HD1M17.

G5PX llfiM

1
Bare
uiror.
Kg v -

-
Agriculture and
Economic Development
Agriculture and
Economic
Development

Subrata Ghatak
Senior Lecturer in Economics
University of Leicester

Ken Ingersent
Senior Lecturer in Economics
University of Nottingham

Wheatsheaf
Books
DISTRIBUTED BY HARVESTER PRESS
First published in Great Britain in 1984 by
WHEATSHEAF BOOKS LTD
A MEMBER OF THE HARVESTER PRESS GROUP
Publisher: John Spiers
Director of Publications: Edward Elgar
16 Ship Street, Brighton, Sussex

© Subrata Ghatak and Ken Ingersent

British Library Cataloguing in Publication Data

Ghatak, Subrata
Agriculture and economic development.
1. Underdeveloped areas—Agriculture—Economic
aspects
I. Title II. Ingersent, K.A.
338.1'09172'4 HD14I7

ISBN 0-7108-0142-4
ISBN 0-7108-0137-8 Pbk

Photoset in 10 pt Times by Thomson Press (India) Limited, New Delhi and

Printed in Great Britain by Biddles Ltd, Guildford and King’s Lynn

All rights reserved

THE HARVESTER PRESS PUBLISHING GROUP

The Harvester Press Publishing Group comprises Harvester Press Limited
(chiefly publishing literature, fiction, philosophy, psychology, and science
and trade books). Harvester Press Microform Publications Limited (publish¬
ing in microform unpublished archives, scarce printed sources, and indexes
to these collections) and Wheatsheaf Books Limited (a wholly independent
company chiefly publishing in economics, international politics, sociology
and related social sciences), whose books are distributed by The Harvester
Press Limited and its agencies throughout the world.
To
Anita and Elizabeth
■
Contents

Preface x'

Chapter 1 Introduction 1

Chapter 2 Structure and Characteristics of Agriculture in

LDCs 4
1. Traditional agriculture 4
2. Access to non-labour resources 18
3. The farming environment: natural hazards
and economic uncertainties 21
4. Summary 23
References 24
Notes 25

Chapter 3 Role of Agriculture in Economic Development 26

1. A framework of analysis 26
2. Product contribution 27
3. Market contribution 39
4. Factor contribution 43
5. Foreign exchange contribution 66
6. Conclusions 69
Appendix: Estimation of labour force growth
rates 21
References 21
Notes 23

Chapter 4 Theory of Rent and the Concept of‘Surplus’ 25

1. Introduction 25
2. Economic rent 75
3. The theory of rent 76
4. Rent and quasi-rent 78

vii
viii CONTENTS

5. The Ricardian ‘corn rent’ 78

6. The rental market 79
7. Agricultural surplus 82
8. Characteristics of landownership in
underdeveloped agriculture 86
9. The theory of share tenancy 87
10. Some extensions of the share tenancy model 94
References 95

Chapter 5 Agriculture in Dualistic Development Models 97

1. Introduction 97
2. The Lewis model 98
3. The Fei-Ranis (FR) model 105
4. The Jorgenson model 112
5. Kelley, Williamson, Cheetham model 114
6. Some concluding remarks on the dual
economy models 119
References 121

Chapter 6 Resource Use Efficiency and Technical Change

in Peasant Agriculture 123
1. Efficiency of resource utilisation 123
2. Technological change in agriculture 142
3. Generation of new agricultural technology 147
4. Factor-biased technological change and its
distributional consequences 155
5. Agricultural technical change and
agricultural employment: empirical
evidence 160
6. Agricultural resources and technical change
in LDCs: policy conclusions 164
References 169
Notes 171

Chapter 7 Supply Response 172

1. Introduction 172
2. The Cobweb model: an illustration 174
3. ‘Perverse supply response’ in backward
agriculture 179
4. A simple supply response model 181
5. Supply response in the underdeveloped
agricultural labour market 185
 CONTENTS ix

6. The concept of ‘marketed surplus’: some

methods of estimation 189
7. Some criticisms of Krishna’s method and
the alternative approach of Behrman 192
8. Perennial crops 199
9. Conclusion 214
References 214

Chapter 8 Institutional Constraints on Agricultural

Development and Remedial Policies 217
1. Inequitable landownership and land reform 217
2. Capital and finance in underdeveloped
agriculture 227
3. Marketing imperfections and marketing
policy 237
References 250

Chapter 9 Population and Food Supplies 252

1. The classical model 253
2. Contra-Malthusian model 256
3. Ecological disequilibrium 262
4. Synthesis of population and food supply
theories 263
5. Malnutrition in developing countries 265
Appendix: Formalisation of Boserup’s model 275
References 276
Notes 278

Chapter 10 Agriculture and International Trade 279

1. Introduction 279
2. Main features of trade in agricultural
goods 280
3. Trade policies in developed countries and
their impact on agricultural trade 282
4. Welfare gains from price stabilisation 286
5. Export instability and economic growth:
some macro issues 289
6. A survey of the literature 290
7. Some measurement problems 293
8. Prebisch’s hypothesis 294
9. The agricultural self-sufficiency argument 296
 CONTENTS
X

10. Cartels in commodity trade and welfare

gains and losses 299
11. Integrated commodity agreement (ICA)
schemes 301
12. The compensating financing schemes 306
References 307
Notes 308

Planning Agricultural Development 309

Chapter
1. Introduction 309
2. Meaning and objectives of economic planning
in LDCs 309
3. Macro-planning: comparison of Soviet and
Chinese models 311
4. Choice of planning strategy for agriculture 317
5. Planning techniques 322
6. Agricultural project planning 329
References 362
Notes 363

Chapter 12 Agricultural Development: an Overview 364

Index
372
Preface

Many textbooks are available in the general field of ‘development

economics’ and several on ‘agricultural economics’ with a developed
country orientation. But very few texts have been written centring on
the economic problems of agriculture in developing countries and
virtually none in recent years. This book is our attempt to fill this gap
in the literature.
The principal objectives of the book are to analyse agriculture’s
role in the development of Third World countries, to identify barriers
to agricultural development and to examine critically remedial
agricultural policies to foster more rapid development.
The book is primarily written for students of economics and
agricultural economics who are already reasonably well versed in the
principles of economics as taught to first and second year under¬
graduates in British universities. We expect that many of our readers
will already have some knowledge of the wider field of development
economics, or will be taking a parallel course in that subject. Thus our
primary purpose is to illuminate the economic problems of agricul¬
ture in less developed countries (LDCs) and to analyse policy
alternatives by applying well-known tools of economic analysis.
However, as in other fields of applied economics, the student of
agricultural development economics must acquire an adequate
knowledge of the relevant institutional framework. Thus, an import¬
ant secondary objective is to provide readers with a basic knowledge
of the institutions through which reformist policies for agriculture in
LDCs are commonly conducted. These institutional aspects include
land reform, the provision of credit and marketing services, price
stabilisation, and the advancement of agricultural research and
education. We envisage that the book might serve either as the main
text for a full-length course on agricultural development economics,
or as a supplementary text for less specialised courses in the economics
of development or agriculture.

xi
xii PREFACE

The analytical sections of the book consist primarily of verbal

arguments, supplemented in some cases by simple diagrams. Our
sparing use of mathematics (mainly algebra) and statistics is confined
to tools of analysis forming an integral part of virtually all modern
economics degree courses.
In common with most textbooks, this one is largely an amalgam of
many people’s ideas. We accordingly gratefully acknowledge our
considerable intellectual debt to the many authors cited in our
references, as well as to our numerous unnamed professional
colleagues. SG wishes to thank David Pyle and Clement Ayisa in
particular for their help and co-operation. A special word of thanks is
due to Edward Elgar of Wheatsheaf Books for his help and en¬
couragement in publishing the book. We also thank Olwen Bradley,
Jane Dewick, Gill Limbery, Dorothy Longsdon, Janet Wimperis and
Marian Wearmouth for typing our manuscript with so much patience
and skill. We are also indebted to our students for the stimulus they
have given to our teaching and indirectly to the writing of this book, by
their many pertinent observations and critical comments over a period
of years. But we naturally accept full responsibility for all such errors,
omissions, inconsistencies and other defects that may remain in the
text.
1 Introduction

In this book, an attempt has been made to provide an up-to-date and

comprehensive account of the interaction between agriculture and
the economic development of the less developed countries (LDCs).
We give a detailed account of both theory and techniques to help the
readers to evaluate policy implications. The plan of the book is as
follows:
First, we consider how developed countries (DCs) and LDCs differ
in respect to the structure and organisation of agriculture and the
behaviour of agricultural producers. Some features of traditional
agriculture receive special attention and a few important peculiarities
in the behaviour of peasant farmers in LDCs are emphasised. Then
some comments are made on such peculiarities of the farming
environment and its effects on the behaviour of peasant farmers.
In chapter 3, the role of agriculture in economic development is
thoroughly investigated in the light of the Kuznets’s analysis.
Agriculture provides both food and raw materials to the rest of the
economy; a growing agricultural sector provides an enlarged market
as it expands aggregate demand; it also provides labour for employ¬
ment in the industrial sector; and agriculture is often a principal
source of capital for investment elsewhere in the economy. Exports
from the agricultural sector are important to earn foreign exchange
which is critical for imports of capital goods and other equipment for
rapid industrialisation and economic growth.
In chapter 4, we analyse the concept of agricultural surplus and the
theory of economic rent. The interchangeability of product, land and
labour surpluses is mentioned, and the pattern of distribution under
alternative tenure systems is noted. Alternative systems of land
tenure are described and attention given to their role in the
agricultural systems of LDCs.
Next, in chapter 5, the dual economy models and the role of
‘surplus’ agricultural labour is evaluated critically, particularly in the

1
2 AGRICULTURE AND ECONOMIC DEVELOPMENT

light of the works of Lewis, Fei and Ranis, Jorgenson and Kelly, and
Williamson and Cheetham. Results of the ‘tests’ of some of these
models are stated to derive policy implications.
Chapter 6 is an analysis of the important issue of efficiency in
resource use in underdeveloped agriculture. In this context, Schultz’s
‘poor but efficient' hypothesis is examined. The behaviour of poor
farmers in the presence of risk and uncertainty is also analysed. The
important relationship between farm size and efficiency is examined
in this context to formulate land policies. Next, technical change in
underdeveloped agriculture is discussed in detail. Here, we have
analysed some characteristics of ‘modern’ agricultural technology
and the impact of their adoption in underdeveloped agriculture.
The barriers to adoption of new technology is analysed with the
distributional consequences of their adoption. We argue that
dangers of dualistic development within the agricultural sector of
LDCs should be properly emphasised.
Chapter 7 deals with the problem of ‘supply response’ in LDCs.
The impact of economic incentives on agricultural production and
supply are investigated. Both the theoretical and empirical issues in
the analysis of the problem of supply response are thoroughly
discussed. In this context, supply response of both annual and
perennial crops is mentioned. Finally, the role of agricultural price
policy is considered for mobilising ‘surplus’ from the agricultural
sector.
Chapter 8 is an investigation of the important institutional
constraints on agricultural development. Here, we turn to land-
ownership structure and the case for land reform. The problem of
capital and finance in underdeveloped agriculture is discussed. The
structure of agricultural credit market is analysed at length and the
problem of ‘high’ interest rates in such markets examined in detail.
Credit policy recommendations, including loan security and interest
rate structure, are formulated in the light of our analysis.
Since market imperfections present considerable problems in
LDCs, the restructuring of agricultural markets, and the necessity to
follow sensible price policy, are advocated. Here again, the case for
institutional reform has been specifically emphasised. In chapter 9,
the relationship between population and food supplies is discussed.
In chapter 10, the role of agriculture in the international trade is
analysed from the point of view of LDCs. Problems such as adverse
movements in terms of trade, primary price fluctuations, and so on, are
examined, with special attention to the welfare gains and losses from
primary product price stabilisation; objectives and achievements of
 INTRODUCTION 3

compensatory finance schemes are discussed at length.

In conclusion, we discuss the macro- and micro-planning of
agriculture. Under macro-planning, we discuss the Soviet ‘tribute’
model, the Chinese model, and their relative merits and demerits.
Next, we argue the merits and demerits of uni-modal and bi-modal
agricultural strategies. It has been argued that, from the standpoint of
LDCs, perhaps a uni-modal strategy might be superior to a bi-modal
one. Then the input-output analysis for agricultural planning is
discussed. Under micro-planning, we show how techniques such as
cost-benefit analysis can be used for agricultural project evaluation
in LDCs. Special emphasis is given to the UNIDO method of project
appraisal including a practical example of its application to the
agricultural project. At the same time, readers are made aware of the
limitations to such exercises.
2 Structure and
Characteristics of
Agriculture in LDCs

In this chapter we shall consider how developed and developing

countries differ with respect to the structure and organisation of
agriculture, and the behaviour of agricultural producers. In the first,
major section we discuss the nature and characteristics of the
‘traditional’ agricultural sector in LDCs and some peculiarities of
the behaviour of peasant farmers. Then we consider inequalities in
the distribution of agricultural land and access to capital and their
consequences for the majority of farmers. Finally, we remark on
some peculiarities of the farming environment and its effects on the
behaviour of peasant farmers.

1 Traditional Agriculture
In developed countries, where the process of economic and social
integration between agriculture and other sectors of the economy is
virtually complete, farming is a business and farmers behave like
businessmen. They are both commercially oriented and technically
well-informed. They have at their disposal the services of an array of
financial, marketing, advisory and research institutions, both public
and private. Nowadays, despite the image of poverty which their
political representatives have a vested interest in projecting, the
majority of developed-country farmers are not notably worse-off
than people in other occupations. Indeed, when account is also taken
of their relative independence, the amenity value of country living
and, tor many farmers, the social advantages and economic rewards

4
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 5

of being a rural landowner, it might be considered that in developed

countries farmers as a class are rather better-oflf than their non¬
farming compatriots. But in LDCs it is otherwise.
The term ‘traditional agriculture’ is used to describe the character¬
istic farming type in countries where agriculture is the dominant
employer (including those who are self-employed), though not
necessarily the largest sector in terms of contributions to GNP. But
before examining the attributes of traditional agriculture, let us be
clear that in most LDCs the agricultural sector is not uniform, but
marked by considerable economic and social diversity. Typically, a
‘traditional’ sub-sector of agriculture, which is large in terms of
employment but smaller in terms of production for the market, co¬
exists with a ‘modern’ agricultural sub-sector, which is relatively
small in terms of employment, but much larger in terms of its
contribution to the country’s marketed output of agricultural
products. This is referred to as ‘economic dualism’. Alternatively if,
instead of applying the dual economy concept to the agricultural
sector alone, we apply it to the economy as a whole, we can view the
traditional sub-sector of agriculture as but one element in a larger
traditional sector (which also includes other categories of self-
employment using little capital) whereas the modern agricultural sub¬
sector belongs to a larger ‘modern sector’ (including other ‘non-
traditional’ industries).

1.1 Agricultural production unit: form and function

We shall now proceed to examine the main attributes of traditional

agriculture. The first point to emphasise is that peasant agriculture is
typified by the small, family farm. As a production unit this is
characterised not only by its small size - whether measured in terms
of the volume of resources employed or the volume of output - but
also by its high degree of self-sufficiency. The main factors of
production are labour and land; few purchased inputs are employed.
The farm workforce consists principally of family labour. Despite the
possibility of a direct correlation between farm size and family size, at
least where land is relatively abundant, some hiring-in and hiring-out
of labour may occur amongst neighbouring farms according to the
differing circumstances of individual farmers, including seasonal
patterns of labour demand. Where the correlation between farm size
and family size is not direct but inverse, the fragmentation of small
farms exacerbates inequality in the distribution of land and agricul¬
tural income. On the output side, the emphasis on self-sufficiency
6 AGRICULTURE AND ECONOMIC DEVELOPMENT

implies that a proportion of farm output is not sold but retained on

the farm for consumption by the farmer’s household. In a purely
subsistence agriculture farm households produce only for their own
consumption.
In a semi-subsistence agriculture farmers produce partly for their
own consumption and partly for the market. The amount of the
marketed portion of total output may either be planned or un¬
planned. If it is planned the farmer may grow one or more cash crops
specifically for the market; if it is unplanned it is likely to consist of
food which is surplus to the farm household’s subsistence needs.
Unplanned food surpluses may be confined to years when the harvest
is unusually good. Pure subsistence farmers, and semi-subsistence
farmers of both types described, may co-exist in the same locality.
Whereas a purely subsistence farmer is not necessarily fully self-
sufficient, a fully self-sufficient farmer may also produce a market¬
able surplus. However, if no market exists, or if the farmer does not
desire a cash income, the surplus may be reabsorbed by the farm
family which consequently increases its consumption to match the
larger-sized crop - Engel’s Law does not apply to a subsistence
economy.
The specification of the peasant farmer’s objective function has
been the subject of much debate amongst economists. Whereas some
have held that his primary concern is to maximise profit (like the
modern sector farmer) others have argued that, above all, he seeks
economic security or the minimisation of risk. The pure subsistence
farmer is concerned only with managing a household, not a business,
so it is to be expected that his primary motivation will be food security
for himself and his family. The semi-subsistence farmer is both a
household manager and a businessman, so that he may be motivated
by the goal of profit maximisation, but subject to the constraint of
withholding sufficient resources from cash crop production to
provide for the subsistence of the farm household. But since, for
reasons to be discussed, his cash income is typically very low, it would
not be surprising if risk-aversion took precedence over profit maxi¬
misation in deciding which crops to grow and how to grow them.
However, allocating resources in a way which trades a marginal profit
for a marginal gain in security does not signify economic irration¬
ality, since such behaviour is fully consistent with utility maximisa¬
tion. Moreover, the results of empirical studies show that peasant
farmers who do produce for the market are not unresponsive to
changes in the relative prices of production alternatives.
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 7

1.2 Resources structure and labour reward system

Since land and labour are the dominant factors of production in

traditional agriculture, the level of production tends to be limited by
the availability of those factors. The labour input includes labour
embodied in farm-produced capital inputs, such as simple buildings,
drainage channels and land reclamation. Although, in principle, the
static constraints on production imposed by available supplies of
land and labour can be relaxed by dynamic improvements in the
quality of both inputs, this consideration is of little relevance in the
short term either to individual farmers or to those concerned with
government policy for agriculture. Moreover, for reasons yet to be
explained, there are formidable institutional and other obstacles to
improving the quality of the inputs used by farmers in the traditional
sector.
Although the scarcer factor may be either land or labour it is more
frequently land due to population pressure and limited non-
agricultural employment opportunities, combined with landowner-
ship concentration. (Restrictions on access to land in LDCs are
discussed in the next section.) The combination of scarce land and
plentiful labour favours labour-intensive systems of agriculture
giving detailed attention to crops in order to secure higher yields. The
relative scarcity of land is signified by the rent - characteristically
‘high’ under labour-intensive agricultural systems - whereas the
relative abundance of labour is signified by ‘low’ wages. But given
sufficient land, and freedom of access to it for farmers, labour may be
the scarcer factor. This favours a land-intensive system of agriculture
which places more emphasis on output per worker than on output per
unit of land. There is consequently much less incentive to grow crops
requiring detailed attention in order to achieve an economic yield.
Moreover, it is feasible to vary the cultivated area according to
changes in the available labour supply. In this situation the rent of
land is characteristically low and, indeed, individual property rights
in land may be ill-defined or even non-existent, but agricultural wages
may be relatively high, at least by the standards of the country or
region concerned.
The peculiarities of the labour reward system in traditional
agriculture derive from distinctions between hired labour (the
modern sector norm) and family labour. Whereas in the modern
sector, under competitive conditions, the firm’s equilibrium level of
employment is where the marginal productivity of labour equates
with a wage which is exogenously determined by market forces,
8 AGRICULTURE AND ECONOMIC DEVELOPMENT

family enterprises in the traditional sector are characterised by work-

and income-sharing. Work-sharing explains why open unemploy¬
ment is absent from traditional agriculture. Because the farm family
shares the work-load, the family (farm) income is also shared
amongst its members. Thus the individual worker’s implicit ‘wage’ is
set endogenously by the average productivity of the family rather
than by the marginal productivity of his labour. There has been much
discussion in the literature about the implicit wage of family workers
in the traditional sector and a minimum subsistence wage. It has been
held both that agricultural wages cannot fall below the minimum
subsistence level and that, due to Malthusian pressures, they are
unlikely to rise above that level. However, both these arguments are
oversimplistic. In the one case, no allowance is made for farm family
workers supplementing their incomes from outside sources, such as
part-time or casual employment on modern sector farms. In the
second case, it is unrealistic to suppose that the ratio of labour to land
is equally unfavourable throughout traditional agriculture because
the norms of perfect competition, including the perfect mobility of
resources, are violated in practice. Where, for a variety of reasons, the
ratio is relatively favourable, either for an individual family, a village,
a region or some even larger social group, the expectation must be
that the implicit family farm wage will be above the minimum level
for physical survival.

