who use the figures so produced both to legitimise

72 Limits to Measurement

past actions and to guide their plans. Enough has been said here to
stress that our understanding of individual adaptability with regard
to food is extremely sparse. That clearly limits our ability to make
quantitative statements about nutrient requirements. Yet over the
years, governments and international agencies have produced a
stream of reports listing minimum physiological requirements, safe
levels, and recommended dietary allowances, which are supposed
to have a firm objective basis. In subsequent chapters, we shall
describe an approach to the use of quantitative data which we
believe can be much more helpful in understanding malnutrition
and in informing action to reduce its incidence.

4 Food system indicators*

Variables in food systems
Two kinds of measurement used in monitoring food systems have
been discussed in previous pages: anthropometric assessments of
nutritional status (Chapter 2), and measurements of food intake
related to ideas about requirements (Chapter 3). In considering
food supply and utilisation over a whole nation, as well as the
nutritional status of its people, there are obviously many other
variables which might usefully be measured: food stocks and
production, for example, or people's expenditure on food. Figure
4.1 is developed from the food system diagram presented earlier
(Figure 1.1) in order to show whereabouts in the system observa-
tions are commonly made or data collected. All the variables
concerned are examples of system variables. The values they take at
any one moment reflect the state of the system at that moment.
Normally, a variety of processes and changes will be continuously
under way, and measurements of these variables specify the state of
the system in 'frozen' form, rather like photographs taken at
particular instants.
It will be noticed that these system variables are of two kinds.
Some, like body weight or food stocks, are characteristic of a

*This chapter is based on material prepared by Peter Cutler, Elizabeth Dowler,
Philip Payne, Young Ok Seo, Anne Thomson and Erica Wheeler. Fuller versions of
some of the material have been published by Dowler et al., (1982) and also by
Dowler and Seo (1983).

74 Limits to Measurement

Hygiene and
housing cond-
itions for
storing, cook-
ing, preparing
food (S)

Individual -
food intake Utilisation of food
(F) by the individual

of clinical
symptoms, e.g.
oedema, eye
(S) Nutrition/health
status of individual

Figure 4.1 The food system of a region or nation as it was defined in
Chapter 1, but now showing some of the system variables which are
commonly measured
Notes: F denotes variables which measure flows through the system; S
denotes variables which reflect conditions within the system.

Food system indicators 75
condition in the system that is liable to change fairly slowly. Others
measure flows through the system, and like the flow from a tap, they
can fluctuate widely or even be abruptly turned off. One example is
cash income, which drops to zero overnight when an individual loses
his job. Clearly, these two kinds of variable reflect on the state of the
system in different ways. Much confusion surrounding data on
energy or protein intakes arises because these latter are flows, yet
some people insist on treating the figures as if they represent an
established condition of either adequate or inadequate nutrition.
The distinction can be understood by referring again to a leaking
bucket that is being filled at a pump. Measuring the flow of water
from the pump cannot tell us how full the bucket is we have to
observe or measure that directly.
The interpretation of measurements often involves checking
them against critical levels of the variables to which they refer.
These are levels below (or above) which it is generally agreed there
is evidence of an unacceptable degree of immediate or impending
distress. It is essentially the use of cut-off points defining such levels
which distinguishes the particular variables which we choose to
define as indicators, from other system variables.
The levels at which cut-off points are set should ideally depend on
the purpose for which the indicator is being used. We have seen how
this works for anthropometric measurements, where a working
definition of a cut-off point for, say, weight-for-age, is defined to fit
particular circumstances, and may vary according to whether the
purpose is surveillance or screening. With regard to food intakes,
the critical points are the requirement figures and estimates of
adaptability limits, and which ones we use may again depend on our
purpose. If the aim is to identify social groups with the most
seriously inadequate diets, we might use an estimate for the lower
critical limit of adaptation. If, however, the aim is to buy
sufficient food for residents in an institution, we would certainly
use a higher requirement figure.
To sum up, then, an indicator is a variable such as food intake or
weight-for-age for which critical points have been agreed. A simple
example from outside our subject is provided by the gauge which
indicates steam pressure in a boiler. Here, the 'system', or set of
components, consists of boiler, fire-box and engine. Steam pressure