1.3 Labour market dualism

The different processes of wage determination in the traditional and

modern sectors give rise to ‘labour market dualism’. Although the
implicit wage of workers in the traditional sector (based on their
average productivity) is not necessarily lower than the modern sector
wage level (based on marginal productivity), it generally is so, due to
the low opportunity cost of labour in the traditional sector. Large
numbers of workers remain in traditional agriculture despite low
wages, due either to ignorance of better opportunities outside
agriculture, or to their inability to obtain a modern sector job despite
wishing to do so, or to the costs of moving being unacceptably high
(including the cost of giving up the relative security of remaining at
home) in relation to the expected wage premium. An important
element of a well-known and influential theory of the rural-urban
migration process in LDCs is that potential migrants discount the
higher modern sector wage by the probability of remaining unem¬
ployed (Todaro, 1969).
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 9

A possible counter-argument to labour market dualism is that

modern sector employers, including large-scale farmers, behave
monopsonistically in the labour market. Thus, it may be predicted
from the theory of monopsony that the volume of modern sector
employment will remain below the level of the social optimum, with
modern sector wages consequently being depressed below the
social opportunity cost of labour. However, proponents of this
theory, such as Griffin (1979), concede that due to pressures on
modern sector employers to increase wages the modern/traditional
sector wage gap is unlikely to be entirely eliminated. Certainly, there
is ample empirical evidence from LDCs that for those traditional
sector migrants who are lucky enough to obtain a modern sector job a
substantial wage premium exists (e.g. Ritson, 1973).
Some obvious consequences of price distortions in the market for
agricultural labour ensue. The cheapness of labour in the traditional
sector causes it to be used extensively there. That is, extra labour is
employed to perform tasks which would be unprofitable at the
modern sector wage rate. Moreover, cheap labour favours the
retention of labour intensive methods of production, such as cultivat¬
ing by hand rather than mechanically. Price distortions in the
markets for land and capital (see section 2 below) accentuate the
intensive use of labour. Now let us assume that the technology known
to, and employed in, traditional agriculture is given and fixed (a not
unrealistic assumption for reasons to be explained below). Then, if
access to additional farm land, or the acquisition of capital inputs
such as new tools or even fertiliser, are not feasible options due to
either their physical inaccessibility or their too high price, the
application of additional labour may be the only means left to the
peasant farmer who wants to produce more. With the number of
workers comprising the family labour force fixed, in the short-run
at least, any increase in work hours must augment the total product to
be divided amongst the family, provided only that the marginal
product per man hour is positive.
It follows from this argument that, provided the opportunity cost
of labour time is zero, it is rational to intensify labour use to the point
of zero marginal productivity. But there are two reasons why labour
time may have a positive opportunity cost, even in peasant agricul¬
ture. First, supplementary employment may be available off the
family farm as, for example, on modern sector farms that employ
hired, part-time or casual workers. Second, because the marginal cost
of extra work time is leisure forgone, it is quite possible - or even
likely - that the equivalence of the marginal utility of extra income
10 AGRICULTURE AND ECONOMIC DEVELOPMENT

and the marginal disutility of extra work will be reached at a lower

level of labour intensity than that corresponding with zero marginal
productivity of labour time. We may reason, a priori, that for leisure
to be valueless, even at the margin of consumption, a society must
either be extremely poor or surfeited with leisure time. Though either
of these conditions could occur as special cases in backward agrarian
societies, there is no compelling argument or evidence for treating
them as general case conditions applying to all traditional agriculture.
(We return to the trade-off between work and leisure, for deeper
analysis in chapters 3 and 6.)

1.4 Access to information and technological choice

Having dealt with the factor reward system in traditional agriculture

we now turn to the questions of access to information and choice of
technology. To observers from developed countries, or even to those
from the modern sector of LDCs, farmers in traditional agriculture
appear to be ignorant of modern farming methods and the tech¬
nology they do use appears to be primitive. There is little reason to
doubt that these are accurate observations, at least from a western
viewpoint. Although it would be naive to equate knowledge and use
of modern agricultural technology with economic and social well¬
being without qualification, there is ample evidence that peasant farm
families in developing countries do aspire to a higher material
standard of living. So, enquiring into the reasons for technological
backwardness in traditional agriculture, and its economic conse¬
quences, are questions of some importance.
Some analysts regard technological stagnation as the definitive
characteristic of traditional agriculture, for example, ‘Farming based
wholly upon the kinds of factors of production that have been used
by farmers for generations can be called traditional agriculture.’
(Schultz, 1964). This raises the three interrelated issues: (1) the type of
technology used; (2) why it does not change; and (3) the con¬
sequences for agricultural labour productivity and farm incomes.

1.4.1 Characteristics of Traditional Agricultural Technology

The technology itself is simple, even primitive to western eyes, though
well-tried and reliable through its use by many generations. It is also
labour-intensive and self-sufficient, with little reliance on outside
suppliers of materials or services. Thus, for example, the digging-stick
or hoe is used instead of the plough, and human draught-power, or
perhaps animal-power, is used instead of the internal combustion
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 1 1

engine. Any capital inputs which are employed, like simple imple¬
ments and containers for transport or storage, are home-produced -
either on the farm itself or in the village. There are, of course,
different degrees of technological backwardness, according to
systems of cultivation, the natural environment and the extent of
any outside influences on farming methods. Few rural societies,
nowadays, are entirely cut off from the outside world, but some are
more isolated than others. Thus, other things being equal, the plough
is more likely to be used where farmers are settled in one place than
where shifting cultivation is practised: technological change is more
likely to penetrate where cash crops are grown, perhaps with some
for export, than where production is only for subsistence. Moreover,
because, even in a traditional society, individual farmers differ in
their resource endowments, and some are better informed or more
interested in farming success than others, different methods of
performing the same task may co-exist in the same locality.

1.4.2 Reasons for Technological Stagnation

The reasons for technological stagnation can be grouped into four
main categories:
(1) the lack of an appropriate alternative technology;
(2) farmers’ ignorance of ‘better’ methods;
(3) farmers’ lack of motivation - the risks and costs of adoption;
and
(4) barriers to adoption due to other market failures.

Lack of appropriate alternatives. Could the technological backward¬

ness of traditional agriculture be due to the absence of alternative
technologies with a higher production or profit potential which are
also properly adapted to its conditions of plentiful labour, very scarce
capital, very small scale of production, and limited literacy and
knowledge of modern technical skills amongst farmers? Much has
been written, in a dual economy context, about the ‘inappropriate¬
ness’ of attempting to transfer modern sector technology to the
traditional sector: the transfer of technology from DCs to LDCs, by
multinational companies and through trade and aid, has also been
condemned. So, for example, it may be feasible to evolve tractors,
power-tillers, harvesters, irrigation pumps, and other items of farm
equipment which are better adapted than large-scale, expensive and
labour-displacing ‘Western models’ to the requirements of small-
scale producers in LDCs. The hallmarks of a more appropriate
agricultural technology might include technical simplicity, low cost
12 AGRICULTURE AND ECONOMIC DEVELOPMENT

and capacity to raise farm productivity without displacing labour.

The case for fostering the development of more appropriate
alternative technologies which, if successful, would help to remove
economic dualism by reducing and ultimately closing the present gap
in per capita income in favour of the modern sector, has been well
made (Stewart, 1978). A number of specialised institutions, such as
the Intermediate Technology Group, have been established to collect
and exchange information, to encourage research and to foster the
actual creation of more appropriate technology, for use in both
industry and agriculture, in other ways. Yet to carry this argument to
the point of saying that there can be no technological advance in
traditional agriculture without the invention of ‘new tools’ is
unconvincing. By no means all technical improvements involve
substituting capital for labour, or call for a high order of technical
skill or literacy in their application. Nor need they be particularly
expensive to introduce. For example there may be scope for raising
crop yields by changing seed varieties or sowing dates. Or it may be
feasible to raise farm income by producing a crop which has not been
grown before. This line of argument suggests that farmers in the
traditional sector may not be fully informed about existing alterna¬
tive techniques which are appropriate and have an output and
income-raising potential.
Lack of knowledge: barriers to spread of technological information. In
seeking to explain why farmers may be ignorant of better methods it
is tempting to blame their low levels of literacy and formal education.
Although this notion cannot be entirely discredited, a number of
other factors also need to be considered. However receptive people
may be to new knowledge and new ideas, actual technological
innovation is dependent on effective channels of communication
between the sources of new knowledge and its recipients. In
agriculture there are two major impediments to the dissemination
of new information. First, in most countries much of the farm
population is located in relatively remote and inaccessible areas.
However, their ‘remoteness’ and ‘inaccessibility’ are, to a consider¬
able degree, expressions of the under-development of the rural
communications infrastructure, particularly in terms of the number
and quality of roads and railways. The underprovision, by developed
country standards, of other means of communication such as postal,
telegraph and telephone services add to the relative isolation of the
rural population. Newspaper circulation is low, partly because of
widespread illiteracy but also because of the difficulties of distri¬
bution in remote areas. A comprehensive analysis of the reasons why
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 13

the rural communications infrastructure is deficient extends well

beyond the scope of our subject-matter, except to make two points:
first, the per capita cost of providing services tends to be higher in
rural areas than in towns and cities because of their relative
geographical isolation and lower population density (this occurs in
developed as well as in developing countries); and secondly, the
problem is exacerbated by the relative poverty of the rural popul¬
ation. In most LDCs, although poverty is not confined to the rural
sector it is concentrated there: farm incomes are substantially below
non-farm incomes on a per capita basis. The relative poverty of
farmers is the second of the major barriers to the spread of in¬
formation about new farming methods, and technological stag¬
nation can be viewed as an aspect of the vicious circle of poverty.
Farmers fail to adopt better methods because they cannot afford to
travel to find out about them - they may even be unable to afford a
radio.
But is it necessary for farmers to travel in search of new
information? Can it not be conveyed to them at home by their
professional advisers? In developed countries farmers get advice from
both public and private sources. Advice is available (usually free)
both from government advisers and from the suppliers of the many
purchased inputs used in modern agriculture. In LDCs agricultural
advice is much less readily available for several reasons. Dealing first
with the government side, there are two major problems. First, there
is the difficulty of recruiting and training adequate numbers of
agricultural extension personnel. Suitable recruits need to be of
better-than-average education and also able to command the respect
and confidence of farmers. They should also be motivated towards
working in agriculture and living in rural surroundings. However,
due to agriculture’s inferior status as a vocation in most LDCs, it
is often difficult to find adequate numbers of people with this moti¬
vation amongst those with sufficient education to compete in the
urban professional job market. Whereas university degree courses in
medicine and law are invariably over-subscribed, would-be univer¬
sity entrants in LDCs rarely express a preference for specialising in
agriculture. The second problem is the high cost of disseminating
advice to large numbers of small farmers, many of them in remote
parts of the country and/or in areas where access to villages and
individual farmers, using modern methods of transport, is impeded
by infrastructural underprovision as already discussed. In particular,
all-weather roads (either main or feeder roads) are frequently lacking.
Thus, handicapped as they are by inadequate numbers of trained
14 AGRICULTURE AND ECONOMIC DEVELOPMENT

personnel, and subject also to budgetary constraints on the cost of

travel and transport, it is small wonder that government agricultural
extension services in LDCs tend to concentrate their meagre re¬
sources on the bigger farmers and those in the more accessible areas.
Farmers in the modern sector - small in number but large in terms of
their contribution to marketed agricultural output - tend to hold the
advantage with respect to both farm size and farm accessibility.
Farmers in the traditional sector tend to lose out on both counts.
The second potential source of agricultural advice is the private
sector where advice complements trade. Profit-motivated commer¬
cial firms tend to concentrate on giving advice to large-scale
commercial farmers in the modern sector who are their major
customers. Attempting to expand trade and accompanying advice
into the traditional sector is fraught with much greater risks and
uncertainties. The initial demand for modern inputs in the traditional
sector must inevitably be small and the costs of servicing it
correspondingly high. The high costs of travel to remote areas, and
the difficulties of reaching farmers where there are no proper roads,
are unlikely to be appreciably less daunting to private enterprise than
to the government. In order to make such business profitable it might
be necessary to raise the prices of goods supplied above the level that
peasant farmers would be willing to pay, or the cost of recovering
debts might be prohibitive. We conclude that in considering whether
to offer services to small-scale farmers in remote areas the private
sector is confronted with the same high costs as government, and the
benefits are very uncertain. The difference is that, being profit-
oriented, the private sector is even more likely to reject the option.
Inadequate motivation for adoption. We now consider whether
technological stagnation could be due not to ignorance, but to
farmers’ inability to perceive the benefits of adopting new tech¬
nology. The perception of adequate benefits provides the motivation
for adoption. The main barriers to perception are the risks and costs
of adoption. These are interrelated. The risks of adoption include the
potential costs of making mistakes due to inexperience. Although
traditional methods of agriculture suffer from low productivity their
margins of error also tend to be low, because the traditional methods
have been refined and improved by use over many generations.
Mistakes are less likely to be made in repeating familiar tasks than in
applying new and unfamiliar ones. For example, to an inexperienced
user of chemical herbicides that method of weed control would be
more risky than hand weeding, in the sense that the risk of damaging
the crop would be higher. Similarly, in adopting a new crop variety
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 15

which, under experienced management, is capable of out-yielding

more traditional varieties, an inexperienced farmer might suffer a
lower yield, or even a complete crop failure, due to mismanagement.
These are examples of production risks but farmers embarking upon
commercial enterprises also face marketing risks as discussed in
section 3.
But why should risk deter farmers in traditional agriculture from
adopting new methods? It is not so in the West where the acceptance
of risk is regarded as a normal consequence of being in business.
Indeed, under capitalism, the function of profit is to reward enterprise
and risk-bearing. But, as already discussed in this chapter, peasant
farmers are not western-style business entrepreneurs. Their objective
function is different: some produce only for their own subsistence;
others produce partly for subsistence and partly for the market. Most
are poor, so poor in fact that they cannot afford to take unnecessary
risks. A crop failure could result in hunger or even starvation for
members of the farm household, so it is safer to grow relatively low-
yielding crops by well-tried methods than to hazard diverting
resources to a new enterprise or technology with a higher expected
yield but with a higher risk of failure also. Whether the unfamiliar
technology concerns a subsistence crop or a cash crop makes little
difference. If it is the former, then a crop failure curtails the family
food supply; if it is the latter, then the family’s cash income to buy
food and other essentials is reduced.
The extent to which the adoption of new technology does in fact
expose the adopter to additional risk is, of course, arguable. To
suppose that the degree of risk varies according to both the particular
technology and the individual adopter is a plausible hypothesis.
Some modern agricultural technologies are ostensibly designed to
reduce risk; irrigation to offset the effect of drought, for example. But
all this is really a matter for empirical verification, which we return to
in chapter 6. For our present purpose the important point is that
farmers commonly believe that some, though not necessarily all, new
methods are more risky than traditional techniques. In a sense,
contrary empirical evidence is irrelevant unless they can be persuaded
to change their minds, but this is bound to take a long time.
We now turn from the risks to the costs of adopting new
technology. Even though a farmer may be convinced that a new
method of production is of potential benefit to himself, he may still be
unable to fund its adoption. If, in reality, adoption is without risk this
seems perverse. At positive odds, there can.be no scarcity of backers
for a ‘sure winner’. It is understandable that the farmer may lack
16 AGRICULTURE AND ECONOMIC DEVELOPMENT

sufficient savings to raise the stakes himself, either because he is too

poor to save or because, in the past, he has lacked an adequate
incentive to save. This could be explained by, for example, a very low
marginal productivity of capital using traditional methods of
agriculture.
But this does not explain a peasant farmer’s inability or unwilling¬
ness to borrow in order to fund the adoption of new technology
offering a favourable marginal return on investment. However, such
a state of affairs can be explained by capital market imperfections.
Outside investors may over-estimate the risks of lending to agricul¬
ture due to imperfect knowledge. Hence, they may either be unwilling
to lend to agriculture at all, or only at a risk premium that farmers are
unable or unwilling to pay. Even without misinformation about the
true risks of lending to agriculture, the price of credit may be biased
against small farmers for other reasons. Expressed as a proportion of
the value of the loan, the transactions cost of advancing credit to
small farmers in small packages must tend to be higher than the cost
of making large advances to large-scale farmers. Therefore, small
farmers in the traditional sector may either be refused institutional
credit or have it offered to them only on terms they find unacceptable.
Their only alternative source may be the private moneylender whose
transactions costs are probably lower (because of his proximity to the
borrower) but whose price may include an element of monopoly
profit. These, then, are the capital market imperfections which
restrict the supply and raise the price of capital to farmers in
traditional agriculture. Credit restrictions hamper the adoption of
new technology and therefore help to explain technological
stagnation.
Other barriers to adoption. Let us now suppose that a farmer in the
traditional sector knows about better methods of production, desires
to adopt them and is able to fund adoption, either from his own
savings or by borrowing. His adoption of new technology may still be
blocked by bottlenecks in the supply of physical inputs such as seeds,
fertilisers, pesticides and irrigation equipment. These bottlenecks
may be caused by inadequate domestic manufacturing capacity, by
import constraints, or simply by an under-developed distribution
network’s inability to cope with an increased demand for its service.
Moreover, in situations of excess demand caused by supply bottle¬
necks, farmers in the modern sector may have preferential access to
available supplies. Not only do modern sector farmers buy in larger
lots, they also tend to have greater influence with both distributors
and government. Political influence is of particular value to farmers
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 17

if the distribution of modern agricultural inputs is directly controlled

by government, which commonly occurs in LDCs. In contrast to the
strong bargaining position of modern sector farmers in competing for
scarce input supplies, farmers in the traditional sector are commonly
found to be in a position of relative economic and political weakness.
The market failures underlying these weaknesses are the fourth and
last of our explanations of technological stagnation in traditional
agriculture.

1.4.3 Consequences of Technological Stagnation: Labour Pro¬

ductivity and Farm Incomes
Although traditional methods of agriculture are ‘safe’ and well-
adapted to the relative scarcities of land, labour and capital in the
sector, they suffer from the drawback of low productivity. Where
traditional crop varieties are grown by traditional methods the
expected output per unit of land area is usually lower than for modern
crop varieties grown by modern methods. This explains the low
productivity of land in traditional agriculture. But because of labour
market dualism (section 1.2, above) the productivity of labour is
depressed as well as the productivity of land. A low level of farm
income follows as an inevitable consequence of low labour productiv¬
ity. Poverty is the price of technological stagnation in traditional
agriculture.
The relative poverty of peasant farmers explains their aversion to
risk and their unwillingness or inability to innovate and invest, as
already discussed. But there are also other side-effects. Extreme
poverty can cause malnutrition and ill-health. Moreover, these
consequences of poverty may have feedback effects upon labour
productivity. Because they are underfed or suffer from debilitating
diseases, peasant agricultural workers may lack sufficient energy to
intensify their work effort during periods of peak labour demand.
Thus, the sowing of crops may be delayed beyond the optimum date
or necessary weeding may remain uncompleted, with inevitable
adverse consequences for crop yields and labour productivity. To
western eyes the ratio of working to non-working time in traditional
agriculture may appear to be low. For example, it has been reported
that on small farms in Africa ‘adult family workers typically work
only twenty to forty hours a week on their land, even during peak
months’ (Lele, 1975, p. 24). Although it may be suspected that part of
the non-working time may be enforced rest due to an inadequate diet
and poor health, there is unfortunately insufficient systematic
evidence to confirm or reject this hypothesis.
18 AGRICULTURE AND ECONOMIC DEVELOPMENT

Poverty in agriculture is also a cause of other social ills, such as

high infant and neonatal mortality. On the other hand, the fertility
rate in poor rural populations tends to be high, possibly in
compensation for high infant and child mortality. In a society where
every newborn baby is seen as a potential addition to the family
labour force, and where parents are anxious to ensure that their
offspring survive to care for them in their old age, parents may tend to
‘over-insure’ against the risk of losing their children by premature
death. Hence the problem of poverty in traditional agriculture is
exacerbated by excessive population growth. Poverty is perpetuated
by the pressure of population on the land in seeming confirmation of
the Malthusian doctrine.
We return to the subject of population and food supplies in
chapter 8. In this chapter we confine ourselves to describing tra¬
ditional agriculture and its inherent problems. All discussion of
policy options for reforming and modernising traditional agriculture,
in order to ameliorate its economic and social ills, is deferred to later
chapters.

2 Access to Non-labour Resources

In many developing economies, pronounced inequality in the

distribution of resources - particularly land - is a major character¬
istic of the rural sector. The distribution of land amongst families
dependent on agriculture for their livelihood is important for two
reasons. First, since land tends to be the scarcest factor of production
in traditional agriculture, there is a close and direct correlation
between farm size (land area) and family income. In other words, the
economic well-being of the farm household depends, to a large
extent, on the amount of land to which it has access. Although there
tends to be an inverse relationship between farm size and land-use
intensity, diminishing marginal returns to non-land inputs, parti¬
cularly labour, prevent farmers from fully compensating for the
handicap of small farm size by means of more intensive cultivation.
Regardless of their efficiency in terms of output per unit of land area,
many traditional sector farmers are thus condemned to a low income
by the inadequate size of their holdings. Second, in many LDCs land
distribution coincides with the distribution of political and economic
power. Feudalism is characterised not only by the high concentration
of landownership but also by government by the landowners. Thus
although the principal motivation for land reform may appear to
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 19

derive from a mass desire for greater equality in the distribution of

land, a shift in the locus of political power away from the traditional
large landowners is usually also implied.
Empirical evidence of inequality in the distribution of agricultural
land is available from a wide cross-section of LDCs. But, due to
incomplete or defective land registration in many countries, most of
the evidence relates not to the distribution of landowner ship rights per
se, but to the distribution of cultivation rights. Although it has been
held that the only way of ensuring access to land is to own it (Griffin,
1979, p. 18) this distinction may be judged not to be of major
importance. Until more precise statistics become available we can
only presume a reasonably close correlation between landownership
distributions and distributions of cultivation rights. However, in
countries where the landlord-tenant system is widespread, it is to be
expected that ownership will be more concentrated than the right to
cultivate, because large estates are typically sub-divided into many
separately tenanted holdings.
Statistics on the distribution of cultivation rights in India reveal
that, in 1970-71, approximately half of all farm occupiers were on
holdings of less than 1 hectare. But such holdings accounted for less
than 10 per cent of the total holdings areas of the country. At the
upper end of the holding size scale, approximately 15 per cent of
occupiers on holdings of 4 or more hectares controlled more than 60
per cent of the total area (See Table 2.1). These statistics exclude
landless rural households, estimated at 22 per cent of all rural
households in India in 1959. Similar statistics on the size distribution
of cultivation rights are available for Kenya, Pakistan, Bangladesh,
Nigeria (Kano state), Colombia and Ecuador (ILO, 1972, Table 2;
1977, Tables 7 and 53; Morris, 1971, Table 3; Griffin, 1974, Table 5.7;

Table 2.1: Size and distribution of operational holdings in India, 1970-1

Holding size Number of holdings Holdings area

(hectares) % /°

Less than 1 50.6 9.0

19.0 119
1-2
2-4 15.2 18.2
11.3 29.7
4-10
3.9 30.9
10 +
All sizes (000s) 70493 —

Source: Compiled from B. Dasgupta (1977), Table 6, who cites the Indian Agricultural Census,
1970-1.
 20 AGRICULTURE AND ECONOMIC DEVELOPMENT

1979, Table 4.4). Statistics on the distribution of land-ownership

include India (Tamil Nadu State), Morocco, Gautemala and
Turkey (ILO, 1977, Table 44; Griffin, 1974, Tables 2.4, 4.6 and
pp. 243-4).
It is important to recognise that the degree of concentration in the
ownership and control of agricultural land is not uniform amongst
countries. The highest concentrations appear to exist in those Latin
American and Asian countries where feudalism in the countryside
still prevails.1 Concentration is lower where redistribution has
already taken place through land reform. Although land reform
tends to be associated with socialism it has also occurred in non¬
socialist countries such as Taiwan. Land concentration is also lower
in countries where traditional systems of communal landownership
still prevail, as in parts of Africa (World Bank, 1975, pp. 243-8).
Contrary to what might be expected, a priori, the degree of land
concentration does not appear to be directly correlated with density
of population. Restricted access to land for the majority whose
livelihood depends on it is certainly not confined to densely
populated countries, as exemplified by the highly concentrated
patterns of landownership and control in a number of Latin
American countries with quite sparse populations (World Bank,
1975, Annex 1, Tables 1.2 and 1.9).
A further aspect of inequality in access to agricultural land in
LDCs is that rents may be distorted in favour of a few large-scale
farmers and against large numbers of small-scale operators. Whereas
monopsonistic modern sector farmers with preferential access to the
land market can rent land relatively cheaply, small-scale farmers in
the traditional sector must pay more than the competitive equili¬
brium rent due to their weak bargaining position vis a vis a powerful
landlord. A parallel distortion occurs in capital markets if, having
allowed for differential risks and transactions costs, modern sector
farmers can still borrow more cheaply than those in the traditional
sector. The contention that the prices of land and capital are in fact
raised against small farmers is backed by empirical evidence from a
number of LDCs (Griffin, 1979). A theory of ‘inter-locking factor
markets’ has been advanced to explain the dominance of large
landowners over small-scale peasant farmers and landless labourers
in poor agrarian economies (Bardhan, 1980). The structure and
conduct of the markets for agricultural land and capital are discussed
at greater length in chapter 8. But due to variation amongst
developing countries in the extent of economic and social inequality
between farmers and landowners in the traditional and modern
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 21

sectors, further empirical investigation is needed to establish the

extent of price distortions in factor markets.
The implications of restricted access to land and distorted factor
prices for small-scale traditional sector farmers in LDCs are clear.
Because they start with so little land and the costs of acquiring
additional land and capital are so high (even if supplies are available)
farm production and farm income must both be low. Thus the
problem of low farm income arising from technological stagnation,
as discussed in the previous section, is compounded.

3 The Farming Environment: Natural Hazards and

Economic Uncertainties

We have already discussed the risks of agricultural production, and

the risk-averse behaviour of peasant farmers, in relation to the
adoption of new agricultural technology. Before concluding this
chapter we refer briefly to the nature and causes of risk and
uncertainty in agriculture, and their effects on the behaviour of
farmers in a broader context.
Agriculture is profoundly affected by an unstable natural environ¬
ment as well as by major economic uncertainties. The biological basis
of agricultural production, and its exposure to the elements, pose
special problems in attempting to forecast yields; these are affected not
only by extremes of weather but also by damage caused by pests and
diseases against which farmers have only limited defences. Most of
LDC agriculture is situated in the tropics where extremes of weather
(temperature, wind and rainfall) are greater than in temperate areas.
The main economic effect of an unstable physical environment is yield
uncertainty. Although complete crop failure can occur under the
most extreme conditions of wind, flood or drought, the risk of this
happening to a particular farmer may not be all that high. But the risk
of a low yield is considerably greater. Whether his crop is substan¬
tially reduced in size or is lost altogether, the consequences for the
subsistence farmer can be extremely serious. Unless he can buy,
borrow or beg food from another source he and his family risk
starvation. For the semi-subsistence or commercial farmer intending
to produce a market surplus, the risk of starvation may be lower,
though the loss of money income due to crop failure or a low yield can
have serious repercussions for the purchase of food and other
necessities. Although production uncertainty can rarely be
eliminated at a feasible cost, it can be reduced by the avoidance of
 22 AGRICULTURE AND ECONOMIC DEVELOPMENT

r unnecessary risks. In view of the uncertainty of their natural

environment it is quite rational for farmers in traditional agriculture,
who are mostly very poor, to be risk-averse.
A major source of economic uncertainty per se is price uncertainty,
ij Agricultural market prices are inherently unstable due to short-term
1 shifts in aggregate demand and supply which farmers are powerless to
control. The root cause of price uncertainty is that farmers are unable
to predict price changes. Although market prices tend to be inversely
correlated with levels of aggregate supply - prices being high during
shortages and low during gluts - the individual producer has little
reason to be confident that the market price will happen to be
favourable in those years in which he experiences a bad harvest. The
^ causes of crop failure can be quite localised, so that a bad harvest in
one region may coincide with more than normal production else¬
where in the market supply area. So, despite the individual’s
misfortune, aggregate supplies can be normal or even above normal,
with the obvious implication for the level of market price, all other
things remaining equal.
Market prices are also affected by shifts in demand. A shortfall in
the aggregate supply of a particular product will not necessarily cause
its price to rise, provided adequate supplies of an acceptable (to
consumers) substitute product are available. Cross-elasticities of
demand are likely to be relatively high for non-staple foodstuffs, such
as fruit, vegetables and some livestock products, though possibly less
so for staples such as rice, millet, maize and wheat. A further
complication is that in an open economy market prices are affected by
shifts in imported supplies and export demand, as well as by domestic
market forces. This adds to the price uncertainty facing the producer.
Price variability and price uncertainty are exacerbated by poor
communications. Due to the lack of a cheap and efficient communi¬
cations system agricultural markets in LDCs tend to be narrow,
particularly in the remoter regions. Farmers are aware of local prices,
but it is difficult for them to get reliable information about prices in
more distant markets and, in any case, it may not pay to send supplies
there. Transport and handling costs, including wastage in transit, are
uncertain. Thus there are serious obstacles to reducing local gluts and
shortages of produce, and to the smoothing of price disequilibria
between markets by engaging in arbitrage.
Because of agriculture’s competitive structure - many predomi¬
nantly small producers producing mostly perishable and relatively
homogeneous products - farmers are price takers. Unlike non-
agricultural producers in oligopolistic industries, they are unable to
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 23

set their selling price. Moreover, due to the perishability of many

agricultural products and the burden of storage costs, there is little
scope for accumulating stocks to bridge periods of low prices.
Price stabilisation schemes to damp price variability and price
uncertainty exist in some LDCs for some commodities. Such schemes
are usually operated directly by government or by a producers’
marketing board with delegated compulsory powers. But for reasons
of administrative convenience and economy, minimum prices tend to
be fixed at the wholesale or retail level rather than at the farm level.
Thus small producers and those without direct access to wholesale or
retail markets for other reasons, such as geographical isolation from
urban areas, may be denied the benefits of price stabilisation due to
their weak bargaining position vis a vis the intermediaries to whom
they sell. Moreover, price stabilisation is generally restricted to a
small number of commodities, particularly those produced for
export. A more extended discussion of price stabilisation schemes
appears in chapter 10.
Price uncertainty compounds yield uncertainty and is an ad¬
ditional reason for risk-aversion amongst traditional sector farmers
producing for the market. The expectation that peasant farmers
might tend to prefer enterprises with low yield and price variances to
more risky but possibly more profitable enterprises, is confirmed by
empirical evidence (Wolgin, 1975).

4 Summary

Poverty is the outstanding characteristic of traditional agriculture.

The causes of poverty are resource constraints and technological
stagnation, both of which limit production. Whilst labour is generally
abundant, with work-sharing the norm on family farms, land is
extremely scarce due to population pressure and the unequal
distribution of cultivation rights. Moreover, access to capital is very
restricted due to low internal saving (linked with low farm income)
and capital market imperfections. Farming methods and equipment
are primitive by western standards, but technological progress is
hampered by a lack of appropriate small-scale alternatives, barriers
to the spread of technological information, the risks of adoption and
input supply bottlenecks.
The scarcity of land encourages the intensive use of labour,
especially family labour, to obtain high crop yields: but the
productivity of labour is reduced. The problem of low labour
24 AGRICULTURE AND ECONOMIC DEVELOPMENT

productivity in traditional agriculture is exacerbated by technological

stagnation. Although wages in the modern sector tend to be higher,
due to labour market dualism, the number of workers in traditional
agriculture remains abundant due to rural isolation and constraints
on rural-urban migration, such as the risk of prolonged unemploy¬
ment in the urban sector. A low level of per capita income in
traditional agriculture is an inevitable consequence of low labour
productivity.
The distribution of land in LDCs tends to be distorted in favour of
a minority of landlords and large-scale farmers. Most rural house¬
holds are restricted to only a very small plot of land, or none at all in
the case of landless labourers. The distribution of wealth and income
in the rural sector is directly correlated with the distribution of land.
A majority of peasant farmers are further handicapped in the capital
market because of large farmers’ preferential access to cheap credit.
These distributions in access to non-land resources greatly exacerbate
low labour productivity and farm income for the majority of
households in traditional agriculture.
Due to agriculture’s biological basis, and an unstable natural
environment, forecasting agricultural production is particularly
hazardous. Yields are therefore very uncertain, particularly at the
level of the individual producer. Farmers producing for the market
are also confronted by major price uncertainties. Because of their
poverty, farmers in traditional agriculture may tend to place security
before profit maximisation. Those who depend on agriculture for their
survival are exposed to numerous inescapable hazards, such as
drought, flood, hurricane and damage by wild animals and pests: it
is wholly rational should they choose not to expose themselves to
further avoidable risks such as growing an unfamiliar crop by an
untried method for an unknown market.

References

Bardhan, P.K. (1980), ‘Interlocking factor markets and agrarian

development: a review of issues’, Oxford Econ. Papers , 32 (1).
Dasgupta, B. (1977), Agrarian Change and the New Technology in
India, UNRISD, Geneva.
Eicher, C.K. and Witt, L.W. (1964), Agriculture in Economic
Development, McGraw-Hill, New York.
Griffin, K. (1976), Land Concentration and Rural Poverty, Mac¬
millan, London.
 STRUCTURE AND CHARACTERISTICS OF AGRICULTURE IN LDCS 25

Griffin, K. (1979), The Political Economy of Agrarian Change,

Macmillan, London (2nd edn).
International Labour Office (1972), Employment, Income, and Equ¬
ality: A Strategy for Increasing Productive Employment in Kenya,
Geneva.
Griffin, K. (1977), Poverty and Landlessness in Rural Asia, Geneva.
Morris, J. (1981), ‘Agrarian structure: implications for
development - a Kano case study’, Oxford Agrarian Studies, X,
pp. 44-69.
Norman, D.W. (1977), ‘Economic rationality in traditional Hausa
dryland farmers in the north of Nigeria’, in Stevens, R.D. (ed.)
Tradition and Dynamics in Small-Farm Agriculture, Iowa State Univ.
Press, Ames, Iowa.
Ritson, C. (1973), ‘A framework for analysing the contribution of
the agricultural sector to economic development’, J. Agric. Econ.
XXIV(l).
Schultz, T.W. (1964), Transforming Traditional Agriculture, Yale
University Press, New Haven.
Stewart, Frances (1978), Technology and Underdevelopment, Mac¬
millan, London, 2nd. ed.
Todaro, M. (1969), ‘A model of labour migration and urban
development in less developed countries’, Amer. Econ. Dev.,
pp. 138-48.
Wolgin, J.M. (1975), ‘Resource allocation and risk: a case study of
smallholder agriculture in Kenya’, Amer. J. Agric. Econ., 54(4)
World Bank (1975), The Assault on World Poverty, Johns Hopkins
University Press, Baltimore, pp. 187-261 (Land Reform).

Notes

1. The term ‘feudalism’ is admittedly ambiguous. Following Warriner, in

Eicher & Witt (1964), we use it loosely merely to denote any system of
large land ownership giving rise to rigid social stratification and economic
inequality in favour of the land-owning class and detrimental to the
economic and social welfare of a large subservient non-landowning class.
3 Role of Agriculture in
Economic Development

1 A Framework of Analysis

Following the classic analysis of Kuznets (1961) the agricultural

sector in LDCs may be seen as being potentially capable of making
four types of contribution to overall national economic growth and
development.
(1) Expansion of the non-agricultural sector is strongly reliant on
domestic agriculture, not only for a sustained increase in the
supply of food, but also for raw materials used in manufacturing
products such as textiles. Kuznets terms this the ‘product
contribution’.
(2) Because of the strong agrarian bias of the economy during the
early stages of economic growth, the agricultural population
inevitably forms a substantial proportion of the home market
for the products of domestic industry, including the market for
producer goods as well as consumer goods. This is termed the
‘market contribution’ by Kuznets.
(3) Because the relative importance of agriculture in the economy
inevitably declines with economic growth and development,
agriculture is seen as a principal source of capital for investment
elsewhere in the economy. Thus the development process
involves the transfer of surplus capital from agriculture to the
non-agricultural sector. Similarly, development also entails the
transfer of surplus labour from agriculture to non-agricultural
occupations, especially over the long term. Kuznets terms
these agriculture’s ‘factor contributions’.
(4) Domestic agriculture is capable of contributing beneficially to

26
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 27

the balance of overseas payments either by augmenting the

country’s export earnings or by expanding the production of
agricultural import substitutes. This ‘foreign exchange contri¬
bution’ is not explicitly identified by Kuznets but is implicit in
his market contribution.

2 Product Contribution

An expression derived from Kuznets (1964) shows how the agricul¬

tural sector’s share of GDP growth is related to the product of
agriculture’s initial share of GDP and the relative rates of growth of
agricultural and non-agricultural net products. Define:

Pa = agricultural net product

Pn = non-agricultural net product
P = total national product
Then:

(1)
and

(2)

Writing ra for (SP^P,, rn for <5Pn/Pn:

(3)

(4)

and (5)

Substituting for <5P on the RHS of equation (5) from equation (3):
28 AGRICULTURE AND ECONOMIC DEVELOPMENT

1
= (Para+Pnr„)/Para

=---... (6)
1 + Pnrn/Para
Kuznets’s formula expressing an inverse relationship between
agriculture’s share of GDP growth (Para/<5P) and the product of the
ratio of sectoral shares of GDP (P„/Pa) and the ratio of sectoral
growth rates (rn/ra), is given by equation (6).
In Table 3.1 empirical estimates of Para/(5P in selected countries at
different stages of economic development are presented. These have
been derived from data published by the World Bank (1980), using
Kuznets’s formula. Separate estimates are shown for 1960 and 1978
corresponding with the available data on sectoral shares of GDP
(Pa, Pn) for those two years. The data on sectoral growth rates (ra,rn)
are trend estimates based on the periods 1960-70 and 1970-78.
Overall growth rates during these periods are shown by countries on
the extreme RHS of the table.
The empirical evidence on the A sector’s share of the growth rate
(Para/<5P) is in broad agreement with a priori expectation. Thus the
evidence confirms that:

(1) the share declines over time as a consequence of economic growth

and development, as shown by the comparative values of Para/^P
in 1960 and 1978 in particular countries; and
(2) the share is inversely correlated with the level of economic
development measured in terms of GNP per capita, as shown by
cross-sectional comparisons of Para/<5P amongst countries at
different stages of development at a particular time, such as
1960 or 1978.
So, for example, between 1960 and 1978, the A sector’s share of the
growth rate fell both in Bangladesh, a low income country, from 33 to
22 per cent, and in Japan, an industrialised country from 5 to 1 per
cent. In Egypt, a middle income country, the share fell from 20 to 11
per cent over the same period. Although there is clearly considerable
diversity amongst countries, both in the rate at which the A sector’s
share of the growth rate declines over time and in the size of the share,
even amongst countries at a similar level of GNP per head, average
values for low income, middle income and industrialised countries
(shown at the top of the table) suggest that these relationships are
reasonably systematic.
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 29

Op
Os. VO r** 04 Ov O' tt 00 00 vO Tto 00 © O' ©
1 oj co 04 m m s 04
3
g o o O O o o O O O O o o o O © © ss
Os d d d d O o d d d d d d d d © © © ©
<o 1

s-o 1 £5
Os Os _ vO vO VO 0- SO) so) 04 04 04 m Ov so) ov m
1 m s 105 CO m vO m 00 Tt SO) h SO) 04 © Tt Tt
$a © o o o o s O p S o O O p O © © ©
<> © d d d d d d d d d o d d d © © © ©
**•.

m OO
R <vi Os 00 O' so) so. o vC SO)
•“ 04 8 04 O
04 ©
Os 04 o © 04 04 04 *—■ —~ p g © © ©
© d d d d d d d d d d d d d © © © ©
win.
04 m
O ©
co m o vO m ■'t ov o O' SO) Tt Ov O' m SO) © ©
Os m — o so) 04 04 rn 04 04 04 04 04 O o © © © ©
■"ni © d d d d d d d d d d d d d © © © ©

| 04 © vO 04 00 vO p so) 04 SO) p rn VO p SO) © — p

05 04 04 rn co — — rn oj rn O'* 04 vb oi SO) oi rn

Os

c
W g
1 O Os SO) tv- oo oo vO oo oo oo Ov O' O p m © O'
o> 04 't — 04 — —’ 04 — — d —* 04 04 — oi oi *o)
>0) —
s
■—

VO rn p oo •O) Os — 04 Tt o^ vq 04 — rn © © © rn
Os
_.* SO) tt d —‘ oi 04 rj 04 Tt 04 OO rn Ov Ov Ov 04
OJ r~“ Tt •— m
C

a.' © O •O) o- vO O Os 04 Os m SO) rn Ov m O' © O' — ©

co SO) d d 04 rn «o vb Tt vb © Tt'
s
04 — 04

R m vO vO O' O' ov Tt O' rn Tt O © SO) Os 04

i 5 vO CO •o, 3 m SO) O' g oc O' vO SO) m 04 SO) Tt CO
©
o.
o O O O o o o o o O O O p © © © © 3

Os o d d d d d d d d d d d d d © © © ©
■* •*«. £
J=
<P
Ow © SO) Tt vO m rn Os — 04 oo o OO O' vO Ov m 04 O'
1 SO) vO SO) f •O) SO) O' O' SO, o Tt SO) O' SO) 04 SO) Tt
o © o O o o o o p o p o p O o © © ©

$ © d d d d d d d d d d d d d © © © ©
•'«.