76 Limits to Measurement

is the system variable which we select to describe the state of that
system. In addition to measuring pressure, the steam gauge may
also have certain critical points marked on the pressure scale, below
which pressure is inadequate, or above which is a danger of
explosion. It is these critical points which make steam pressure an
indicator of the state of the boiler.
A normal characteristic of an indicator is that it is used to show a
number of different aspects of the performance of a system. In this
example, the situation is relatively simple because steam pressure is
quantitatively related in a known way to other variables such as
temperature, and strain in the walls of the boiler. Steam pressure
also tells us about the balance between steam output from the boiler
and fuel input, but it does not allow us to estimate these flows
In an instance like this, where the relationship between variables
is unambiguous and clear, we may use one variable as a 'shorthand'
indicator of another without hesitation. In biological and social
systems, however, matters are rarely so simple. We may know that,
within a particular kind of system, certain variables have usually
assumed values which bear consistent mathematical relationships
with each other, but we may not know how this happens. In such
circumstances, a variable may not have a sufficiently well defined
significance to be regarded as a shorthand indicator, but it might be
generally accepted as a proxy for a set of system variables, or even
for certain aspects of the system which cannot as yet be precisely
Nutritional indicators based on anthropometry are to some
extent shorthand indicators, when weight and size are taken as
estimates of body nutrient stores, and size at a given age is regarded
as a measure of the previous pattern of growth. But in addition,
anthropometric indicators are commonly used as proxies for
'nutritional status', an imprecise notion that, as we have seen,
covers the outcomes of a wide range of different processes,
including the effects of different nutrient deficiencies, and of
non-nutritional factors such as infection.
Defining the critical points which are associated with the use of
indicators must always involve the expression of value judgements,
and attempts to achieve consensus about acceptable levels of risk.

Food system indicators 77
With regard to the boiler, for example, the owner of the plant may
see advantages in terms of efficiency in running it at a high steam
pressure, and provided he can afford the insurance premium, he
may be happy to accept an increased risk of explosion. The men
who stoke the boiler, however, may see no advantage in the higher
pressure, but will be aware of a greater risk of injury or death. Their
trade union might seek a much more stringent standard of safety
than the management wishes to accept.
This illustrates a question which was raised in previous
chapters but has not so far been answered. When we define
malnutrition in terms of unacceptable levels of dysfunction, who
decides what is unacceptable? The scientist cannot do this alone,
despite the fact that he or she possesses some of the necessary
information. With regard to the boiler, for example, an engineer
might be called in to advise, but is in no position to judge what
insurance premium the firm can afford, nor what level of risk the
stokers (and their families) are willing to accept. Thus the
engineer has to explore with the management, or perhaps with
the trade union, what would be the consequences for them of a
higher steam pressure and a greater risk. In other words,
valuations of risk and of what is acceptable or otherwise should
arise from a process of dialogue, either explicit or indirect, in
which the scientist or technician is only one of the parties
In nutrition, as in other fields, the role of the scientist is defined
partly by his or her expertise in using a range of specialised
techniques for observation, measurement and prediction.
However, in no sense is his or her responsibility limited to this.
Besides the work of measurement there are questions of social
valuation, and whereas the former is a purely technical activity,
the latter entails dialogue in some form with other sectors of
society with planners, no doubt, with those who fund research,
and also, we might hope, with the ultimate consumers of the
food. One of the differences between measuring a variable and
using an indicator is that the former is a more or less objective
technical activity while the latter always involves the use of
critical points which express social valuations.

78 Limits to Measurement

Indicators commonly used
We have already commented on the use of anthropometric
indicators in the food system. We will now look at two other sets
of indicators commonly used: food supply and consumption, and
expenditure on food items.

Food supply and consumption
Among the more important variables that reflect on the opera-
tion of food systems are data on food supplies collected at
national level, which are commonly expressed as per capital
protein or energy availability. Such data are based on national
food balance sheets (FBS) compiled regularly for many countries
in a manner that takes account of food production, imports,
exports and increases or decreases in food stocks. Grain used for
seed or animal feed is subtracted, and so are estimated losses.
(For detailed accounts of methodology, see FAO, 1980; Pole-
man, 1981.)
Despite several decades of effort, errors in these estimates are
still known to be significant, and some authors suggest that for
low income countries especially, they tend towards the under-
statement of food available (Poleman, 1981). Figures expressed
as per caput supply introduce additional errors wherever popula-
tion data are inaccurate. Moreover, as the food system diagram
emphasises (Figure 4.1), FBS data represent flows through the
system at a point which is rather remote from the individual
eating a meal. Thus the figures do not represent human
consumption, nor do they represent food distribution over time
or over population. With good reason, the FAO cautions against
misunderstanding their basis (FAO, 1980).
The usual way of estimating actual food intakes is through
household consumption surveys (HCS, Figure 4.1). These
employ a range of methods including weighed intake surveys
(larder stocks or plate/pot consumption), diary recording and
recall interviews (Marr, 1971; Burk and Pao, 1976, 1980).
Among the difficulties involved in using this method are
problems of sampling and of seasonality in food supply or usage.