00
tv. o o sO vO m O' —* VO f— Os — SO) OO •— m ov
1 04 to 04 04 ? m SO) 04 o 04 ©
o> o p p © O O o o o o g s O O g © © ©
tv
© d d d d o d d d d d d d © © © © ©
o>
»•» ■***
w° o>
tv SOI rf 04 O' Os o Os SO) Ov SO) r- •O) 00 OO m © — m
04 04 ■*t
©
1 04 n
O o ©
© d d
04
o o o
d d d
m •<t 04
p O
d d
04 SO)
o O s
o d d d d
m m
o O © © © gg
© © © © ©
|a
•*•*
<

oc 04 Tt sO m SO) 00 Ov m 04 Ov O' m 00 SO) SO) O'

tv SO 00 O' tt s vO vO SO O' r- OO vC 00 Ov Os Ov Ov Ov
•*«■ © d d d d d d d d d d d d © © © © ©

a.' s£> —
O o oo Tt Os o sO •t VO o Ov O' VO Tt O' vO O'
g
Os
SO) O' Os ro SO) vO SO) •St O'; vO O' vO OO oo Ov OO Ov Os
d d d d d d d d d d © © © © ©
'•«. d d d
a £
p_- ,
oo oo sC Tt O' o SO) 04 Ov O' oo O' 04 SO) soi m
—- © © © © ©
tv rn o SO) Tt m m m 04 04 —•rn
Os
d d d d o d d d d d d o d © © © © ©
oj3 m Tt vO m m
o o 04 vO o tT vO •^t o a
g SO) 04 o vO SO) m SO) m s 04 rn —■ S © S
©
Os
d d d o d O d d d o d d d © © © © ©

c aa
*^3
lu §2
c
a
-
a. ^
<u
E •T3
8>J
S£
gpS: £
<u o 15
-C

•§
J to v-

S
i
o
(J
c
c
"cd
•C
C/5
15
•O
cd
cd

C
c
-2
.2
15
c
-o o
c
ct
IS o
cd 8 £ o C
>N
cd
la
O’ § R C53 </) p
s* M ' -5 Os t3
T3 TJ 3
c •6 J IS TD
00 .2 o o.
pn to O
u >< CT3 4B
UK

D. O C/2
o 15

(J.S o ^ •S
o
j 2 C
3
CQ c
c cd C
0-
0£ -C
W H s U s
3 cd
•“> Z 3 II
30 AGRICULTURE AND ECONOMIC DEVELOPMENT

The magnitude of any change in the ratio of non-A sector to A

sector product is determined in part by the difference in sectoral rates
of growth, thus:

where the superscripts relate to time

^Pn°(l+rn) Pn°
(8)
Pa°0+ra) Pa°‘"
P,°[(l+r„)-(l+r.)]
(9)
Pa°(l+ra)

PnVn ~ Q (10)
P.°0+ra)-
Two major deductions follow from equation (10). First, the
magnitude of the change in sectoral shares depends upon:

(i) the initial ratio (P°/Pa°);

(ii) the agricultural growth rate (ra);
and (iii) the difference in sectoral rates of growth (rn — ra).

Other things being equal, A(Pn /ra) will be ‘large’ (‘small’) if either:

(a) P„/Pa is large (small)

or (b) ra is small (large)
or (c) rn — ra is large (small).

Second, the sign of the change in sectoral shares depends only upon
the sign of rn — ra: rn — ra > 0 signifies an increase in the non-A sector’s
share of the total product, whereas rn — ra < 0 signifies a decreasing
share.
The highest rates of decline in the relative importance of agricul¬
ture in the economy tend to be associated with the combination of a
relatively high initial product shares ratio in favour of the non-A
sector, a relatively low rate of agricultural growth and a relatively
high rate of non-agricultural growth (making for a large positive
difference in favour of the non-A sector). A specific example of this
type is Japan where the A sector’s share of GDP declined from 13 per
cent in 1960 to only about 5 percent in 1978 (Table 3.1). Conversely,
the lowest rates of decline in the A sector’s relative importance tend to
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 31

be associated with a low initial ratio in favour of the non-A sector

combined with a relatively high rate of agricultural growth and a low
rate of non-agricultural growth. This situation is exemplified by Sri
Lanka where the A sector’s share of GDP marginally increased from
34 per cent in 1960 to 35 per cent in 1978.
In a developing economy, where per capita incomes are rising,
growth in the A sector can be expected to lag behind non-A sector
growth for three reasons. First, the demand for food and other
agricultural products is generally less income-elastic than the demand
for non-agricultural products, due to the Engel effect. Second, due to
scientific advances and associated technological innovations in
agriculture farmers become increasingly reliant on inputs purchas¬
ed from the non-farm sector of the economy: this is termed the chang¬
ing resources structure of agriculture effect. Third, because the
demand for off-farm marketing services - distribution, storage and
processing - is more elastic than the demand for agricultural pro¬
ducts at the farm gate, the farmer’s share of food expenditure at retail
prices declines with time (the urbanisation effect).
In recognising the development trends making for the declining
relative importance of the agricultural sector in the long term, it is
necessary to avoid the trap of overlooking the critical importance of
domestic agriculture’s product contribution to the maintenance of an
adequate rate of economic growth in the short term. This is the trap
that a number of developing countries have fallen into in opting for a
strategy of rapid industrialisation without parallel development in
agriculture.
There are two basic reasons why, in most LDCs, development
based on structural diversification of the economy is constrained by
the rate of growth in the marketed output of domestic agriculture.
The first reason is that the domestic farm sector is an important
source of raw materials for use in industries such as textiles and food
processing, as well as being the principal source of food for
consumption by growing numbers of non-food producers employed
in industry. Secondly, as agriculture becomes more and more closely
integrated with other sectors of the economy - due to the changing
resources structure and urbanisation effects - the multiplier effects of
increased agricultural production and incomes assume an increasing
importance in relation to the growth in demand for the products of
domestic industry and the associated demands for labour and other
industrial inputs. The multiplier effects of domestic agricultural
expansion are discussed later under agriculture’s ‘market contribu¬
tion’ where the inter-industry linkage effects of the adoption of
32 AGRICULTURE AND ECONOMIC DEVELOPMENT

modern agricultural technology in LDCs are analysed. In the

meantime we undertake a closer examination of the direct food and
raw materials contribution.

2.1 Food contribution

In most developing countries the domestic agricultural sector is the

principal source of food for consumption by non-agricultural
workers. Diversification of the economy is therefore contingent upon
domestic food producers producing a surplus, in excess of their own
subsistence, which is large enough to feed a growing number of non¬
food producers. Although, in principle, shortfalls in domestic food
supplies can be made good by expanding food imports, in practice
such imports are frequently severely constrained by the scarcity and
high cost of foreign exchange. Unlike imports of capital goods, food
imports are consumed and do not augment the capital stock. So,
where a choice has to be made between importing food and importing
capital goods, the opportunity costs of food imports may be very
high in terms of lower investment and a consequently reduced rate
of economic growth.
The marginal opportunity costs of domestic food supplies are
derived from the opportunity costs of the additional agricultural re¬
sources employed. In countries where the principal agricultural inputs
are land and labour, rather than capital, and where alternative em¬
ployment opportunities for agricultural workers are very scarce, these
opportunity costs will tend to be ‘low’. Thus, in LDCs, there is
typically a marked contrast between the high opportunity cost of
food imports and the relatively low marginal opportunity of costs of
domestic food production. However, an adequate rate of domestic
agricultural expansion is often difficult to achieve in practice due to
institutional and policy constraints, as already mentioned chapter 2,
and discussed at more length in subsequent chapters. Meanwhile, we
consider how industrialisation and urbanisation affect the demand
for food and possible consequences of failing to match the increased
demand with a corresponding increase in supply.
As industrialisation and urbanisation proceed, the rate of increase
in the off-farm demand for food tends to exceed the rate of growth in
industrial employment for two reasons: first, because the real
earnings of industrial workers are usually higher than those of
agricultural workers and, secondly, because in LDCs the income-
elasticity of demand for food is relatively high. So the migration of
workers from agriculture to higher paid industrial employment may
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 33

cause a relative scarcity of food unless the productivity of those

remaining in agriculture rises fast enough to provide the migrants
with a higher level of per capita food consumption than previously.
Failure to expand food supplies quickly enough to meet the growing
demand inevitably results in higher prices, if market forces are
permitted to operate, or rationing in some other form if government
intervenes to prevent prices from rising. Whether rising food prices
can actually cause general price inflation is a question to which we
shall return shortly. But, whatever the answer is, there is no doubt
that due to the primacy of food as a wage good, as reflected by the
high ratio of food expenditure to total household expenditure in
LDCs, rapid food price inflation frequently leads to serious social
and political instability which is inimical to economic growth. Apart
from the untoward indirect effects of social and political unrest on the
rate of industrial development, the principal economic effect is that,
because rising food prices increase the pressure for employers to
concede higher wages which are unrelated to productivity, industry’s
terms of trade deteriorate. Hence, due to falling profits and lower
reinvestment there is a drastic decline in the rate of industrial
development.
Where consumer incomes are relatively high, there is scope for
substituting less costly types of food for more expensive ones as food
prices in general rise. However, in LDCs the majority of consumers
are relatively poor, their diet consisting mainly of the cheapest
possible starchy foods (grains or tubers). The price elasticity of
demand for these foods is typically very low and if their price rises, the
quantity demanded is little affected due to the absence of less
expensive substitutes. Thus, in an LDC, a substantial rise in the price
of a staple foodstuff can cause a substantial loss of real income for
large numbers of non-food producing consumers. Conversely, if
staple prices fall, due to a technologically induced shift in supply, the
real incomes of food consumers will rise. Thus, in contrast to the
consequences of lagging food production, success in closing the gap
between production and the growing demands of industrial and
urban consumers will remove a major obstacle to progressive
industrial development. Moreover, lower food prices will leave
relatively more in the pockets of urban consumers for the purchase of
industrial goods, as well as making such goods more competitive on
external markets.
34 AGRICULTURE AND ECONOMIC DEVELOPMENT

2.2 Food supplies and inflation

There are a number of arguments linking lagging food supplies with

inflation in LDCs. First, because in LDCs supply bottlenecks are
more characteristic than underutilised productive capacity due to
deficient demand, the effect of income augmenting policies, such as
employment creation, may be excess demand and inflation, rather
than higher real output. In countries where food consumption
accounts for a high proportion of total household consumption, a
correspondingly high proportion of excess demand is likely to relate
to food. Second, because agricultural product prices are largely
determined by market forces, and because the demand for most food
products is relatively price-inelastic, excess demand for agricultural
and food products generally results in steeply rising prices. Thirdly,
although industrial product prices tend to be ‘administered’ by firms
rather than being determined by the market, they are usually
inflexible downward (due to trade union resistance to wage reduc¬
tion, for example). However, because higher food prices are likely to
result eventually in higher industrial wages, industrial product prices
can also be expected to rise after a time-lag. With food and industrial
product prices both higher, a rise in the general price level (i.e.
inflation) is virtually inevitable. Thus a process of cost-push inflation
may be visualised in which lagging supplies and rising prices of food
provide the initial impetus, but other prices follow after a time-lag
(Maynard, 1961; Chakrabarti, 1977). Because of the time-lag,
agricultural prices may appear to ‘lead’ other prices at one stage of
the inflationary process whereas non-agricultural prices appear to be
leading at a later stage. If farmers purchase substantial amounts of
industrial inputs, such as fertilisers and machinery, and in countries
where they can also claim compensation from the government for
increased costs of production (as under schemes of minimum
guaranteed prices for farm products), this gives a further twist to the
inflationary spiral (Hathaway, 1974).
However, even in LDCs it is naive to attribute inflation to a single
cause, such as rising food prices. Theoretical considerations and
empirical evidence both point to the conclusion that inflation is due
to a number of causes, including government policy variables. In
particular, it seems likely that a country’s rate of inflation is affected
by macro-economic policy decisions on exchange rates, interest rates
and changes in the money supply. Although food supplies and prices
may play a significant role in determining the rate of inflation, in
practice their effect may be masked by these and other factors (Edel,
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 35

1969; Chakrabarti, 1977). Such complications may explain why

comparatively few studies of the influence of food supplies on
inflation in developing countries have been attempted. But two such
studies are worth mentioning.
A study based on a number of Latin American countries in the
1950s and 1960s reached the conclusion that ‘although agricultural
improvement alone would not automatically curb inflation single-
handedly ... eliminating pressure from food prices would lessen the
number of factors which can set off inflationary spirals’ (Edel, 1969,
p. 95).
This conclusion is based on a country-by-country comparison of
the difference between actual rates of growth in domestic agricultural
supply (symbolised by a) and the so-called autonomous rate of
agricultural growth required for price stability (a*). The autonomous
rate of growth is derived from a structural model of food pricing and
inflation, with certain parameter restrictions imposed. The model
postulates inflation if a < a*, and general price stability if a ^ a*.
Empirically, it was found that whereas some countries had experien¬
ced rapid inflation despite an ‘adequate’ rate of agricultural growth,
in others the rate of inflation had been lower despite failure to achieve
a*. Hence the above quoted conclusion to the effect that at least in
Latin America, rising food prices alone fail to explain inflation.
Although this is an interesting study it also contains a number of
conceptual weaknesses. First, inflation is defined simply as an
increase in the relative price of food without formal behavioural
justification. Second, the empirical application of the structural
model is limited by a number of questionable simplifying assump¬
tions, including a closed economy and a constant level of population.
A study of price behaviour in India during the period 1962-6 was
based on a more complete model than the Latin American study
(Chakrabarti, 1977). It was postulated that the overall rate of
inflation is functionally related not only to changes in the prices of
both agricultural and non-agricultural goods, but also to the rate of
growth in the money supply.1 The empirical findings point to the
conclusion that, in India, rising domestic food supplies and falling
prices have exerted a downward pressure on inflation (as expressed by
a general price index), whereas rising industrial production and, at
best, static prices had the opposite effect. Given the generally low
price-elasticity of demand (or the high coefficient of price flexibility)
for food grains, which apparently exists despite the extent of poverty
in India, any increase in aggregate industrial output and incomes is
bound to result in steeply increasing agricultural prices, all other
36 AGRICULTURE AND ECONOMIC DEVELOPMENT

things being equal. The results of this study point to the conclusion
that given the upward movement of prices built into industrial
expansion, a continuous fall in the relative price of basic foodstuffs is
necessary to permit overall price stability to prevail in the long run.
Thus, given a static or rising demand for food, real food prices will
show a declining trend only in response to a sufficiently rapid upward
trend in output. This implies that, where containing inflation is a
major policy objective, inducing the agricultural sector to produce
more for the market, despite declining prices, must be a central
objective of government agricultural policy. The attainment of this
goal may well demand measures to enhance productivity in agricul¬
ture, especially of labour, so that farm incomes may be maintained,
or even improved, despite falling product prices. Since, for numerous
reasons discussed elsewhere in this book, it may not be feasible to
achieve an adequate rate of agricultural productivity gain in practice,
a government intent on a cheap food policy may need to resort to
redistributional measures such as food price subsidies.
Like all partial analytical studies, the study of Indian price
behaviour is unrealistic in the sense that it concentrates on but one
aspect of a more general problem. In practice, the achievement of
price stability may not always be of paramount importance to
government policy-makers. If a measure of inflation is the price that
has to be paid for attaining a target rate of economic growth, or even
for achieving the more limited objective of ‘getting agriculture
moving’, it may be decided that this price is worth paying, at least for
a limited period.
There is also some empirical evidence supporting the view that
inflationary expectations are influenced by historic rates of food price
inflation (Holden and Peel, 1977). This evidence comes from the UK,
a developed country: it might be expected that in developing
countries, where food accounts for a higher proportion of the
consumer’s total budget, this association would tend to be even more
pronounced. To the extent that there is a direct link between actual
rates of inflation and inflationary expectations, the policy implication
is that anticipatory measures to contain food price rises may
contribute towards the achievement of greater general price stability.

2.3 Raw materials contribution

In many LDCs the early stages of industrialisation are marked by a

predominance of industries based on the processing of agricultural
raw materials. One possible measure of the relative importance of a
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 37

country’s agro-industries is food and agriculture’s share of value

added by manufacturing industry. Statistics published by the World
Bank show that, within a group of 17 ‘low income’ countries
(<US$390 GNP per capita in 1978) in the mid-1970s, this share
ranged from 95 per cent (Madagascar) to 17 per cent (Sri Lanka)
with a median value of 46 per cent (Pakistan); within a group of
43 ‘middle income’ countries (US$390 to US$3500 in 1978) the
equivalent shares were 92 per cent (Nigeria), 7 per cent (Singapore)
and 41 per cent (Morocco). In contrast, within a group of 18
‘industrialised countries’ (US$3500 to US$12,100) the comparable
figures were 31 per cent (Ireland), 8 per cent (Japan) and 14 per
cent (UK, Canada and Norway). However, for the purpose of
compiling these statistics, only the food manufacturing, beverage
and tobacco industries (ISIC Major Groups 311, 313 and 314)
were defined as agro-industries. Because industries like textiles (ISIC
Major Group 321) and footwear (ISIC Major Group 324) are
excluded, these figures underestimate the relative importance of
industries using agricultural raw materials. Actual figures to indicate
by how much the inclusion of these additional categories increase
the relative importance of industries that are importantly dependent
on agriculture for their raw materials are regrettably unavailable.
However, it is clear that in some LDCs, notably those with industries
based on the domestic production of natural fibres, hides and skins,
the exclusion of these categories results in serious underestimation
of the relative importance of agro-industries.

2.4 Production linkages between agriculture and other industries

As discussed at the begining of this chapter, there is a ‘natural’ long¬

term tendency for the A sector’s relative share of GDP to decline as
real GDP per capita rises. However, because the income-elasticity of
demand for the value added to agricultural raw materials by agro¬
industries is usually higher than for the raw materials themselves, the
much broader based ‘agri-business’ sector, including the food
manufacturing, clothing and footwear industries as well as farming
per se, declines at a much slower rate.
This broader view of agriculture’s product contribution via related
industries leads on to an application of the concept of inter-industry
linkage (which describes inter-sectoral inter-dependence). Formally
expressed, inter-sectoral production linkage measures the affect of an
autonomous increase in final demand for the product of a given
industry not only on the output of that industry and the outputs of
38 AGRICULTURE AND ECONOMIC DEVELOPMENT

industries supplying the first industry with inputs, but also the
outputs of yet other industries ‘supplying the suppliers’ in a second
round of transactions. The second round leads to a third, and so
on, for an indefinite number of further rounds. The first round of
linkage effects are sometimes termed the direct effects to distinguish
them from the indirect effects represented by the second and
subsequent rounds.2
Distinctions may be made between backward, forward and total
linkages. Backward linkage measures the ratio of intermediate input
purchases from other industries to a particular industry’s total value
of production; forward linkage measures the ratio of intermediate
output sales to other industries to a particular industry’s total sales
(including sales to the final consumer); total linkage is the sum of
backward and forward linkages (Yotopoulos and Nugent, 1976,
ch. 16). Thus a hypothetical industry which purchased nothing from
other industries and sold the whole of its output directly for final
consumption would register zero coefficients for its backward,
forward and total production linkages. In practice, even in develop¬
ing countries, virtually all industries add value to some quantity of
purchased inputs, as well as selling a proportion of their output in the
form of intermediate output. Hence the direct linkage effects are non¬
zero and positive: the sum of the direct and indirect effects is even
greater, due to the multiplier effect of successive rounds of inter¬
industry trade.
It has been held that, as a primary industry, agriculture lacks
backward linkages by definition: also that because much agricultural
production goes directly for home consumption or for export,
without intermediate processing, especially in LDCs, forward link¬
ages are likely to be weak (Hirschman, 1958, ch. 6). However, even
in LDCs, the ‘modernisation’ of agriculture entails the growing usage
by farmers of industrial inputs such as fertilisers, pesticides and
simple mechanical aids. Thus, in practice, LDC agriculture is rarely
so primitive as to be entirely without backward linkages with input
suppliers. Moreover, on the downstream side of LDC agriculture, it is
a common observation that industries based upon the processing of
agricultural products such as grain milling, textile manufacture and
fruit and vegetable canning, are often pioneer industries in the
industrialisation process. Empirical evidence exists to support the
view that, although some other industries may have stronger inter¬
industry production linkages, those exhibited by agriculture are not
negligible. For example, total inter-industry linkage coefficients
estimated for various industries in Taiwan, in 1966 were as shown in
Table 3.2:
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 39

Table 3.2: Production linkage coefficients

Industry PL

Agric. production for food 1.90

Agric. production for raw material 1.50
Working capital for agriculture 2.46
Food processing 2.16

Source: Yotopoulos and Nugent, 1976, Table 15.4.

These data indicate that although a marginal unit of agricultural

production (food or raw material) generated less extra production in
linked industries than that generated by a marginal unit of production
in the agricultural supply or food processing industries, the production
linkage coefficients of agriculture as such are nevertheless substantially
in excess of unity (PL = 1 would signify the complete absence of inter¬
industry linkage effects). Although the agricultural supply and food
industries may have somewhat stronger linkages with other indus¬
tries (including agriculture), the development of industries directly
linked with agriculture is clearly dependent on parallel development
in agriculture itself, apart from the possibility of procuring raw
materials from abroad or producing agricultural requisites for
export. However, these latter alternatives may not be feasible or
economically attractive to producers due to the higher risk and for
other reasons (Hirschman, 1958, pp. 99-100).
Inter-industry production linkages are complemented by linkages
with respect to both employment and income generation. We shall
refer to these again under agriculture’s factor and market
contributions.
Empirical estimates of inter-industry linkage coefficients are
general based on an application of the technique of input - output
analysis. In adopting this approach, the usual assumption is that
production, employment and income generation linkage effects occur
in response to autonomous changes in demand. Despite its conve¬
nience, this assumption rules out the initiation of growth from the
supply side. Another limitation of the technique is that it does not
embrace the effects of supply constraints such as input price inflation.

3 Market Contribution
Consider, hypothetically, a closed, single-sector agrarian economy
on the threshold of sectoral diversification. Even though the per
40 AGRICULTURE AND ECONOMIC DEVELOPMENT

capita incomes of those employed in the ‘new’ industries may be

expected to be somewhat above those of farmers, the farm sector,
because of its sheer size, must initially be the major market for
domestic industrial products. Farmers’ expenditures on industrial
goods - both consumer goods (clothes, furniture, household utensils,
building materials) and producer goods (fertilisers, pesticides, tools
and implements) - represent one aspect of agriculture's market
contribution to general economic development (through sectoral
diversification). But, as farmers’ purchases of industrial goods have
their counterpart in inter-sectoral sales of agricultural goods, the A
sector’s market contribution also includes the sale of food or other
farm products to the non-A sector. In discussing these two aspects of
the market contribution, Kuznets (1964) describes the first as
‘marketization of the production process’ and the second as ‘market-
ization of agricultural net product’. Both are accelerated by the
adoption of new agricultural technology. Although this does not
necessarily entail increased farm expenditure on industrial inputs,
many agricultural innovations, some mechanical others chemical, are
of this type. On the income side of the agricultural sector account,
because of the favourable effect of the adoption of new agricultural
technology on the productivity of the farm resources (land and
labour) adoption almost invariably results in higher agricultural
output and a large marketable surplus of farm products.
The effect of dropping the restrictive closed economy assumption
is to break the circular flow of income between domestic agriculture
and domestic industry: domestically produced goods can, in principle,
be sold abroad instead of on the domestic market. This applies to both
agricultural and non-agricultural goods. In a perfectly competitive
world, competitive forces would ensure that the extent and pattern of
each country’s external trade conformed with comparative advan¬
tage. A country specialising in manufactures could export these in
exchange for food imports produced by countries specialising in
agriculture: it could, in principle, meet the whole of its domestic
demand for food with imported supplies. By the same token, a
country specialising in agriculture for export, as well as to meet
domestic market demand, might buy all its manufactured goods
abroad. In reality, of course, international trade is not conducted in
conformity with the norms of perfect competition: competition is dis¬
torted by numerous restrictions upon both the exchange of goods
and the mobility of resources. Moreover, because of the risks of being
wholly dependent on overseas sources for the supply of food or other
‘strategic’ goods, it is not surprising that most countries choose to be
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 41

self-sufficient in some degree, regardless of the extent of specialisation

indicated by considerations of comparative advantage alone. Our
over-simplified hypothetical examples serve to show, nevertheless,
that agriculture’s market contribution remains in the open economy,
but its form is modified.
In a closed economy, the ultimate benefits of sectoral diversifica¬
tion would include the diversification of consumption, i.e. consumers
have a wider choice of goods. In an open, single-sector agrarian
economy exporting agricultural products in exchange for manufac¬
tures, the sectoral diversification is by-passed (or postponed), but the
diversification of consumption remains. The example is artificial
because, in the real world where practically all national economies are
open in some degree, production is never so specialised as to prevent
the diversification of consumption being derived from both forms of
agriculture’s market contribution. The importance of agriculture’s
market contribution in an open economy clearly depends on the
relative importance of the agricultural sector itself, which is typically
‘large’ in LDCs, though generally declining as economic development
proceeds. In order to find a country where agriculture’s market
contribution ‘does not count’ one would have to look for a country
with no agriculture, apart from farmers producing exclusively for
their own subsistence. Despite the existence of a few small, open
economies specialising in the production of light manufactures, such
as Hong Kong and Singapore, it may be doubted whether such a
country exists anywhere in the world.
All other things being equal, the effect of affording protection to
domestic manufacturing industries will be to reduce agriculture’s
market contribution. Because protection of infant industries shifts
the terms of trade against agriculture, agricultural production,
income, savings and investment may all decline. With agriculture
both selling less to and buying less from other sectors, its market
contribution must inevitably decline (Nicholls, 1963). Further aspects
of the benefits of agricultural trade are discussed in a subsequent
section, under agriculture’s ‘foreign exchange contribution’.
Because, due to inter-industry linkage, the expansion of produc¬
tion and income in one sector generates increased production and
income in other sectors, production linkages (as previously discussed)
are complemented by income-multiplier or income-generation link¬
age effects. Moreover, these may be regarded as proxy measures of
the sectoral market contribution to economic growth and develop¬
ment. Further results of the empirical study of inter-industry linkages
in Taiwan showed agricultural production of both food and raw
42 AGRICULTURE AND ECONOMIC DEVELOPMENT

Table 3.3: Income linkage coefficients

Industry YL

Agricultural production for food 1.64

Agricultural production for raw materials 1.80
Working capital for agriculture 1.14
Food processing 1.09

Source: See Table 3.2.

materials exerting stronger income-generation linkage effects than

either the agricultural input supply or food processing industries. The
estimated coefficients are shown in Table 3.3.
Agriculture’s high ranking in terms of income generating capacity
outside the sector reflects the relatively high labour component of
agricultural value added. Additional net output means additional
employment which, in turn, generates additional income, a propor¬
tion of which (depending on the marginal propensity to consume) is
spent on the products of other industries. Because per capita income
in agriculture tends to be lower than elsewhere in the economy, the
agricultural population’s MPC might tend to be relatively high.
So far in this section, we have implicitly assumed that agriculture is
fully integrated with the remainder of the economy. This overlooks
the distinction between farmers who produce primarily for the
market and those who produce either primarily or even exclusively
for their own subsistence. Although the agricultural sector is partially
commercialised in virtually all LDCs, a large subsistence or semi-
subsistence sub-sector is characteristic of many. Although the
commercialisation of agriculture is promoted by general economic
growth and development, the rate of commercialisation is inevitably
slow where the linkages between agriculture and other sectors are
weak. Very weak linkages are characteristic of the ‘dualistic
economy’, where the bulk of domestic agricultural production is for
subsistence, where farmers have a minimal cash income to spend on
non-agricultural products and where the urban population relies
upon imports for the bulk of its food supplies. Clearly, agriculture’s
market contribution can only be small while dualism prevails,
whereas for it to be large agriculture must be fully integrated with the
rest of the national economy. Most LDCs are to be found between
these extremes and accelerating the commercialisation of agriculture
is a prominent agricultural policy objective, at least implicitly, in
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 43

many. In other words, a major short-term objective is to expand the

absolute magnitude of agriculture’s market contribution. In the long
term, although the absolute magnitude of the market contribution
may continue to grow, its relative importance in fostering economic
growth and development naturally decreases as agriculture’s share of
national product declines.

4 Factor Contribution

Whereas agriculture’s product contribution derives from agricultural

production per se (other than subsistence production), and the
market contribution derives from trade with other sectors, the factor
contribution derives from resource transfers to other sectors (Kuz-
nets, 1961). The resources transferred are capital and labour,
including human capital. We deal first with the transfer of capital.

4.1 Capital contribution

Before proceeding to discuss alternative means of transferring capital

from agriculture to other sectors and criteria for use in judging the
most appropriate amount or rate of transfer, let us briefly consider
why the net transfer of capital from agriculture is a credible means of
development. The main argument against inter-sectoral capital
transfers, if they are compulsory, are based on considerations of
equity. Is it not unfair that farmers should be deprived of part of
their wealth in order to fund developments in other sectors from
which they derive no direct benefits? The main arguments in favour
of transferring capital out of agriculture, compulsorily if necessary,
are fourfold. First, even assuming that capital-output ratios in agri¬
culture and non-agriculture are identical, the incremental demand for
capital in the non-agricultural sector may be higher in a developing
economy because the demand for non-agricultural products and
services is generally more income-elastic than the demand for food
and other agricultural commodities. In effect, then, the transfer
reflects the declining relative importance of agriculture in the
economy. Second, incremental capital-output ratios in LDC agricul¬
ture may, in fact, tend to be lower than in LDC industries. Certainly,
scope apparently exists for raising productivity in agriculture by
using methods requiring only a moderate outlay of capital, such as
adopting higher yielding crop varieties and improved strains of
livestock, and intensifying the use of fertilisers and pesticides.
44 AGRICULTURE AND ECONOMIC DEVELOPMENT

Another example is the adoption of simple mechanical aids to

supplement, rather than displace, human labour (intermediate
technology). Third, as the dominant sector in the economy of an
LDC, agriculture is virtually the sole domestic source of savings and
investment during the initial stages of development: foreign private
investment and overseas aid are only supplementary and not the
principal sources of investment capital, with the possible exception of
a few of the poorest countries in the world. Fourth, farmers are likely
to benefit indirectly from non-agricultural-type investments such as
the improvement of communications and provision of public utilities.
Except in a dualistic economy, with a wholly uncommercialised
agriculture, indirect benefits should also derive from growth and
development in the industrial sector. If sectoral diversification is
beneficial to the nation, by raising average living standards, it should
also benefit those making their livelihood in agriculture (or any other
sector), provided inter-sectoral resource mobility and income distri¬
bution are not unduly impeded by market imperfections and
discriminatory government policy interventions.
Although realism compels us to recognise that market imperfec¬
tions are prevalent in most countries, including LDCs, and that
government intervention is sometimes inimical to the interests of
farmers (although this is not always intended), there is no convincing
evidence pointing to the acceptance as an LDC norm that the
agricultural sector is excluded from deriving any indirect benefits
from incremental investment in the non-agricultural sector.
Having argued a case for some transfer of capital from agriculture
to other sectors, we proceed to a brief discussion of alternative means
of transfer. The broad choice for governments lies between relying on
the voluntary decisions of private investors in a free market to effect
the transfer, or resorting to compulsion. The conditions governing
the free market transfer of capital from the agricultural to the non-
agricultural sector have been neatly summarised by Griffin, as
follows:

(i) Farmers must sell part of their output outside their own sector,
i.e. a market surplus of agricultural products must exist.
(ii) Farmers must be net savers, i.e. they must consume less than
they produce.
(iii) Farmers’ savings must exceed their investment in agriculture.

‘If these three conditions are satisfied agriculture will have a

“balance of payments” surplus with the rest of the economy’ (Griffin,
1979, p. 109). The same author draws attention to the possible effect
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 45

of changing the land tenure system on the amount of agriculture’s

market surplus and hence upon the inter-sectoral transfer of capital.
Whereas, under the conditions of peasant agriculture, owner
farmers may not produce a market surplus (being content with their
lot as subsistence farmers) farm landlords exact a rent (in cash or in
kind) from their tenants. Thus tenant farmers are obliged to produce a
surplus in order to pay the rent. If the rent is paid in kind, and
provided it exceeds the landlord’s own subsistence requirements
(which it surely must do save in exceptional cases), the landlord will
sell part of his share of the surplus to urban-industrial consumers. If
the rent is paid in cash a proportion is likely to be saved and invested
outside agriculture. Agricultural landlords as a class are often
credited with a high marginal propensity to save (due to their relative
wealth) and are considered to have ‘city connections’. The crux of the
argument, then, is that the landlord-tenant system cannot only
generate an agricultural surplus (which might not otherwise exist),
but also enforce a net transfer of capital from agriculture to the non¬
farm sector. It follows that ceteris paribus land reform - whereby
former tenant farmers become owners and so cease to pay rent - may
lead to a fall in agricultural output and so also in the size of the
agricultural surplus corresponding with the inter-sectoral transfer of
capital. The policy implication for avoiding or reducing this ‘hidden
cost’ of land reform is clear: farmers must have adequate economic
incentives, adequate resources and, if necessary, technical assistance
to maintain or even increase output when the changeover from tenant
to owner-occupier takes place.3
Without the complication of land reform, relying on the market to
transfer capital from agriculture at any required rate-through
voluntary savings and investment - means giving the agricultural
population adequate incentives to accumulate and hold savings in
forms that can be used to fund non-agricultural type investments.
For example, an agricultural credit co-operative might receive deposits
from farmers for reinvestment in industrial enterprises, with appro¬
priate provision for liquidity: this has worked in Japan. However, in
most countries the direct purchase by farmers of industrial securities
must generally wait until a relatively advanced state of development,
when joint stock companies and an organised securities market are
both well established.
Having adopted the device of national economic planning, the
governments of LDCs generally prefer to intervene in the inter¬
sectoral allocation of capital and other resources, rather than relying
on the aggregate outcome of the decisions of large numbers of
 46 AGRICULTURE AND ECONOMIC DEVELOPMENT

independent private investors and entrepreneurs. On the particular

issue of transforming agricultural savings into non-agricultural
investment capital, government may consider that it is in a better
position than farmers, or even agricultural landlords as well, to
perceive the long-term benefits of sectoral diversification: left only to
the voluntary savings and investment decisions of farmers and
landlords, the rate at which capital was transferred out of agriculture
would be ‘too slow’.
Having chosen to intervene in the inter-sectoral transfer of capital,
the broad choice for government lies between indirect and direct
methods of control. Indirect methods, such as price controls, indirect
taxes and exchange rate manipulation, have the common objective of
changing the inter-sectoral terms of trade ratio (i.e. the farm: non¬
farm product price ratio). If the objective is to extract capital from
agriculture by this method, then the inter-sectoral terms of trade must
be shifted against agriculture - by raising duties on imports of
manufactured goods, for example. Some possible dangers and
disadvantages of this include:

(1) The creation or, more likely, aggravation of rural-urban income

inequality conflicting with broader objectives of income
distribution.
(2) The creation of ‘infant’ industries which impose a burden on the
economy by continuing to require protection from outside
competition, even in the long term.
(3) Stagnation or even a decline in the amount of the agricultural
market surplus, due to the reduced purchasing power of agricul¬
tural products at the lower price level; the corresponding decline
in agriculture’s product and market contributions could eventu¬
ally lead to industrial collapse as well.

Direct methods of control include the direct taxation of

farmers and landowners (usually based on income or property
values); compulsory deliveries of agricultural products to the state
(usually at less than the prevailing market price); product-input
barter exchange schemes with government acting as the monopoly
supplier of inputs, such as seeds and fertilisers.
Direct and indirect methods of control are not mutually exclusive;
they can be successfully combined as exemplified by what happened
in Japan during the late nineteenth and early twentieth centuries.
The Japanese experience suggests that it is feasible to extract a
substantial surplus from agriculture, without resort to measures as
drastic as physical confiscation, provided certain conditions are
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 47

satisfied. First, agriculture itself needs to be technologically progres¬

sive and on a rising plane of prosperity. Second, an appropriate
framework of financial institutions is needed to channel agricultural
savings into industrial development, without unacceptable losses of
security and liquidity. Third, a land tax is a feasible and very effective
method of forcibly extracting wealth from agriculture to fund
sectoral diversification, provided government has the will and ability
to levy and collect the tax. Over 80 per cent of central government
taxation in Japan during the last two decades of the nineteenth
century came from this source (Kuznets, 1964, p. 115). The relative
importance of the land tax and other factors in explaining Japan’s
first industrial revolution is open to debate: evidence exists that a
proportion of land tax revenues was ‘wasted’ on military expendi¬
tures rather than being used for industrial investment (Bird, 1977).
But the fact that the tax was successfuly collected to form a high
proportion of the government’s total ‘tax take’ at that time is not in
dispute.
Compared with other forms of taxation in LDCs, a land tax is
thought to possess several advantages (Lewis, Jr, in Southworth and
Johnston, 1967, pp. 464-7):
(1) It is relatively easy to collect and difficult to evade (providing
landowners hold legally valid titles which are officially
documented).
(2) It discourages the speculative holding of idle land (earning
zero rent).
(3) It induces farmers to market more of their output - due to their
higher cash requirements to pay the tax (even if output itself
remains unchanged).
But the political problem remains of getting the landowning
classes to vote for legislation increasing their own tax burden. For
example, an empirical study of the relative tax burdens of agriculture
and non-agriculture in India concluded that the agricultural sector
was ‘undertaxed’ (Southworth and Johnston, p. 482). Although the
study was confined to the incidence of direct taxation, thus ignoring
the inter-sectoral distribution of indirect taxes, this finding is
consistent with the widely held view that in India tax evasion is rife in
the rural sector.
India is a mixed economy. However, the socialist solution of
abolishing private landownership in order to mobilise surplus
agricultural savings has not been notably more successful. In Russia,
although the peasants obtained independence and private ownership
48 AGRICULTURE AND ECONOMIC DEVELOPMENT

of their land for a brief period following the 1917 Revolution, land
reform was followed, within a few years, by forced collectivisation
and compulsory deliveries to the state. Although the squeeze on
Soviet agriculture in the late 1920s and 1930s was intended to
accelerate industrial development, especially of heavy industry, the
unintended effects on agricultural output, especially in the livestock
sector, were disastrous and long lasting. In China, where the
communist revolution came later, the lessons of Russia’s adverse
experience probably influenced the choice of a different development
strategy affording a better balance between industry and agriculture.
Although some of the familiar devices have been employed to extract
a surplus from Chinese agriculture - especially rigid price controls
and compulsory deliveries - agriculture’s terms of trade have been
allowed to improve in recent years through policy adjustment, such
as:

(1) Raising the prices paid to agricultural producers relative to the

prices charged to urban food consumers.
(2) Allowing the real prices of manufactured goods to decline in
response to improved productivity.
(3) Keeping direct taxation of agriculture at a comparatively ‘low’
level and exempting agricultural productivity gains from tax
altogether, by basing tax liability on historic rather than current
production (Reynolds, 1975, pp. 21-2).

Thus in capitalist and socialist countries alike, government inter¬

vention to influence and control the transfer of capital from
agriculture to the remainder of the economy has been fraught with
many difficulties. The choice between intervention and reliance on the
market and voluntary transfers is difficult to make. Whereas the
benefits of sectoral diversification and improvement of the economic
and social infrastructure are more apparent to government than to
private investors, the ‘costs’ of intervention cannot be ignored. In
addition to those already discussed, these include:

(1) The inflexibility of intervention measures in response to changing

conditions.
(2) The extra administrative burdens of operation and policing.
(3) The wide scope for corruption.
(Griffin, 1979, p. 134)
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 49

4.2 Labour contribution

4.2.1 The Turning-point in the Growth of the Agricultural

Labour Force
The absolute size of the agricultural labour force cannot decline until
the growth rate of the non-A labour force exceeds the growth rate of
the total labour force. The conditions underlying this turning point
are of interest.