Food system indicators 79

A further problem is that precise methods of weighing meals or
ingredients tend to be intrusive, and may induce behavioral
change and untypical intakes. Variation in eating patterns from
day to day compound the problem, and household-based
measurements usually omit food eaten outside the home,
although in some instances, an estimate of food and/or nutrients
in this category is included when results are published.
Do these two sources of data on food give the same informa-
tion? Since FBS and HCS measurements of per capital food flows
are made at different points in the system, we would not expect
them to yield precisely the same results. Furthermore, FBS
estimates cover 356 days; while HCS measurements typically
cover 3-7 day periods in each household. Yet the differences
which do occur have proved difficult to explain. FBS figures are
always greater by about 1 MJ per capital in less developed
countries, and by 3-4 MJ per capital in developed regions.
These differences may arise partly because alcohol and
casually gathered fruit and nuts are not always included in the
surveys as 'food'. Such mismatching cannot account for the
discrepancies fully, however. Much emphasis is placed by some
authorities on underestimation of wastage in the home as a
source of error in HCS data in developed countries, but studies in
Britain have failed to demonstrate such underestimation (Dow-
ler, 1977; Wenlock et al., 1977, 1980).
In Japan, a nationwide survey of household food consumption
based on seven-day weighed intakes has been carried out
annually for over thirty years, and an interesting comparison can
be made between HCS and FBS figures over the whole of that
period (Seo, 1981). Figure 4.2 shows the results graphically,
plotted against per caput gross national product (GNP) on a log
scale. Average energy intake as measured by HCS remains
reasonably constant over the period considered. However, per
caput energy supply as judged by FBS figures rose rapidly as
GNP increased. Comparison with cross-sectional data from
other countries shows that not only is the FBS estimate always
higher than the household estimate, but the discrepancy in-
creases as the economy of a country, and thus its food system,
becomes more extensive and complex, with more opportunities




GNP per caput
($ US log scale)

200 300 |' 500 700'' 1,000
1951 1958 1965

2,000 4,000 6,000 10,000
1972 1974

Figure 4.2 Longitudinal comparison Japan (energy and GNP/caput: energy supply data (food balance sheets) and
energy intake data (household consumption surveys))
Source: Seo 1981.

-13 MJ

(Food balance sheets)

(FAO country requirement)

(Household consumption surveys)

2,600- 11





Food system indicators 81

for loss or waste within the system as a whole, before reaching
the household level.
Given the difficulties associated with these food supply data,
and the uncertainty in estimates of 'requirements', efforts to use
FBS figures to identify countries with an 'energy deficit' are
bound to run into difficulties. Yet official bodies continue to
describe global and national energy deficits in this way. The US
Department of Agriculture produced two World Food Budgets
relating to the 1960s (USDA, 1961, 1964) which concluded that
almost the entire population of the developing world lived in
'diet deficient' countries. The World Bank (1976), using data
from the 1960s, saw a problem of roughly the same magnitude
involving about 1200 million people, or some 60-70 per cent of
the total population of the countries represented. FAO has tried
to modify FBS to estimate intakes by imposing an apparent
distribution on the mean figure; this process leads to an
estimated total number of people affected by 'protein energy
malnutrition' of just over 400 million (FAO, 1974, 1977).
When FBS data are used to produce figures of this latter sort,
what has happened is that the system variable measured by food
balance sheets has been turned into an indicator by the applica-
tion of a cut-off point. This latter is either a minimum level of
food consumption or a recommended intake adjusted for such
characteristics of the local population as age structure or average
body weight (as in FAO's 'country requirement' figures).

Expenditure on food items
Since purchases of food intervene between the points in the food
system where FBS and HCS measurements are taken (Figure
4.1), it might seem that a further way of exploring the operation
of this part of the system might be to enquire into family food
expenditures. As usual, there are several difficulties to be
overcome. Spending on food by low income families as quoted to
survey teams is quite often greater than the reported income of
the same families; the former is over-estimated and/or the latter
is under-reported (Scott and Mathew, 1983). In addition, there is
little agreement about how data on family income should be

82 Limits to Measurement
used. Again we are turning a system variable into an indicator by
defining some particular low income as a cut-off point, which
then becomes a 'poverty line'. But as we saw with the indicators
previously discussed, unless the cut-off point is carefully defined
in relation to some clear, practical purpose, its use may be
With indicators derived from anthropometry, we have seen
cut-off points can be set at levels at which specified risks of
physiological dysfunction begin to rise steeply. An attempt to
introduce similar 'objective' criteria into the definition of a
poverty line has been made by Lipton (1982). He notes that the
poorest people in a wide range of countries commonly spend an
irreducible 20 per cent of their income on non-food essentials.
He further takes account of the tendency of food requirements
figures to over-estimate food needs, especially in relation to
'small but healthy' people living in warm climates, and with
Sukhatme's (1961) work in mind, he suggests that it might be
more realistic to work with 80 per cent of the conventional
requirements figures. This leads him to a 'double 80' rule,
according to which people are judged to be 'ultra poor' if 80 per
cent of their income is insufficient to purchase 80 per cent of
conventionally defined food requirements.
Lipton points out that people living just above this double-80
standard may be to some extent undernourished, and probably
experience hunger quite frequently. He calls them 'moderately
poor' on the grounds that the risk of physiological impairment
they face is likely to be much less than for the 'ultra-poor'. He
further estimates that 'the ultra-poor... form 10-20 per cent of
populations' in most low-income countries, and for India finds a
startling contrast between urban and rural populations. Calculat-
ing on the basis of monthly expenditure per person (MEP), he
comments that in 1972-3: '41% of India's rural households but
under 2% of urban fell into MEP groups averaging food/outlay
ratios above 80%' (Lipton, 1982, p.45).
In urban areas, more money is probably needed for non-food
items such as fuel, but this does not fully account for the
rural/urban difference.