Define: rT = growth rate of total labour force

rA = growth rate of A labour force
rN = growth rate of non-A labour force
WA, WN = sectoral weights, 0 ^ W =5 1
Then, rT = rA(W*W£)* + rN(WgW?)‘
where the subscripts B and C on WA and WN denote base and current
periods, respectively.
Suppose that rA = 0; then
rT = rN(W»W»)*

Therefore, since (WjW^ < 1

rN > rT
Two deductions follow. First, if Wjj and W£ — Wg are both ‘small’,
(WbWc)* is also small, accentuating the degree to which rN must
exceed rT to be consistent with rA = 0. For example, given Wg = 0.15,
Wc = 0.20, (Wg W£)* = 0.17. That is, rN must exceed rx by a factor of
more than 5. But given Wg = 0.6, Wg = 0.8, (Wg W£)* = 0.69 and the
factor is reduced to only 1.4. It is unnecessary here to analyse the
determination of rT except to make the obvious link with population
growth. Similarly with rN, which must be closely linked with the
growth in demand for non-A products. The important first deduction
is that it is inherently much easier for a country to achieve zero
growth of its agricultural labour force when the non-A sector has
already captured a relatively large share of the total labour force, than
when agriculture is still the dominant sector in terms of employment.
Thus, in most LDCs, the turning point is unlikely to be reached until
a relatively ‘late’ stage of their development.
The second deduction is that, ceteris paribus the higher the
exogenously determined value of rT, the higher the ‘required’ value of
rN to be consistent with rA = 0. In other words, the higher the rate
at which the total labour force is growing, the more ‘difficult’ it is to
50 AGRICULTURE AND ECONOMIC DEVELOPMENT

reach the turning point. By implication, the higher the population

growth rate the more difficult the task. Our empirical estimates of
rates of growth in the agricultural labour force, the non-agricultural
labour force and the total labour force in a number of countries over
the period 1960-70 are presented in Table 3.4. The countries are the
same as those shown in Table 3.1, dealing with sectoral contributions
to GDP, except that Bangladesh is omitted due to missing inform¬
ation, and Nepal is included instead. The present table also contains
estimates of average labour force growth rates within two large
groups of countries, namely, 38 low income countries and 52 middle
income countries (again, corresponding with groups shown in Table
3.1). The method of estimation is explained in the appendix.
Amongst the 38 low income countries, the agricultural labour force
was still growing at an average rate of 2 per cent per annum over the
observation period. Within this group, amongst the five countries
shown in the table, the rate ranged from a low of 0.9 per cent
(Indonesia) to a high of 2.9 per cent (Pakistan) (Table 3.4). Even
amongst the middle income countries, the average rate was 1.4 per
cent per annum. However, amongst the nine individual countries,
shown in the table, although several exceeded the average rate, the
rate was negative in four, namely, Algeria, Argentina, Yugoslavia
and Spain. This signifies that these countries had reached the turning
point by 1960 or before. Moreover, all are amongst the ‘better-off’
countries in the middle income group. Thus empirical observation
confirms our earlier a priori conclusion that few countries reach the
turning point until a relatively late stage in their economic
development.

4.2.2 Transfer of Labour from Agriculture to Industry

The feasibility of accelerating development through the transfer of
labour from agriculture to industry is subject to the following
constraints:

(1) The size of any reservoir of‘redundant’ agricultural labour, i.e.

MP ^ 0 (or the rate at which labour can be made redundant by
productivity enhancement).
(2) The ‘quality’ of rural migrants as potential industrial workers (or
the costs of industrial training).
(3) The supplies and prices of non-labour inputs and other compo¬
nents of the demand for industrial labour.
A vailability of surplus agricultural labour. Although realism compels
us to recognise that in many LDCs the rate of rural exodus to towns
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 51

Tt —« —< UO <N <N OO ^<NvOOrO^<e(N--

■JSj £ <N <N<NCNCO<N CN <N cd <N Tt <N CO -—‘ *

£
Q)
oo
V I 3
K p Tt TtpoocNpcoosp©
~<3 )0 Os p 00 vo Os
Tt CO'AicOvdvdlO^'^-CO
C =>\ CO <N <N <N CO Tt

I5 O © ^ co Os Os Tf oois;‘noq^vois*^'0
<N <N<N<N<N<O —^ —! (N — © <N — © <N CO
I III
<n ^ 00 00 <n »n O ONTficiOOtfOhM
R
Ov Os O CO co so O O'e^OvOO^tONON
Os Os —■ CO 3
<N
c
k ©
s o co co r- O Os ©<NcOTtOs©<NcO<N
<S._ 0\ V" O J <N 06 cn —' d ri <N to ■—i 00 o d '—<
o
I s
•8;§
°0 Tt O 00 »n q> O p <N O co SO p CO °° <N
Table 3.4: Employment in LDCs: Shifting sectoral shares and relative growth rates, 1960-78

tv. <N d vd Tf <N id vd OO d <N cd <N Tf

sO <N —1 — 1—H r—H
<N <N

Co
SO Os uo O SO p p <N CO 00 p 00 00 SO
Tt <N CO CO 00 OO 00 cd cd cd Os <N OO
1 S
CO
OO CO iO

Sources: World Bank, World Development Report, 1980, Appendix, Tables 1 and 19.
R 00 r- SO SO <N Q 10
10
Os CO r- O O SO r- <N
r- iO OO SO OO
Os <N <N <N
c
a s
-© C C> CO <N <Nvoooxro<oor-oo
$ <N 3 Tt Tt — CO Tt CO Tt 00 ro in

^ s
©
gp-S
a a R <N co Tt Tt oo l/^ - hcOOO^'tCOM
Ov r- Os t </“> <o Tt l/lh-<C|fOfOCO — CO —
United Nations, Statistical Yearbook, 1961, Table 1.

£ c> SO Tf sO __ OO 00 Tt <N <N 10 0 CO <N

ov r- ON r- ST) SO r- m 00 SO i^l so i/'i <N SO Tt

SO 1/0 p CO Tt <N O co p p CO p Tt SO p Tt
R 0 OO Os SO d <N cd Os Tt 00 cd sd Tt cd
£p O *1 Ov — ■o co r» 00 <N <N CO •—< •—• <N
— c
.C 'C CO Tt
-5 a .25
5 32 CO <N p p °o 00 CO
^ §•? c© CO <N p co CO Os
s 00 d vd d <N p Tt cd sd d <N d
ft, Ov sC Tt <N 10 o> •—i *—« —H
Tt <N <N

"<3 •> 0

-* -ft B
V o
o
§ *
fee ^ Ov eo cfe G
’O
ed C
M c
.S! Su
Ck .C
o
c S cd <i5 JJ G
o3 d _ • E =9
•§1 ad cfe Ji .2 5 §. •c<D .2 fig
X
z a-p ^ T3 feo G 8) |
© ^ 5d O « 12 *c 15 c
80S z m3 C/D Oh n3 1 W ■h2o<S<>-c8
OO <N
<j .£ Vi CO uo
52 AGRICULTURE AND ECONOMIC DEVELOPMENT

and cities is currently higher than the growth of urban and industrial
employment - as signified by very high rates of urban unemployment
and underemployment - the economic implications of removing
labour from agriculture, under the limiting assumptions of various
models, has been a popular topic of research and a source of
controversy amongst development theorists. We therefore undertake
a brief survey of the relevant theoretical issues, and some empirical
evidence.
Several development models (discussed at more length in chap¬
ter 4) are based on the mobilisation of redundant agricultural
workers for socially-productive employment outside agriculture.
Labour redundancy in LDC agriculture is supposed to take the form
of disguised unemployment. This term has been defined in several
ways. First, there is a definitional distinction between redundancy in
the sense that
(1) the productivity of the marginal agricultural worker is zero
MPL = 0), and
(2) marginal productivity is positive but below average consumption
or the subsistence wage (0< MPL < SW).
Second, a distinction can be made between two forms of redundancy
depending on whether the length of the working day in agriculture is
fixed or variable.
With the existence of disguised unemployment conditional upon
MPl = 0 and, all other things being equal, some workers could be
permanently withdrawn from the land with no consequent fall in
aggregate agricultural output (Ragnar Nurkse, 1953). But, given that
0 < MPl < SW, and a fixed agricultural work norm, a marginal but
permanent contraction of the agricultural labour force would cause a
decline in output and, ceteris paribus, an improvement in the
agricultural sector’s terms of trade (Lewis, 1954; Ranis and Fei,
1961).
Given a flexible agricultural work norm (i.e. a variable length of
working day subject to a maximum constraint), and 0 < MPL < SW,
then permanent contraction of the agricultural labour force need not
cause a reduction in output (Sen, 1975 ch. 4). These arguments are
clarified in Figure 3.1 where TP is a total product function of
agriculture with labour, the sole variable input, measured on the
horizontal axis in hours. The number of workers is measured on the
vertical axis, reading down from the origin. Supposed that TP is
maximised at Op2 and that, initially, the labour time input is 013
hours. The MP of labour at this point is zero (MPT = 0). Suppose
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 53

TP

SW

Laboir Hours

Figure 3.1: Disguised unemployment

further that the total number of workers is initially Ow2. Then the
agricultural work norm is 013/0w2. But, since the labour time input
could be reduced from 013 to 012 without loss of output, and since
012 labour time could be provided by Owx workers without altering
the work norm (013/0w2 = Olj/Owj, given that w2l3 and Wjl2 are
parallel) Ow2 — Ow, workers are disguisedly unemployed in
Nurkse’s sense. However, even if initially the labour time input was
only 012, but the number of workers was again Ow2, the number of
disguised unemployed in Sen’s sense would again be Ow2 — Owl5
provided that agricultural workers were prepared to raise their work
norm from 012/0w2 to 012/01j.
Suppose now that equilibrium exists where the marginal produc¬
tivity of labour time equates with the subsistence wage (MPT = SW).
corresponding with a labour time input of Ol, and a total product
of pt. Suppose further that the work norm given by dividing
the initial number of agricultural workers into Olj is lower than the
maximum acceptable level. Then, the number of workers can again
54 AGRICULTURE AND ECONOMIC DEVELOPMENT

decline without loss of output, even though MPX > 0. This case
combines disguised unemployment in Lewis’s sense with Sen’s notion
of a flexible work norm.
Choosing between Nurkse (MPL = 0) and Lewis (MPj__ ^ SW)
depends on the reward system in agriculture. Neither model is in the
neo-classical tradition in which wages are determined by market
forces. Nurkse’s model fits a society in which farmers are prepared to
retain completely non-productive ‘employees’. This seems inherently
implausible, except possibly in a highly paternalistic traditional
agricultural system with a labour force consisting exclusively of
family labour. The Lewis model also precludes the employment of
hired workers according to commercial criteria for two reasons. First,
the subsistence wage is determined institutionally and not by market
forces. Secondly, although the marginal productivity of labour may
be positive it cannot exceed per capita consumption (the subsistence
wage).
Sen’s model has aroused substantial controversy due to its inherent
assumption that, up to some limiting input value, the labour supply
function in LDC agriculture, at the subsistence wage, is perfectly
elastic, above the limiting value it is completely inelastic. The limiting
value of labour input is set by the minimum number of hours in a day
(or other time period) needed for eating, sleeping and other non-work
activities, including leisure. Although no reward, however large, is
sufficient to induce exceeding the limit, no extra inducement is needed
to move up to the limit from some lower level of working hours. Thus,
provided the limit on individual working hours is not exceeded, it
may be feasible to obtain either extra output from a given-sized
labour force, or a given output from a smaller labour force. The
second of these alternatives signifies Sen-type disguised
unemployment.
Sen’s model contrasts with neo-classical wage theory which
postulates that the supply price of labour is a positive function of
work effort, as measured by the length of the working day, for
example. Since workers need a ‘bribe’ to induce them to work longer
hours, the supply function is upward sloping, i.e. it is less than
perfectly elastic.
The crux of the argument between Sen and adherents to the neo¬
classical model concerns the peasant agriculturalist’s marginal
valuation of leisure. Is leisure a superior good, as assumed by the neo-
classicists? Or does it alternate between worthlessness above a critical
level (corresponding with the limit on labour input) and pricelessness
below that limit, as postulated by Sen?
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 55

The theoretical issues raised by notions of disguised unemploy¬

ment have been neatly summarised and clarified by Berry and Soligo
(1968), who attempt to demonstrate that, ceteris paribus, aggregate
agricultural output cannot be maintained in the face of a decline in the
stock of agricultural workers, except in three special cases which they
identify. Berry and Soligo formulate the theoretical problem
posed by the concept of disguised unemployment as follows: given a
representative family farm worker’s production function and the
average product (AP) per man hour corresponding with a given total
farm output (P) and total man hours per farm (W), find the conditions
under which total output (and total man hours) can remain constant,
despite the loss of one (or more) members of the farm labour force.
In Figure 3.2,

Bw! x slope of BE = OS

where the slope of BE measures AP, and OS is the subsistence age.

At point C on BE,
MP per man hour = 0
Total farm output (P) = OS x n
Total work hours (W) = Bwj x n

Figure 3.2: Berry and Soligo's model

56 AGRICULTURE AND ECONOMIC DEVELOPMENT

with n signifying the number of workers comprising the farm labour

force.
Suppose that one worker leaves the representative farm perma¬
nently: the production function shifts up from SCB to TDB. To
maintain total farm output at P (and total work hours at W) it is
necessary to move up BE from C to D, where BE intersects TDB,
At point D on BE,
MP per man hour =0

ensuring that total farm output remains at P

Bw2 x slope of BE = OT

where OT is the daily wage

P = OT x (n - 1)
W - Bw2 x (n — 1)
So, provided D is the new equilibrium point on TDB, rather than
some lower point corresponding with a shorter working day, n — 1
workers will produce the same output P as was formerly produced by
n workers. But what demand conditions need to be fulfilled to ensure
that the new equilibrium is at D?
Consider a general utility model of the from U = f[F, L], where
F = food and L = leisure. The sign of the marginal rate of substitution
between the arguments is assumed to be negative. The objective
function is to maximise U subject to an income constraint, where
income (= food) is a function of labour time (or leisure time
forgone). The maximand on the utility surface is where the produc¬
tion function is tangent to the highest attainable indifference curve. In
Figure 3.2, TDB is the relevant production function and I0 and It are
indifference curves.
Special case 1: leisure an inferior good. If leisure is an inferior good,
food has no opportunity cost. Instead of being a function of food and
leisure, utility is merely a function of food (or the work expended in
producing it) i.e. U = f[F], Clearly, in this case we can re-label the
vertical axis of Figure 3.2 ‘utility’, and moving from C to D on BE is
clearly optimal as it maximises U, subject to the constraints on
production.
Special case 2: leisure satiation. If the representative worker is
satiated with leisure at wage OT and leisure hours Ow2, his
indifference curve at D will be flat. But since TDB is also flat at D
(where MP = 0), D must be the point of tangency between TDB and
Ij, the highest attainable indifference curve.
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 57

Figure 3.3: Sen's model

If Ow2 does not represent leisure satiation, the point of tangency

will be on some sloping portion of TDB to the right of D. Thus work
day length Bw2 will not be reached and total output will fall.
Special case 3: food and leisure perfect substitutes (Sen’s model). If
food and leisure are perfect substitutes, with a constant marginal rate
of substitution regardless of the consumption ratio, the indifference
curves will be linear, as shown in Figure 3.3.
It can be shown that if I0 is tangent to SCB at C, 11 must be tangent
to TDB at D. So, even if MUW < 0, MUL > 0, the required shift from C
to D will occur provided F and L are perfect substitutes.
It is worth nothing that, in all three special cases, extending the
length of the working day raises the daily wage above the original
subsistence level OS. This violates a commonly made assumption that
real agricultural wages remain constant even when the number of
agricultural workers is falling (Ranis and Fei, 1961). But provided
they cannot earn more by leaving agriculture, it is theoretically
feasible to hold down the wage at level OS by imposing a poll tax of
TS on all agricultural workers.
From their analysis, Berry and Soligo conclude that ‘only under
rather special circumstances does output remain constant when
labour is withdrawn from agriculture.’ But are the limiting con¬
ditions of the special cases ever met in practice? This is clearly a
question which can be answered only in the light of empirical
58 AGRICULTURE AND ECONOMIC DEVELOPMENT

evidence. In particular information is needed to determine the form

of the peasant farmer’s utility function.
Although there have been several attempts to measure disguised
agricultural unemployment in LDCs, some of the results obtained
fail to carry conviction. Some studies have been of the post hoc ergo
propter hoc variety. For example, Schultz found that in India
agricultural output declined in the year following the influenza
epidemic at the end of the first world war. He also observed that flu
epidemic fatalities had reduced the size of the agricultural popul¬
ation. Had there been disguised agricultural unemployment before
the epidemic, the reduction in manpower would not have affected the
level of output. The fact that output had fallen pointed to the rejection
of the disguised unemployment hypothesis (Schultz, 1964 ch. 4). This
conclusion is open to several criticisms. First, although the winter
crop which Schultz observed was smaller than usual, the summer crop
which preceded it (but not the epidemic) was of normal size. Second,
the size of the winter crop may have been affected by low rainfall.
Third, the smaller size of the winter crop may have been in response
to a shift in demand (caused by the higher death rate). Fourth,
whereas epidemic fatalities ‘prune’ a population unselectively, the
rural-urban migration process may tend to remove the less produc¬
tive members of the farm population. Fifth, even if Schultz’s
conclusion carried conviction for India in 1918-19 it would not
necessarily be valid now (Mehra, 1966; Sen, 1975 ch. 4; Thirlwall,
1978 ch. 3).
An important distinction exists in agriculture between peak season
unemployment and unemployment at other times of year. Unless a
worker is unemployed during the peak season, as well as at other
times, output is likely to decline if he permanently leaves agriculture.
The results of a rural employment survey in Egypt appeared to
discredit the existence of peak season unemployment of agricultural
workers there (Hansen, 1966). But, in reaching this conclusion,
insufficient weight may have been given to institutional rigidities
which hinder transfers from farms with surplus labour to farms with a
labour deficit, where the former are generally ‘small’ and the latter
‘large’ farms (Mabro, 1967).
A different approach was adopted in another study of disguised
agricultural unemployment in India (Mehra, 1966). Instead of
attempting to measure marginal productivity per man day or per man
hour, Mehra sought to observe the number of working hours per
capita of the agricultural population. The actual number of working
hours was then compared with the maximum possible number, subject
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 59

to the constraints identified by Sen. The minimum number of workers

needed to accomplish the agricultural task represented by current
output was found by dividing the maximum/actual hours ratio into
the actual number of workers employed. By subtracting the mini¬
mum number of workers from the actual number, the estimated
number of disguisedly unemployed workers, judged according to the
criterion proposed by Sen, was found.
Mehra had at her disposal comprehensive farm management
survey data on average working hours per capita on farms classified
by land area. The largest farms were assumed to have no surplus
labour time, i.e. all workers on such farms were assumed to be putting
in a full working day. We adopt Sen’s symbol x* to denote a full
working day. Assuming that x* is the same for all workers on all
farms irrespective of their size. Mehra then calculated the numbers
of workers required on all other (smaller) farms. The number of
‘surplus’ workers in each farm size class was calculated by subtracting
the required number from the actual number, and a national total
of surplus agricultural workers was obtained by aggregation. This
total, expressed as a proportion of the actual number of people
employed in Indian agriculture, yielded the required measure of
disguised unemployment. Because of the discrete numbers of people
working on farms, maximum and minimum estimates of labour
requirements and disguised unemployment, respectively based on
rounding up and rounding down to the nearest whole number of
workers, were obtained. The resulting estimates of the disguisedly
unemployed proportion of the total agricultural workforce in India
in 1956-7 (the data period) ranged from 29 per cent (maximum) to 6
per cent (minimum), giving a mean value of approximately 17 per
cent. This suggests that, at the time of the survey, more than one
Indian agricultural worker in six was redundant, in the sense defined
by Sen. The results also revealed substantial inter-state variation in
disguised unemployment, with mean values for individual states
ranging from nearly 40 per cent to zero. However, despite its virtue of
thoughtful planning and painstaking execution, this study also is
open to several criticisms. First, although the assumption of zero
surplus labour on the largest farms has a logical basis, and is certainly
convenient, it is nevertheless arbitrary. If disguised unemployment
exists, its causes include a scarcity of non-farm jobs as well as
inequality in the distribution of agricultural land. Second, in this as in
other contexts, averages can be misleading by concealing the variance
of empirical observations. Differences in working hours per capita
amongst individual farms must inevitably be greater than the
60 AGRICULTURE AND ECONOMIC DEVELOPMENT

differences between the means of size classes. Thus, the use of class
means causes the labour surplus to be underestimated. Third, although
farms with surplus labour might have sold part of the surplus to
farms with a labour shortage, it was not feasible to account for such
transfers in the study. Thus, to the extent that, in fact, family
members on small farms went to work as part-time labourers on large
farms, Mehra’s method of measurement overestimated the overall
labour surplus in the system (Sen, 1975, p. 13). Fourth, further errors
could have arisen from sources such as differences in attitudes to
work and leisure based on a social class, culture or, of major
significance to economists, income. There is also the failure to
distinguish between differences in work effort intensity due to
variations in the land: labour ratio, and those due to factors such as
ill-health and malnutrition, choice of technology and the quality of
non-labour inputs (including land) - any or all of which may be
correlated with farm size (Thirlwall, 1978, p. 103). Although, a priori,
these various sources of error are not uniformly positive or negative,
it requires an act of faith to assume that their net effect is zero. Thus
the interpretation of Mehra’s findings on disguised agricultural
unemployment in India, and the policy implications of her results,
remain difficult.
The results of another empirical study of the use of labour in Indian
agriculture suggested that, whereas marginal productivity is gener¬
ally positive on farms where hired labour is employed, it is zero, or
even negative, on farms without hired labour. Moreover, farms of
both types were found to co-exist in the same geographical area
though not in the same village (Desai and Mazumdar, 1970). This
suggests, that at the micro-level underemployment in agriculture can
be due to imperfections in the market for agricultural labour, as well
as to more familiar causes such as labour immobility between
agriculture and other sectors.
A further problem arises from the difficulty of distinguishing
between people who are involuntarily unemployed or underem¬
ployed and those who prefer not to have a full-time job, or even
not to work at all. In the West, the idea of voluntary unemployment
(or even underemployment) is a largely alien concept. But the results
of a recent study of unemployment in rural Bangladesh show that in
at least one LDC voluntary underemployment is of major sig¬
nificance (Ahmed, 1978).
In this study the rate of unemployment was defined as the sum of
three components. Expressed as an identity these were