Food system indicators 83

Social valuations
Discussion of poverty, even more than of undernutrition, brings
us back to the question of what is acceptable or otherwise to a
particular nation or social group. Among several ways in which
this question might be tackled, one is for the investigator himself
to lay down minimum expenditures on food, clothing, fuel,
housing and so on for families of different size, and to consider
that those with incomes less than this are living in poverty. This is
what Rowntree (1901) did in his pioneer work on poverty in
Britain. In using this method, the investigator makes his own
valuation of what should be regarded as unacceptable, in the
light of what he knows of social norms and physiological needs
for food and warmth.
Table 4.1 summarises a wide range of estimates concerning the
extent of poverty in India at three dates for which information
from the National Sample Survey (NSS) enables such calcula-
tions to be made. The estimates differ considerably, and have
been the subject of great and continuing controversy. However,
it is important to note that they are not all attempting to measure
the same thing. Sukhatme (1981a,b) deals specifically with
malnutrition; but Rao (1981), Dandekar (1981) and others
include undernutrition within their definitions of poverty.
The table also illustrates some of the many different energy
requirements figures that have been used in arriving at poverty
lines. These figures are for a standard 'consumer unit', that is, a
moderately active 'reference man' weighing 55 kg. Family food
needs are estimated from this by counting women and children as
fractional consumer units at various levels.
A rather different way of establishing criteria for the estima-
tion of poverty is to observe how the poor express their own
valuations through their behaviour- when purchasing food.
Boyd-Orr used this approach in his surveys of food, health and
income in the UK during the 1930s (Boyd-Orr, 1937). Lipton's
double-80 criterion is based on the empirical observation that in
several countries, including India, almost no group of poor
people allocates significantly more than 80 per cent of household
expenditure to food. The 80 per cent level seems to represent a

84 Limits to Measurement

Table 4.1 Estimates of the extent ofpoverty in India based on National
Sample Survey data on consumer expenditure (compare Kumar 1973)

energy intake
used in Cut-off level Percentage of
calculating adopted as the population
poverty lines poverty line, orto below the cut-
Authors and dates based on define under- off level
of NSSfigures used income nutrition defined
MJ kcal Rural Urban

Bardhan (1970a, b)
Dandekar and Rath
(1971); Dandekar

Ojha (1970)
Rao (1981)
Sukhatme (1981a,b)

FAO (1977)
Lipton (1982)


Income of
11.3 2700 Rs 15
11.3 2700 a

Income of
9.4 2250 Rs 15 (rural)
Rs 22.5 (urban)
9.4 2250 a
9.6 2300 a
Income of
9.4" 2250C Rs 18-21
9.4' 2250C a
Income of
9.6" 2300c Rs 21-24
Food intake
of 8.8 MJ
Food intake
of 9.6 MJ
Food intake
= 1.2 x BMR
80% of spending
on food
Breakdown of
Engels's law

a denotes that the same poverty line was used as for 1960-1, but with
adjustment for inflation; b refers to percentage below stated energy
intake, not below the income poverty line; c based on average
intake; d refers to rural and urban populations.


40 50


20b 25b

17b 21b

41 2

10-30 5-20

Food system indicators 85
critical condition below which people behave as if they have no
real options left: they cannot easily economise further on
non-food items, nor can they manage with less food. If their
income per caput falls further, forcing more cuts in spending,
food and non-food items are reduced in the same proportion.
Another way of looking at this is in terms of Engel's law. When
comparison is made between better-off and worse-off groups as
judged by income per family member, it is found that a
progressively larger percentage of family expenditure is used to
buy food as one goes down the income scale. The relationship is
stated by Engel's law to be a linear one, as Figure 4.3 illustrates
for a particular instance. The graph also shows that below the
point where 80 per cent of spending is devoted to food, actual
expenditures diverge from the sloping line, and the law ceases to
be obeyed. Indeed, the point where Engel's law ceases to be
applicable might itself be used to define an 'ultra-poverty' line
(V.B. Rao, 1981), although this cannot be done with much
precision. Using the NSS data for India, Lipton (1982, p.45)
concluded that 'for the poorest 10-30 per cent in rural areas and
5-20 per cent in towns, Engel's law does not apply'.
The larger proportion of rural people identified by Lipton as
living in ultra-poverty is again significant, especially in view of
the higher mortality rates (including infant mortality) recorded
in rural areas. Yet according to HCS data, rural food intakes are
much the same as urban populations. Noting this, Lipton
suggests that for rural people, food energy requirements are
greater. There is more heavy work to be done, not only in
agriculture but in carrying water and other tasks, and there are
more pregnancies. As we observed in Chapter 1, malnutrition
occurs at different levels of intake in different environments. By
contrast, if a single, fixed food energy requirement is assumed for
everybody, poverty and undernutrition seem more prevalent in
urban India.
It will be apparent that two sorts of data are used in these
various attempts to define criteria for the nutritional aspect of
poverty: observations of people's behaviour, and conclusions
drawn from scientific models of energy needs. The way these
data are combined depends on a variety of assumptions and



0 0 2 40 60 80 100 120 140 160 180
Monthly expenditure per person (Rupees)
Figure 4.3 Monthly consumer expenditure per person (MEP) in rural areas of Maharashtra, October 1972-September
Notes: Each vertical bar represents one MEP class. The modal class in the survey, with 1051 households, was that for Rs.
34-43 per person. There were only 74 households spending less than Rs. 13 per person, and only 25 in the Rs. 150-200
Source: NSS data, twenty-seventh round 1972-3, Sarvekshana, January 1979, quoted by Lipton 1982, table 6.