Ut = UVl + Uy2 T- U,
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 61

where UVl = rate of voluntary unemployment

Uy2 = rate of voluntary underemployment
U, = rate of involuntary unemployment (including under¬
employment)
All the variables were specified in terms of unused labour days,
without regard to season. Hence, due to seasonal variation in the
demand for labour in agriculture, these unemployment rates do not
indicate the number of workers who could be withdrawn from
agriculture without reducing output. Actual work time data were
collected from about 300 rural households by sample survey. These
data were aggregated to find the equivalent number of eight-hour
work days per annum for each individual. Three alternative assump¬
tions were made about the number of working days constituting a
man year.
Taken at their face value the results of the survey appeared to
indicate that a substantial proportion of the work potential of the
rural population remained unused. Under the median assumption of
275 working days per annum, Ux was approximately 42 per cent. But
the results also showed that, under all three assumption, voluntary
underemployment (UV2) was the major source of unused labour time.
Since the voluntarily underemployed consist of self-employed people
who are unwilling to work for anyone else, their unused labour time
cannot be mobilised for outside employment in any form, even on a
part-time basis. Under two of the three working days in the year
norms, the second most important source of unused labour time was
the voluntary unemployment (UVl) of members of village elites
who, for cultural reasons, are unwilling to work at all. Thus, the
unused labour of this group cannot be counted as part of any kind of
mobilisable surplus, short of a revolution in social attitudes. Except
under the largest of the working-days in the year norms, involuntary
unemployment (U,) - including those with too little work, as well as
those with none - turned out to be the smallest source of unused
labour time. Under the median assumption of 275 days in the
working year, only about 8 per cent of total unused labour time (UT)
derived from this source. This result points to the conclusion that, at
least in Bangladesh, the amount of surplus agricultural labour
available for alternative employment is a good deal smaller than
casual observation might suggest.
A profile of seasonal labour demands in the study area was
constructed from information obtained from a sub-sample of
households. By this method an estimate was made of the proportion
of involuntarily unemployed workers on a typical day in the busiest
62 AGRICULTURE AND ECONOMIC DEVELOPMENT

week of the year. This turned out to be fractionally below 4 per cent of
the total potential work force, including the voluntarily unemployed
who comprised about 8 per cent of the same total. The approximate
sum of these two figures (12 per cent) gives an estimate of the
proportion of potential workers who could be permanently removed
from the rural sector without reducing output. However, this figure is
not comparable with Mehra’s estimate of 17 per cent disguised
unemployment in India for two reasons. First, because Ahmed’s
Bangladesh study is unique in counting the voluntarily unemployed
as part of the potential work force. Second, because Ahmed lacked
information about how many more working hours the voluntarily
underemployed might have been prepared to put in on their own
farms, in response to the departure of former colleagues. In other
words, estimates of disguised unemployment based on Sen’s defi¬
nition are not comparable with estimates based on the assumption
that the intensity of work effort is fixed. The overall impression left
by the Bangladesh study is that, even in a country which is notorious
for its over-population, the degree of disguised agricultural un¬
employment, as conventionally defined, is not large. Nevertheless,
because social and cultural attitudes are resistant to change, a
considerable amount of labour in the rural sector appears to be
‘wasted’.
Although accurate observation is obscured by problems of both
definition and measurement, really convincing evidence of major
disguised unemployment amongst those working on the land in
LDCs is hard to find. Though the evidence is too incomplete to justify
an unqualified rejection of the disguised unemployment hypothesis
irrespective of place or time, it appears unlikely that large-scale and
permanent labour redundancy in the agricultural sector is typical in
LDCs. Hence, the notion of‘painless’ capital formation based on the
large-scale mobilisation of surplus agricultural labour may have little
or no policy relevance in most countries. This is not to deny that in the
long term the agricultural sector is bound to surrender labour to the
non-agricultural sector, but only to emphasise that the labour
transfer must be matched or preceded by other changes which violate
the ceteris paribus condition, such as the adoption of new agricultural
technology.
There may also be scope for mobilising seasonally unemployed or
underemployed agricultural workers for non-farm work during part
of the year. Since workers from farms must be free to return there
during peak periods, supplementary employment must derive from
activities which operate efficiently despite seasonal fluctuations in
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 63

manpower. It will also be an advantage if the supplementary

employment is actually located in rural areas so that workers
recruited from farms do not have to leave home. Local ‘public
works’, such as the construction of roads and dams, might meet these
requirements. This happens in China, where industrial workers have
assisted on farms during seasonal peaks and agricultural workers
have been employed on construction projects during the agricultural
off-season. The Chinese have also attempted to integrate agricultural
production with the manufacture of simple industrial products, such
as building materials, and even consumer goods for outside sale,
within the rural commune system (Reynolds, in Reynolds, 1975).
Although agricultural workers who work outside agriculture for part
of the year are not ‘disguisedly unemployed’, in the strict sense, this
may nevertheless be the most feasible short-term means of utilising
surplus agricultural manpower in LDCs.
Considering the inter-sectoral transfer of agricultural labour in its
broadest sense, with the ceteris paribus condition relaxed, we now
address ourselves briefly to two further aspects. These are
(1) The ‘quality’ of rural migrants as potential industrial workers
and the implicit transfer of ‘human capital’ and
(2) The availability and prices of complementary factors of pro¬
duction in the industrial sector.

Quality of redundant agricultural labour and the inter-sectoral transfer

of human capital. The more obvious costs of transferring workers
from agriculture to industry include transport and rehousing, for
example. Somewhat less obvious are the cost of training or retraining
and the reasons why this is necessary. There is also the question of
how these costs are distributed between the two sectors.
Rural migrants will normally require some training in industrial
skills before becoming fully-fledged members of the industrial labour
force. The costs of training for some types of work may be enhanced
by the low formal educational qualifications of many migrants. The
farm to factory transfer may also mean a change in attitudes to work.
The rural migrant may experience difficulty in adjusting to the
standards of punctuality and regular working hours inherent in the
factory system. Although some of the consequential adjustment costs
may be borne by the newly-arrived workers themselves, perhaps in
the form of lower wages, a share may also fall on employers. The rural
migrant’s productivity in industry may also be affected adversely by a
legacy of poor health or weakened physical condition caused by
malnutrition. To the extent that the costs of remedying such ills fall
64 AGRICULTURE AND ECONOMIC DEVELOPMENT

on employers, they add to the ‘hidden costs’ of employing migrants

from rural areas, where standards of health and nutrition are almost
invariably lower than in towns and cities.
Although industrial employers, or government, bear the costs of
training rural migrants in industrial skills, the industrial sector
‘inherits’ from the agricultural sector the capital invested in rearing
and educating them to working age. Schultz has emphasised the
complementary roles of material capital and ‘human capital’ in
agricultural development, and the potential benefits of providing
better educational and health services in rural areas of LDCs to
improve the quality of human resources in agriculture (Schultz, 1964,
ch. 12). Kuznets has also made a notable contribution in stressing the
size of the investment made by parents of rural migrants in rearing,
educating and training them to maturity. He gives an illustrative
example, based on ‘realistic’ assumptions, in which the annual value
of this transfer could amount to as much as 6 per cent of a developing
country’s GNP (Kuznets, 1964, section iv). Although the inter¬
sectoral transfer of human capital is closely linked to the transfer of
labour, discussing it here, rather than under agriculture’s capital
contribution, is arbitrary and undertaken largely for convenience.
Since landless labourers are often the poorest section of the rural
community, they would appear to have the most to gain from leaving
agriculture. However, because of their poverty, they may also be least
able to bear the costs of migration.
Supply of complementary factors of production. Industrial develop¬
ment cannot be based on labour alone, capital inputs are also
needed, even where industries with a low capital: labour ratio, such
as some types of textile manufacturing, are deliberately selected for
development. In practice, industrial development is frequently
constrained by available supplies of capital and other co¬
operant factors - including technical know-how and managerial
expertise - even though unskilled labour may be plentiful. The
existence of high open unemployment in and around large towns and
cities in many LDCs supports the belief that, at least in the short run,
industrial development is not constrained by an inelastic supply of
labour.
For the immediate future, therefore, an important policy objective
in many LDCs might be to increase employment in agriculture, or at
least to slow down the rate of out-migration in order to contain, or
even reduce, the number of urban unemployed. This is the main
policy conclusion of Todaro’s celebrated rural-urban migration
model (Todaro, 1969).
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 65

The case for expanding agriculture in order to create employment

is supported by evidence on employment linkage effects. Employ¬
ment linkage may be defined as the total effect of a unit increase in
final demand for the product of a given industry on the derived
demand for labour, both in that industry and in all others linked with
it in the inter-industry input-output matrix. Like the income linkage
coefficient YL, the employment linkage coefficient EL reflects the
labour coefficient (in value terms) of sectoral output. EL = 0 would
signify a zero labour coefficient. Because of the importance of
employment as a source of income, there tends to be a close and direct
relationship between income linkage and employment linkage coeffi¬
cients. Thus an industry which is strong on income generation can be
expected to come out well in terms of employment creation also. This
is borne out by the results of the empirical study of inter-industry
linkage in Taiwan referred to above. For the same four sectors, the
employment linkage coefficients were as shown in Table 3.5.
These results indicate that agriculture’s employment generating
capacity is superior to that of both the input supply (working capital
for agriculture) and food processing industries. Indeed, amongst a
total of twelve industries, the two sub-sectors of agriculture ranked
first and third, whilst the mining sector ranked second. In contrast,
the agricultural input supply and food processing industries ranked
tenth and twelfth, respectively.
Most of the literature on surplus labour in agriculture is concerned
with densely populated countries. Moreover, development econo¬
mists have tended to assume that in sparsely populated countries or
regions, where labour is scarcer than land, surplus labour is ruled out
ex hypothesi. Any transfer of labour from agriculture to another
sector is therefore contingent upon investment to advance farm
labour productivity. However, Helleiner (1966) has put forward a
model postulating a potential labour surplus in a land surplus

Table 3.5: Employment linkage coefficients

Industry EL

Agricultural production for food 0.852

Agricultural production for raw
materials 0.908
Working capital for agriculture 0.725
Food processing 0.673

Source: as Tables 3.2 and 3.3.

66 AGRICULTURE AND ECONOMIC DEVELOPMENT

economy. In this model a difference is postulated between the amount

the agricultural working population is technologically capable of
producing and the amount it chooses to produce under a ‘limited
material wants’ constraint. Assuming that the potential surplus can
be mobilised - by taxation for example - it can take the form of
either food or labour, but not both. Unlike Sen, who assumes that,
until the maximum length of working day constraint becomes
binding, leisure is valueless to family farm workers, Helleiner
assumes that once their basic subsistence requirements have been
met, such workers have a strong preference for leisure. Although
Helleiner offers historical evidence from Nigeria to back his hypothe¬
sis, more direct evidence backing the limited wants assumption is
clearly needed.

5 Foreign Exchange Contribution

The potential benefits of overseas agricultural trade in an open

economy have been discussed under the ‘market contribution’. In this
section we adopt a more pragmatic approach. Given as an axiom that
development in most LDCs is constrained by a scarcity of foreign
exchange, we examine how agriculture may contribute to relaxing
that constraint. But first we consider briefly whether the foreign
exchange contribution conflicts with the market and factor
contributions.
It has been argued that there is a fundamental contradiction
between

(1) the ‘balanced’ development of agriculture and industry in a

closed economy where agriculture aids sectoral diversification
through its market and factor contributions; and
(2) the ‘unbalance’ of an open economy in which agriculture is no
longer constrained by the size of the domestic market and where
agricultural exports compete with the non-agricultural sector
for capital, or even labour (Myint, in Reynolds, 1975).

Food that is sold abroad is not available for domestic consumption in

the non-agricultural sector or to help contain inflation. Resources
utilised in producing agricultural products for export are not
available for transfer to the non-farm sector.
There is some substance in this argument which cannot be
dismissed out of hand. In a country with a lagging agricultural sector
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 67

and a large and unmanageable food import bill it might well make
better economic sense to expand food production for the
domestic market than to encourage agricultural exports. But
once domestic agriculture is able to meet the basic requirements
of the domestic market, and the food import bill is reduced to a
manageable size, it may be sound policy to exchange agricultural
exports - either of food or other agricultural products - for imported
goods of the type needed to quicken the pace of development through
sectoral diversification. Indeed, exchanging surplus agricultural
goods for imports, particularly of capital goods, may be seen as a first
step in the industrialisation process (Nicholls, 1963). More pragmati¬
cally, in many developing countries which are chronically short of
foreign exchange, selling agricultural products abroad is a feasible
and practical method of earning foreign currency to purchase
‘essential’ imports. Alternative sources of funds for buying from
abroad, such as official aid, foreign loans and private foreign
investment, are all useful for ‘topping up’ but are rarely available in
sufficient amount to meet more than a minor proportion of the bill for
indispensable imports of goods, raw materials and fuels. Although
the difference between indispensable imports and those which it pays
a country to replace with a domestic substitute may be hard to
determine in practice, it cannot be seriously doubted that most LDCs
intent on a higher rate of economic growth and development do have
minimum import needs because of their limited range of industrial
skills and gaps in their resource endowments.
For a country wishing to increase its export earnings, an expansion
of agricultural exports offers a number of potential advantages not
shared by exports of manufactured goods. First, an export crop such
as coffee, cocoa or cotton, can often be added to an existing, largely
subsistence, cropping system, thus avoiding major new investment.
Second, to the extent that new investment is needed, the amount of
capital needed is often only moderate due to the relatively low capital:
output ratios inherent in some types of agricultural product. Farmers
are capable of using highly labour-intensive methods to create capital
assets such as roads, dwelling houses and other buildings, drainage
and irrigation works, and terracing to prevent soil erosion. Much of
the labour is low opportunity cost family labour which would
otherwise be underemployed during slack periods of the farming
year. Such methods of lowcost capital formation in agriculture have
been much used in China. Third, because most agricultural products
are fairly homogeneous, and the market share of ‘new’ exporters is
usually quite small, the demand schedule facing the individual
68 AGRICULTURE AND ECONOMIC DEVELOPMENT

exporter is typically fairly elastic. This implies that, given freedom of

entry to overseas markets, a modest increase in export volume can be
undertaken by an individual country without depressing its terms of
trade. The terms of trade effect would of course be different (and
adverse) were a number of countries to expand their exports
simultaneously.
Another potential advantage of agricultural export trade is that
export outlets in developed countries may provide capital and
technical advice on production and product improvement. The
benefits of such advice may spread from the export sector to the
production of food for the domestic market as well.
Where the objective is to ‘save’ foreign exchange by reducing
imports, rather than earning more by expanding exports, agricultural
import substitution may be especially advantageous. First, there is
the question of which kinds of imports it is technically feasible to
produce at home. Major agricultural imports may score highly on
this count, whereas many industrial imports may not. Second, there
are the marginal costs of import replacement to be considered.
Agriculture will be superior to industry in this respect if domestic
agricultural production can be expanded at a lower real cost - by
bringing reserves of underutilised land and labour into production,
for example. But if expanding domestic agricultural production to
replace imports unavoidably entails diverting capital away from
industrial uses, where even more foreign exchange could be saved,
then the cost of agricultural import replacement will be too high.
Although LDCs are very varied in their resource structures, choice of
production techniques and the extent of competition between
agriculture and industry for the use of resources, a good many appear
still to belong to the first of these categories rather than the second.
In this section we have stressed the short-term, foreign exchange
benefits of agricultural production for export and import replace¬
ment. This emphasis is not intended to disguise recognising that, in the
long term, the relative importance of agriculture’s foreign currency
contribution is bound to diminish as the economy becomes more
diversified, and the relative importance of agriculture itself declines.
Open economies with a pronounced comparative advantage in
agricultural production face a dilemma. Should they give full rein to
their short-term comparative advantage as agricultural exporters?
Or, should they divert resources away from agriculture into industrial
investment and training in industrial skills in order to shift their
comparative advantage, in the long of term, away from agriculture to
manufacturing? In other words will a purely static view of compara-
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 69

tive advantage suffice, or must it be dynamic? Assuming that sectoral

diversification is a long-term policy objective in practically all LDCs,
it is logical that dynamic comparative advantage should not be
disregarded. This is to recognise a trade-off between the short-term
gains from holding back sectoral diversification in order to exploit a
static comparative advantage in agricultural production and poten¬
tial long-term gains from hastening the process of diversification into
new industries in which a dynamic comparative advantage is thought
to exist. Finding the optimum point of trade-off, where short-term
gains are equal to the present value of long-term losses (or short-term
losses are equal to long-term gains), is undoubtedly difficult due to
the dearth of needed information and the many uncertainties
involved. But it would be unrealistic to suppose that, in practice,
long-term trading policy is ever based exclusively on what is
perceived to be a country’s static comparative advantage.

6 Conclusions

This chapter points to a number of conclusions about the role of

agriculture in economic development. First, and most importantly, in
many LDCs the creation and sustained increase of a marketable
surplus of agricultural products is a virtual precondition of sectoral
diversification and hence of development itself. Although some
1 exporters of mineral oil or other natural resources may be exceptional
in this respect, because these are depletable assets it is doubtful
whether even they can afford to neglect agriculture in the long-term.
There are four major reasons why a growing surplus of agricultural
products (in the sense that production exceeds consumption in
agriculture itself) is needed:
(1) to increase supplies of food and agricultural raw materials at non¬
inflationary prices;
(2) to widen the domestic market for industrial goods through
increased purchasing power within the rural sector;
(3) to facilitate inter-sectoral transfers of capital needed for in¬
dustrial development (including infrastructure); and
(4) to increase foreign exchange earnings through agricultural
exports.
Although these four contributions of agriculture to economic
development are conceptually distinct, they are also interdependent:
moreover, they need to be consistent. We have discussed one aspect of
70 AGRICULTURE AND ECONOMIC DEVELOPMENT

consistency in the previous section, where it was concluded that the

foreign exchange contribution need not be inconsistent with the
market and factor contributions. But what about the consistency of
the capital contribution with other contributions? Is transferring
capital from agriculture to the non-farm sector consistent with the
retention in agriculture of sufficient resources to enable industry’s
growing food and raw materials requirements to be met without
inflation? And is it consistent with sustaining agriculture’s important
role as a buyer of industrial output? These questions were discussed at
some length in section 4.1, where it is concluded that although the
capital contribution is not necessarily inconsistent with the other two
‘closed economy’ contributions, the choice of method for transferring
capital out of agriculture, and finding the best rate of transfer, are
critical policy decisions.
A second major conclusion of this chapter is that the importance of
agriculture as a source of redundant labour waiting to be mobilised
for relatively ‘painless’ industrial development has been much
exaggerated. There is a dearth of convincing evidence supporting
disguised agricultural unemployment as an empirically valid concept,
even in LDCs. This means that, apart from the recruitment of
seasonally underemployed agricultural workers for temporary em¬
ployment outside agriculture during the slack season, the agricultural
sector’s ability to yield labour to the non-agricultural sector is
contingent upon the achievement of significant and sustained ad¬
vances in agricultural labour productivity. The supposed importance
of agriculture’s labour contribution, at least in the short term, is
also undermined by a dearth of evidence that, in LDCs, development
outside agriculture is constrained by the scarcity of unskilled labour.
On the contrary, high rates of urban unemployment in most LDCs
point to the conclusion that the demand for industrial labour
typically lags a long way behind the supply of those seeking industrial
employment. The rate of rural-urban migration is too high, so that
an appropriate policy objective might be to attempt to damp it down
by providing would-be migrants with incentives to remain in
agriculture. These might include better education, health and other
social amenities, as well as higher farm incomes.
Our final conclusion is that because most LDCs need to generate a
larger agricultural surplus to enable the agricultural sector to play
its developmental role more effectively, they must also consider how
best to achieve the needed increase in agricultural output. In some
areas the mobilisation of seasonally underemployed agricultural
workers for supplementary employment in agriculture may be a
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 71

feasible option, particularly where adequate supplies of complemen¬

tary inputs, such as irrigation to extend the growing season, are
available. Elsewhere, securing the required increase in agricultural
output may be contingent upon the scope for relaxing any or all of
a large number of production constraints. This may require major
institutional changes, such as land reform, as well as a battery of
measures to equip farmers with the knowledge, confidence and en¬
hanced economic incentives they need in order to use unfamiliar,
output-increasing technology. Successful application of such
measures will yield higher agricultural output in the short term, and
the release of labour to the industrial sector in the long term. As
shown in section 4.2, any decline in the absolute size of the
agricultural labour force is likely to be delayed until a comparatively
late stage of sectoral diversification.

Appendix: Estimation of Labour Force Growth Rates

The following statistical items were obtained from the sources

quoted:

1. Total population levels, by countries, in 1960 and 1978.

2. Percentage of population of working age, by countries, in 1960
and 1978.
3. Percentage distribution of the labour force between agriculture
and non-agriculture, by countries, in 1960 and 1978.

Derived estimates were then obtained for:

4. Total working population levels (including the unemployed),

by countries, in 1960 and 1978.
5. Total numbers employed, or available for work, in agriculture
and non-agricultural, by countries, in 1960 and 1978.
Compound rates of growth in the agricultural, non-agricultural and
total working populations over the 18-year period were then obtained
by substituting the derived 1960 and 1978 population estimates in the
compound growth formula Pt = P0(l -I- r)1 and solving for r.

References

Ahmed, I. (1978), ‘Unemployment and underemployment in Ban¬

gladesh agriculture’, World Development, 6 (11/12).
72 AGRICULTURE AND ECONOMIC DEVELOPMENT

Berry, R.A. and Soligo, R. (1968), ‘Rural-urban migration, agricul¬

tural output and the supply price of labour in a labour surplus
economy’, Oxford Economic Papers, 20 (2).
Bird, R.M. (1977), ‘Land taxation and economic development: the
model of Meiji Japan’, J. Dev. Studies, 13 (2).
Chakrabarti, S.K. (1977), The Behaviour of Prices in India, 1952-70,
Macmillan Company of India, New Delhi.
Chakrabarti, S.K. (1978), ‘Money, output and price: guidelines for
macro-credit planning’, in Reserve Bank of India, Recent Develop¬
ments in Monetary Theory and Policy, Papers and Proceedings of the
Seminar held at Bombay.
Edel, M. (1969), Food Supply and Inflation in Latin America, Praeger,
New York.
Griffin, K. (1979), The Political Economy of Agrarian Change, (2nd
edn.), Macmillan, London.
Hansen, B. (1966), ‘Marginal productivity wage theory and sub¬
sistence wage theory in Egyptian agriculture’, J. Dev. Studies, 2 (4).
Hathaway, D.E. (1974), ‘Food prices and inflation’, Brookings
Papers on Economic Activity, 1.
Heady, E.O. (1962), Agricultural Policy Under Economic Develop¬
ment, Iowa State University, Press, Ames.
Helleiner, G.K. (1966), ‘Typology in development theory: the land
surplus economy (Nigeria)’, Food Research Institute Studies, vi (2).
Hirschman, A.O. (1958), The Strategy of Economic Development,
Yale University Press, New Haven.
Holden, K. and Peel, D.A. (1977), ‘An empirical investigation of
inflationary expectations’, Oxford Bui. Econ. Statist., 39 (4).
Kuznets, S. (1964), ‘Economic growth and the contribution of
agriculture’, in Eicher, C.K. and Witt, L.W. (eds), Agriculture in
Economic Development, McGraw-Hill, New York.
Lewis, W.A. (1954), ‘Economic development with unlimited supplies
of labour’, Manchester School, pp. 130-90.
Mabro, R. (1967), Industrial growth, agricultural under-employment
and the Lewis model: the Egyptian case’, J. Dev. Studies, 3 (4).
Maynard, G. (1961), ‘Inflation and growth: some lessons to be drawn
from Latin-American experience’, Oxford Economic Papers, 13 (2).
Mehra, S. (1966), ‘Surplus labour in Indian agriculture’, Indian Econ.
Rev. (1), (reprinted in Chaudhuri, P. (ed.) (1972) Readings in Indian
Agricultural Development, Allen & Unwin, London).
Myint, H. (1975), ‘Agriculture and economic development in the open
economy’, in Reynolds, L.G. (ed.), Agriculture in Development
Theory, Yale University Press, New Haven.
 ROLE OF AGRICULTURE IN ECONOMIC DEVELOPMENT 73

Nicholls, W.H. (1963), ‘An agricultural surplus as a factor in

economic development’, J. Political Economy, lxxi (1).
Nurkse, R. (1953), Problems of Capital Formation in Underdeveloped
Countries, Blackwell, Oxford.
Ranis, G. and Fei, J.C.H. (1961), ‘A theory of economic develop¬
ment’, American Econ. Rev. 51, pp. 533-65, (reprinted in Eicher and
Witt, 1964).
Reynolds, L.G. (1975), ‘Agriculture in development theory: an
overview’, in Reynolds (ed.), op. cit.
Schultz, T.W. (1964), Transforming Traditional Agriculture, Yale
University Press, New Haven.
Sen, A. (1975), Employment, Technology & Development, Oxford
University Press, Oxford.
Southworth, H. and Johnston, B.F. (1967), Agricultural Development
and Economic Growth, Cornell University Press, Ithaca.
Thirlwall, A.P. (1978), Growth and Development (2nd edn.), Mac¬
millan, London.
Todaro, M. (1969), ‘A model of labour migration and urban
development in less developed countries’, Amer. Econ. Rev.,
pp. 138-48.
World Bank (1980), World Development Report, The World Bank,
Washington DC.
Yotopoulos, P.A. and Nugent, J.B. (1976), Economics of Develop¬
ment, Harper & Row, New York.

Notes
1. The structural equations of Chakrabarti’s model are:
p
— = ay^y*2... (1)
p2
P2 = Bmv... (2)
P =CP|>P^... (3)
where Pj = price of A sector goods, P2 = price of non-A sector goods,
P = general price level, Yj = real A sector output, Y2 = real non-A sector
output, m= per capita money supply, and A, B and C are constants, a, /J,
y, Xx and X2 are parameters, with a priori expectations of a < 0, [l > 0, y < 1,
+ 22 = 1.
The reduced form equation for P is:
P = DYf>Yf>my... (4)
74 AGRICULTURE AND ECONOMIC DEVELOPMENT

where the constant D involves the constants A, B and C and some

parameters. To facilitate parameter estimation, (4) is reformulated as:
P = DR^'Y^^'m’... (5)
where R = Y j/Y 2
(Chakrabarti, 1978).
2. In countries with adequate industrial statistics, inter-industry linkage
effects can be estimated empirically by using input-output analysis. The
input-output model specifies that:
X — AX = Y... (1)
where X is a vector of sectoral gross outputs (Xj), A is a matrix of inter¬
sectoral input-output coefficients (a^), and Y is a vector of sectoral final
demands (Yj). In matrix notation, the simultaneous solution to (1) derives
from:
X = (I — A)_1Y... (2)
where (I — A) -1 is the inverse matrix of Aj of which the elements, Ajj5 are
the demand multiplier or linkage coefficients. Subject to the condition
Hay < 1 for all j, i.e. the column sums of the coefficients of the A matrix
are all less than unity:

(I-A)-1 = I + A + A2 + A3 + ---+An (3)

Hence, by substituting (3) in (2):
X = (I + A + A2 + A3 + • • • 4- An)Y... (4)
Intuitively, if A is ‘large’ X is also large relative to Y. Conversely, if A is
‘small’ X is not much larger than Y. In other words, the A-’s give a
direct measure of the strength of intersectoral demand multiplier or
linkage effects.
Readers are referred to the literature for fuller explanations of the
input-output model. Heady (1962 pp. 263-5) contains a particularly
lucid explanation emphasising agriculture's interdependence.
3. More comprehensive discussions of both landlord-tenant systems in
LDCs and land reform are included in chapters 4 and 8 respectively.
4 Theory of Rent and the
Concept of ‘Surplus’

1 Introduction

In this chapter we shall first discuss the theory of economic rent.

Next, we shall discuss the important concept of ‘surplus’ in agricul¬
ture which is critical for capital formation. A clear distinction will be
made among a ‘product’ surplus, a ‘labour’ surplus and a ‘financial’
surplus. We shall then discuss some characteristics of landownership
in under-developed agriculture. Here, several features of owner-
occupation, share-tenancy, ordinary lease will be described. Given
the absence of specialisation in underdeveloped agriculture, the case
of ‘interlocking’ of factors will be mentioned under which the
landlord may also act as a moneylender and a trader. Finally, the
theory of share-tenancy will be examined in detail.

2 Economic Rent

Economic rent can be generally defined as the difference in reward

actually accruing to a factor in a particular use and the minimum
reward required to keep it in that use (i.e. its minimum supply price or
transfer earnings).
It is generally assumed that agricultural land as a whole has no
alternative use, i.e. its transfer earning is equal to zero. Hence,
agricultural landlords consider any return better than no return. But
agricultural land in a particular use - wheat production, for
example - will generally have positive transfer earnings in terms of
alternative agricultural uses. Note that land in particular locations
may also have alternative value, for example, in housing or industrial

75
76 AGRICULTURE AND ECONOMIC DEVELOPMENT

development. These peculiarities enable the landowner to extract an

unearned surplus or rent from the land user, provided land is scarce.

3 The Theory of Rent

In a free market economy, rents are determined in the same way as

other prices, i.e. by the interaction of supply and demand. Since land
is usually fixed in supply, and ‘immobile’, its price or ‘rent’ is
peculiarly dependent on demand factors, i.e. changes in rent are
mainly due to changes in demand.
Changes in demand for land may be explained by a number of
reasons:

(a) A number of existing farmers may wish to take more land under
cultivation, and

Figure 4.1: Origin of rent

 THEORY OF RENT AND THE CONCEPT OF SURPLUS’ 77

(b) many would-be farmers may decide to buy more land.

If we assume free entry and exit, then the number in both classes of
farmers will tend to vary directly with product prices and farm
incomes or product prices and farm income expectations. Expecta¬
tion of higher farm income will cause upward shift in the demand
curve and an increase in rent and vice-versa. This is shown in Figure
4.1 where we measure rent on the vertical axis and demand and
supply of farms on the horizontal axis.
Assuming that all farms are let rather than being farmed by their
owners, the initial ‘equilibrium’ is obtained at rt where demand is
equal to supply. If the demand for land goes up, the demand curve for
land shifts from Di to D2, ‘rent’ rises by r^EF.
Notice that this rise in rent is due to the complete inelasticity of the
supply schedule. If the supply of land were completely elastic, no rent
would have occurred despite a shift in demand. (Students can verify
this point easily by drawing a diagram.) Note also that if the demand
for land had been very low (as indicated by the demand curve D0),
land would not have reached its extensive margin of cultivation and
no rent would have occurred.

Figure 4.2: Quasi-rent

78 AGRICULTURE AND ECONOMIC DEVELOPMENT

4 Rent and Quasi-rent

It is possible to argue that there are cases where supply of some

factors may take some time to react to a change in demand. In such
cases, a ‘quasi-rent’ emerges.
In Figure 4.2, the quantity of factors available in the short run is
given by OQ at price OP. No more of this factor is available in the
short run. If the demand curve shifts from D to Dt, the price rises to
Pi and a quasi-rent equal to OQRjPj appears. We have implicitly
assumed away any variable cost of production. The short period
supply curve is given by PRS.
In the long run, the supply curve is given by PRN which is perfectly
elastic. The price of factor goes down to the previous equilibrium
level (= OP) despite a shift in the demand curve to the right and
quasi-rent disappears.
Note that in the long run, no rent is earned in excess of cost of
production. The short period return to the factor can therefore be
regarded as ‘quasi-rent’.

5 The Ricardian ‘Corn Rent’

The Ricardian corn model has sometimes been used to determine rent
as a residual. Assume that corn is the only output that can be
produced by available land, and labour is the only input that can be
hired. In Figure 4.3, the average and the marginal productivity of
land are measured by the APL and the MPL. Assume that the supply
of labour is given by the level of population and total ‘wages fund’ is
given by past developments. The rate of wage, w* is then given by
wage fund divided by the population. The supply of labour, N*
determines the MPL. Note that profit is simply the difference between
the MPL and wage rate. Rent (the shaded area) is also a residual as it
is the surplus on intra-marginal units of labour. Ricardo also argues
that rent is high because the price of corn is high and not the other
way round. Adam Smith also argued that rent is a price-determined
surplus. ‘High and low wages and profit are the causes of high or low
price; high or low rent is the effect of it.’ The case for taxation of rent
is not, then, difficult to understand in the light of such an analysis.
It is now well known that the Richardian theory of rent has lost
ground with the advance of the marginal productivity theory of
factor income determination. Thus, rent is supposed to be deter¬
mined by the marginal productivity (MP) of land. However, the MP
 THEORY OF RENT AND THE CONCEPT OF SURPLUS’ 79

Figure 4.3: The Ricardian corn rent

theory explains the demand and not the supply of factors. Land is
usually a fixed factor in the short run and, as such, it can earn ‘quasi¬
rent’. Also, in the long run, from the stand-point of the society as a
whole, the supply of land is fixed.

6 The Rental Market

The important implication of the previous discussion is that the level

of rent is a function of the relative profitability of land use. ‘The more
profitable farming is relative to other comparable occupations, the
higher the rents prospective tenant farmers are likely to be prepared
to pay and the higher the rents landowners are likely to require to
induce them to lease their land’ (Currie, 1981). The rental contract
between the landlords and the prospective tenants can then be
illustrated with the help of Figure 4.4. Let DD stand for the market
demand curve and let SSjSj indicate the supply curve for the land (for
the sake of simplicity we assume away a ‘step’ demand and a ‘step’
supply function (for such analysis, see Currie, 1981). The equilibrium
level of rent is then shown by Re. Units to be cultivated under the
80 AGRICULTURE AND ECONOMIC DEVELOPMENT

Figure 4.4: The Rental market with homogeneous land

landlord-tenant system is given by Qe and the units that would be

farmed by owner-occupier is given by (S — Qe).

6.1 Heterogeneity of land

In the previous analysis, we assumed that all operational units of land

are the same in fertility, location, etc. We now drop that assumption
and introduce cases of different units of land which are hetero¬
geneous. Suppose there are two different units operating land. There
are then those alternatives that a prospective tenant faces: (a) he can
rent a type 1 farm; (b) he can rent a type 2 farm; (c) he can accept off-
farm employment. The highest rent that will be offered on any type of
land will then depend not only on the alternatives to land cultivation,
but also on the rent which a tenant can obtain from the other type of
land. Hence, for a given rent for one type of land, we can obtain the
market demand curve for the other type by ‘ordering’ the limit rent of
the prospective tenants. As Currie argues, that for certain combina¬
tions of rent, an owner of one type may decide to lease out his unit
and become a tenant in another type. His reservation rent may
therefore depend on the market rent for the other type. Therefore, the
THEORY OF RENT AND THE CONCEPT OF ‘SURPLUS’ 81
82 AGRICULTURE AND ECONOMIC DEVELOPMENT

supply curve of one type may depend on the market rent for the other.
In Figure 4.5, we show the simultaneous equilibrium in markets for
two types of operating units. Once again, rent is measured vertically
and quantity of type 1 and type 2 land are measured horizontally.
DjDj and D2D2 land for demand curve for type 1 and 2 respectively.
Similary, and S2S2 stand for supply of units of types 1 and 2 land
respectively. However, DjDj is the demand for type 1 land, given that
the market rent for the second type is R2. Also, SjS} is the supply of
type 1 land, given the market rent for type 2 land is R2. Hence, for
the i th type, (Sj — Q') will be farmed by their owners, but the rest
will be leased to tenants.
The actual land market in LDCs, however, may pose a number of
complications. For one thing, in the previous analysis, we have
assumed complete absence of any monopoly or market power. Such
an assumption may not be valid in many LDCs. For another, it is
necessary to modify the implicit assumptions that costs of contract¬
ing are virtually nothing. Also, the assumption of a perfect flow of
information may be difficult to sustain in a more realistic model of
a land market in LDCs.
In the next few sections, we proceed to discuss the different types
of ‘surplus’ in agriculture which can be mobilised for economic
development.

7 Agricultural Surplus

The concept of ‘agricultural surplus’ can be defined in a number of

ways. It is possible to think about agricultural (a) ‘product’ surplus
(b) ‘labour’ surplus and (c) ‘financial’ surplus. First, we shall define
the agricultural ‘product’ surplus (=M). Such a ‘surplus’ can be
defined as the difference between total production of agricultural
output (= Q) and its consumption (= C), i.e.

M = Q — C... (1)

The importance of food surplus has been well documented in the

economic literature. As Adam Smith (1776) said:

When by the improvement and cultivation of land.. .the labour of half the
society becomes sufficient to provide food for the whole, the other half..'. can
be employed... in satisfying the other wants and fancies of man-kind.

As far as the LDCs are concerned, Nicholls argues that, ‘until under-
 THEORY OF RENT AND THE CONCEPT OF ‘SURPLUS’ 83

developed countries succeed in achieving and sustaining (either

through domestic production or imports) a reliable food surplus,
they have not fulfilled the fundamental precondition for economic
development’ (Nicholls, 1963, p. 1). In the same spirit, Kuznets
argues on the basis of his studies on economic development: ‘an
agricultural revolution - a marked rise in productivity per worker
in agriculture - is a precondition of the industrial revolution in any
part of the world’ (Kuznets, 1959).
The emergence of food surplus can be demonstrated with the use of
a single diagram. Let food production be measured in the vertical axis
and employment be measured in the horizontal axis. Let us assume
that labour (= L) is the only factor of production that produces food
(= Q). Hence we have:

Q = f(L)... (2)

Let us assume that (a) wages remain fixed; (b) agricultural techniques
are given; (c) quality of land is homogeneous; (d) labour and land are
so combined that any given population maximises its agricultural
output. The total production curve is then given by OP as shown in
Figure 4.6(a). The wage line is given by OW. It is assumed that the
wage level is fixed (subsistence wage). The corresponding marginal
(MP) and average product curves (AP) are shown in Figure 4.6(b).
Where MP is = AW, the maximum difference between OP and OW
is obtained (=QS in Figure 4.6(a) with employment =Ob). The
difference between OP and OW can then be regarded as ‘surplus’
available for reinvestment. Note that such a surplus disappears at a
point like E where the total population (Ob0) consumes the whole of
total production and there is no surplus left for capital accumulation.
At this point, average product (AP) is equal to the average wage ONl5
which is just enough to keep ObQ, population alive (see
Figure 4.6(b)).

7.1 The concept of iabour surplus’

It has frequently been pointed out in the literature of economic

development that the marginal productivity of labour (MPLa) is near
or equal to zero in LDC’s agriculture, i.e. MPLa 2* 0. This implies that
in a typical LDC it is possible to remove one or more units of labour
without reducing the total output in agriculture. This phenomenon
has been branded as ‘disguised’ unemployment in agriculture by
writers including Lewis, and Ranis and Fei (see, chapter 3 of this
book; and Lewis, 1954, Ranis and Fei, 1961). It has been argued
84 AGRICULTURE AND ECONOMIC DEVELOPMENT

Figure 4.6: The emergence of an agricultural surplus

 THEORY OF RENT AND THE CONCEPT OF SURPLUS’ 85

that agriculture can contribute to the growth of the industrial sector

by transferring ‘surplus’ labour from agriculture to industry