Food system indicators 87
valuations. Conventional nutritional requirement figures, based
on observing healthy populations whose intakes are not con-
strained by poverty, provide one type of behavioral evidence.
These data are then frequently used in conjunction with some
form of 'genetic potential' model, and in relation to Table 4.1,
lead to the estimates which place 50 per cent and more of the
Indian population below various poverty lines.
By contrast, Sukhatme (1981a,b) uses a version of the
individual adaptability model, with minimum intake fixed and no
allowance made for differences in adaptation between rural and
urban groups and sub-groups. The proportion of the population
below the poverty line is then given as between 20 per cent and 25
per cent (Table 4.1). The same approach was used by FAO
(1977): a critical limit of response or adaptation, estimated at 1.2
x BMR (see Chapter 3 above). The authors themselves applied
this level to the 1971-2 data from India and estimated that
around 17 per cent of rural people should be regarded as
suffering from undernutrition at this level. The estimate is very
close to Sukhatme's figure.
Lipton (1982) relies much more than the other authors on
observed behaviour that is symptomatic of critically low levels of
food intake, and by implication gives less weight to any of the
scientific models. Assuming that the real extent of adaptability in
individuals is reflected in their behaviour, Lipton's results may
well come closest to reflecting real limits of adaptability or
response in the populations he considers. This makes his
conclusion about rural/urban differentials doubly important.
Observing people's food purchasing behaviour falls a long way
short of the 'dialogue' between scientists, planners, and consum-
ers of food we earlier envisaged (p.77). But taking account of
behavioral evidence is an important way of at least beginning an
exploration of social valuations regarding food needs; it points to
some of the more extreme limits of what is acceptable in a given

Problems of cut-off points: a poverty lines
Amartya Sen (1974) has commented that the poor of India 'may

88 Limits to Measurement

not be accustomed to receiving much help', but they are
'beginning to get used to being counted'. Knowing that precisely
X per cent of India's or the world's population falls below a
certain poverty line does not necessarily help alleviate that
poverty. Thus the number of 'ultra-poor' people identified by
Lipton matters less than the way he is also able to point out
concentrations of rural poverty in certain places on inferior
land in Bihar, Assam, Orissa and West Bengal, for example, or
among the low-caste and tribal landless in western India. These
latter are the kind of facts on which some sort of action might be
based. We shall take up these ideas again later.
Scott and Mathew (1983) suggest in a study of poverty in
Kerala that the variables relating to poverty, 'should be express-
ed initially in terms of a distribution', without reference to any
particular poverty line, because the choice of a cut-off point is
always relative to the socio-economic context. They go on to
argue that, 'it may be more helpful to administrators or local
village committees whose job it is to provide relief or ameliorate
conditions, to decide on their own critical points ... in the light of
... resources and needs, and with full knowledge, rather than the
half knowledge provided by a poverty line'.
This is an approach to be commended wherever it can be
adopted, because once cut-off points have been defined and
found to be useful, they are often taken over and used out of
context by other investigators. Poverty lines based on income,
for instance, are not necessarily accurate indicators of nutritional
risk. Rigidity in the use of an indicator implies that no dialogue is
taking place about what cut-off points might best represent
'unacceptable' conditions. Furthermore, in a country in which
development policies are achieving at least some of their goals,
we would expect the borderline between acceptable and un-
acceptable risk to be raised continuously. Poverty lines should
then be adjusted from year to year. Actually the poverty line
used by the Indian Planning Commission is based on 1961-2 real
prices. And another example of what ought not to happen is
provided by Woolf (1946) and Walker and Church (1978):
Rowntree's 'absolute minimum for survival' poverty line con-
tinued to be used in Britain for a period of over forty years and

Food system indicators 89
indeed is still the basis of state social security and welfare
intervention. One danger with Lipton's double-80 criterion is
that its evident usefulness and apparent objectivity may give it a
similar permanence.

Problems of cut off points: b anthropometry
Similar considerations apply to the cut-off points used in
connection with anthropometric data. Gomez et al. (1956) used
hospital mortality results to define a cut-off point at 60 per cent of
standard weight-for-age, and described children below this level
as suffering from 'severe' or 'third degree' malnutrition. This
cut-off point and these terms have been used by many subse-
quent authors, and have become part of the vocabulary of
nutrition; they are summarised in Table 4.2 along with similar
classifications used with other anthropometric indicators.
The use of Harvard standards for heights and weights was
mentioned in Chapter 2. The standards were originally derived
from measurements of several thousand American children, and
are presented as sets of age-standardised centiles. These should
be regarded as probability statements. The 90th centile for any
measurement is a set of numerical values below which we would
expect to find 90 per cent of a sample, with only 10 per cent
above. It thus represents a condition well above average, and in a
set of data representing body weights, the 90th centile would
correspond to about 115 per cent of standard weight-for-age.
At the other end of the scale, the 1st centile might be used as a
cut-off point in weight-for-age to identify children who are at risk
in some way. There is a probability of only 0.01 that a healthy
individual would have this weight, so if we found that all the
individuals in a sample were on or below the 1st centile, we
would have good reason to think that not all of them were 'well
nourished'. In the Harvard distribution, body weights of indi-
viduals in the 1st centile are near 60 per cent of those of
individuals in the 50th centile that is, they are about 60 per cent
of 'standard'. And as we have said, 60 per cent weight-for-age is
very widely taken as the cut-off point identifying individuals who

90 Limits to Measurement

Table 4.2 Panel showing examples of Harvard Standards for
anthropometric data and illustrating some classifications commonly

(a) Harvard 50th centile weights for young children of both sexes at
representative ages.
Age/years 0 0.5 1 2 3 4 5
Weight/kg 3.4 7.4 9.9 12.4 14.5 16.5 18.4

(b) Interpretation of weights below Harvard Standard 50th centile
Weight as % of Gomez
Harvard standard classification
> 90 normal includes individuals in the
1 25th centile and above
75-90 < undernutrition
(1st degree)

60-75 undernutrition
(2nd degree)

<60 severe includes only individuals
(3rd degree) in the 1st centile

have a low probability of being 'healthy', and therefore a high
probability of 'severe malnutrition' (Table 4.2).
These cut-off points are given because they are so widely used,
but we also need to recall that their significance is always relative
to our circumstances, resources and purpose (Chapter 2). In
screening a particular population, for example, we might have
resources to treat all those on the 10th centile and below. This
means that we would use 85 per cent weight-for-age as a cut-off,
regardless of how it relates to the Gomez classification.
It is sometimes possible to attach greater significance to a
cut-off point if it defines conditions where some particular risk
begins to increase steeply. For example, in the group of children
studied by Chen et al. (1980) in Bangladesh mentioned in
Chapter 2, the numbers of deaths occurring in the two years after
anthropometric data were recorded has the distribution relative

Food system indicators 91

Harvard 50th centile heights or lengths for young children of both sexes
at representative ages.
Age/years 0 0.5 1 2 3 4 5
Length/cm 50.4 65.8 74.7 87.1 96.0 103.3 109.0

The Waterlow classification (Waterlow 1972, 1976) combines two
indicators, weight-for-height, and height-for-age. Children of low
weight-for-height are referred to as 'wasted'; those of low height-for-
age are termed 'stunted'.
More than 80% Less than 80%
Harvard Standard Harvard Standard
More than 90% classed as classed as WASTED
Harvard Standard NORMAL

Less than 90% classed as highest risk category:

to weight-for-age which is shown in Figure 4.4. There is an
'elbow' in the curve where mortality risk begins to rise rapidly,
and this break in the curve is located at values of weight-for-age
around 60-65 per cent of Harvard standard. Bairagi (1982)
quotes similar data from a group of Indian children studied by
Kielman and McCord (1978), and discusses statistical considera-
tions to locate the 'best cut-off point' more precisely on the elbow
of the curve. As Figure 4.4 shows, the point is not the same for
the two communities. Moreover, the Narangwal study also
showed that the shape of the curve depended on the age of
children measured: risk of death for a given nutritional status
decreased with age. The possibilities of using anthropometric
measurements as indicators and the statistical, rather than value
issues involved are discussed further by Habicht etal. (1982), and

92 Limits to Measurement


*--O Bangladesh
80- o--- Narangwal
S60 \

5 40- \


55 65 75 85
Percentage of Harvard standard weight for age

Figure 4.4 Variation of death rate vs weight-for-age for two groups
of children: Bangladesh (children aged 12-24 months), Narangwal,
India (children aged 0-36 months)
Notes: The vertical lines at 63 and 73 per cent weight-for-age represent
Bairagi's view of the 'best cut-off points' for nutritional monitoring.
Sources: Bangladesh, Chen et al. 1980; Narangwal, Kielmann and McCord
1978; diagram after Bairagi 1982.
also by Trowbridge and Staehling (1980) and Eusebio and Nube
Another question that arises is whether standards derived in
one country can be used to evaluate individuals in another. Some
studies indicate that there are small differences in growth pattern
between Africans, Asians and Caucasians, although these only
appear after 5 years of age (e.g. Hiernaux, 1964; Jelliffe, 1966;
Habicht et al., 1974). We may ask whether, even if all human
groups have basically the same genetic growth potential, nation-
al standards should not still be used because they take account of
environmental and economic differences.
Much discussion has been given to these questions, but to
answer them we need to come back to a basic point. The actual
value of weight-for-age at which children's mortality rates
increase sharply can be called X kg. On the Harvard standard, X
is 60 per cent weight-for-age. Any standards could be used to

Food system indicators 93
define 'malnutrition'; cut-off points may be chosen according to
measured risks, or to resources available for allocating treat-
ments, or to suit some other purpose. If such procedures are
consistently followed, it does not matter very much what
standard values are used, as long as they are derived from a large
sample, with acceptable methods of measurement.
In any case, as we saw in Chapter 2, there is no unambiguous
way of fixing points on the scale so that 'high risk' are sharply
divided from 'normal' individuals. For any fixed point, there will
be some healthy individuals who fall below it, so in a practical
situation where action is contemplated, there will be some
healthy people who are treated as if they are sick, or at risk of
Although it makes sense to use the 'elbow' of the risk curve as a
cut-off point, we should adopt this procedure with caution, for
two reasons. Firstly, to apply a statistical test in order to establish
the best cut-off point may pre-empt the dialogue we have
mentioned as necessary in making social valuations. Risks are
not zero for children who are better than 60-65 per cent
weight-for-age, and we cannot be sure that the 'elbow' of the curve
coincides with the view of these risks taken either by planners, or
by parents of the children concerned. It is also often the case,
however, that those whose children fall into the extreme, worst
category are excluded from any intervention or any dialogue at
all. Where there is an option of spending more money in order to
provide treatments (or food) for extra children, those who
provide the funds might also have a view they might wish the
treatments to be given even though the children who benefit are
in low-risk category.
Secondly, we should note that a cut-off point defined by
reference to mortality risk in one part of the world may not be
applicable to morbidity risks in the same population groups, nor
applicable at all in any useful way to children in a different region
or continent (Kasongo Project Team, 1983). Investigating the
incidence and prevalence of diarrhoea among children in Niger-
ia, Tomkins (1981) found that the frequency of diarrhoea was no
greater for children of low weight-for-age than for others, but
that such children took significantly longer to recover from each

94 Limits to Measurement

attack. Weight-for-height proved to be more effective as an
indicator of the risk of an attack. Among children who were
below 80 per cent standard weight-for-height, episodes of
diarrhoea were 1.47 times more likely than among those above
this level. Chen et al. (1981c) found similar results. Reddy et al.
(1976), and Cunningham-Rundles (1982) among others, have
looked at the association between resistance to infection and
nutritional status.
In these circumstances there are no clear-cut answers about
which indicator to use, and with what cut-off points. Parents
might be worried about the high incidence of some particular
illness, while educationists point out that slow recovery from
another disease is holding back children's learning. Different
indicators and fixed points would be needed to respond to these
different concerns. Once again, therefore, several points of view
must be considered before social valuations of risk can be fully
appreciated. And as in other instances we have considered, the
deciding factors about cut-off points must include dialogue
concerning the relative importance of risks of different aspects of
malnutrition, and a clear idea about the purpose for which the
indicator is being used and the resources available.


The Causes of

'It... places the emphasis on man, and endeavours to
study him in, and in relation to, his environment...
economic, nutritional, occupational.'
John A. Ryle, on epidemiology as an aspect of 'social medicine'
(see Ryle, 1948, p.11)

5 Multiple causes in

A focus on people
The direction in which previous chapters have pointed is towards
the need for a comparative understanding of malnutrition as it
occurs in different sectors of society, and as it is affected by
geographic, environmental and disease factors. We approached
this conclusion by a negative route, noting the limitations of
more conventional approaches. Measurements of average food
consumption do not tell us about social patterns of undercon-
sumption; requirements figures do not in practice tell us about
which food intakes are likely to lead to malnutrition. Taking a
more constructive approach, we may now say that our goal in
studying the social distribution of malnutrition is to understand
how people become malnourished in terms of processes within
society. We may then hope to find out who these people are and
what policies or changes in society might help them (Joy and
Payne, 1975; Payne, 1982).
Two major themes in any such study must be the distribution of
malnutrition in the population, discussed below in Chapter 6,
and causality, dealt with here. An example of limited but pioneer

*This chapter is based on material prepared by Peter Cutler, David Nabarro and
Philip Payne; see especially Cutler (1982b) and Nabarro (1982). We are also
indebted to a teaching text on epidemiology by Jones et al. (1982) for one way of
presenting the concept of multiple causes.

98 The Causes of Malnutrition
work on both aspects was carried out in the UK in the 1930s,
under the banner of 'social medicine'. In one study, begun in
1932, James Spence studied malnutrition among children in the
northern British city of Newcastle-upon-Tyne. His work
'amongst the poorest classes' led him to think that three factors
were interacting: infectious disease, the effects of poor housing,
and inadequate diet. He described the symptoms he saw in the
children as 'apparent malnutrition' to express his view that
inadequate food was not always the prime cause (Spence, 1960).
The relative importance of infection, environment and food in
causing the problems with which we are concerned has been a
continuing subject for debate. For example, Gopalan (1983)
accepts that environmental factors often influence the growth of
children in ways that lead to malnutrition, but argues that:
'There is no way by which environment can influence growth and
development except by conditioning and modifying the availabil-
ity of essential nutrients... at the cellular level.' Thus retarded
growth: 'is a reflection of under-nutrition and of nothing else'
(Gopalan's italics). However, poor housing can have an effect by
inducing 'stress situations, like exposure to cold or infections
which condition the net availability of nutrients at the cellular
In his work among children in Newcastle, Spence recognized
the same issue. The aim should be to study, 'how disease and
social circumstances are related', and to work at a level that
'places the emphasis on man, and endeavours to study him in,
and in relation to his environment ... economic, nutritional,
occupational' (Spence, 1960; quoting Ryle, 1948).
Our argument is that we chiefly require better understanding
of these factors at a level accessible to policy interventions such
as improvements in housing or more ample food supplies.
Indeed, we might need to know which of these options is likely to
give more benefit to a particular social group, and at what cost.
This is not an easy set of issues to pursue, however. The
environmental circumstances of people's lives are very varied,
and a causal analysis of nutritional problems could rapidly
become an excessively cumbersome analysis of the human
condition (Payne, 1976). One strategy adopted in the past to

Multiple causes in malnutrition 99

avoid such complexities was to identify those few symptoms or
combinations of symptoms for which simple nutritional explana-
tions do seem to work. During the period from 1930 to 1950,
nutritionists following this procedure were successful in identify-
ing several vitamin deficiencies, such as pellagra (niacin deficien-
cy) among impoverished people living mainly on maize in the
United States. They also identified thiamin deficiency among
polished-rice eaters in Japan, and inadequate vitamin A in the
diets of low-income groups in Newfoundland. These investiga-
tions were all followed by fortification projects, and over
subsequent years, deficiency symptoms in the populations de-
clined. A careful historical analysis shows that in all these
situations, however, the decline was as much associated with
rapid social and economic improvements as it was with the
nutritional intervention: indeed in all cases the decline was
already well under way before the programmes were started.
However, these interventions established a pattern for public
health nutrition in the 1950s, and it was against this background
of apparent success that the disease of small children called
kwashiorkor gradually came to be accepted as due to protein
deficiency alone. We saw in Chapter 3 that more recent research
has largely disposed of this view.
Some diseases which seem explicable in terms of nutrient
deficiencies are distressingly widespread. The documentation for
India's sixth five-year plan quotes lack of vitamin A in diets as the
cause of widespread xerophthalmia, leading to blindness in the
most severe cases (Government of India, 1981). Nutritional
anaemia and goitre are often cited as the two other most
prevalent deficiency diseases affecting the poor in many develop-
ing countries (FAO, 1977).
Nothing that is said in this book should be taken as minimising
the significance of these forms of malnutrition. Our argument is
simply that the model of single dietary deficiency causes used to
understand them is of limited application. It is based on an
epidemiological approach, certainly, but one which lacks adequ-
ate means of analysing the multiple causes of disease. For
example, according to FAO (1977), nutritional anaemia affects
at least 20 per cent of children in developing countries and even

100 The Causes of Malnutrition
more adult women; among men the pre% alence could be around
10 per cent. This does not necessarily mean their diets are
deficient in iron however. One stid\ of labourers employed on
road works in north India, for example, found that although
there was no correlation between work output and energy.
intake, reduced levels of output obser\ ed for some labourers
%ere siinificantl\ correlated \ith anaemia (Tandon etal., 1975).
Surpriingl\, perhaps, the diet eaten by these men included
'generous'helping~ of lentils. and supplied more than three times
the recommended dailyb allowance of iron. As with man\ other
people eating largely vegetarian diets, the problem seemed to be
one of malabsorption of iron rather than inadequate intake. In
many social groups there is also a hea\\ burden of parasitic
disease due mainly to poor animationn, and this gives rise to iron
loss. Thus even a ts perficial account of one type of anaemia leads
us to notice other causes apart from deficient food intakes. (For
example, see M.C. Latham et al.. 1tl3.1
With other illnesses that are conventionally attributed to poor
nutrition, causal explanations become even more complex.
Referring again to k\ ashiorkoi in \ ,ung children, it is becoming
increasingly clear that this is not a problem of low food intake
alone (Whitehead. 1977; Lunn etal.. 19"79; Martorell eI al., 1980).
Modern research has come to focus on diarrhoeal infections as a
precipitating factor, and there is a danger that those who think in
terms of single causes \ill now rather easil\ fall into the habit of
speaking about infection as 'the cause' of malnutrition. Howev-
er, the piecise mechanisms b\ whichh diarrhoea leads to this type
of malnutrition are uncertain. Also while e some investigators
(Rutishauser, 1974; Mata et al., 1977; Tomkins 1983) suggest
that anorexia is the main cau-e. others stress that, during
diarrhoea, there is malabsorption of nutl ients due to abnormality\ in
gut flora and function (Gracey et al., 1977; Rowland, 1981); yet
others (Tomkins et al., 1983) have shown that the fever associated
with infections raises the bod\'s requirement foi dietary) energy.
In e\anminin the nutritional problems of young children, we
first have to remember that their energy needs are prop-
ortionatehl higher than those of adults When children are
bi east-fed, mothers produce adequate amounts of milk to

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs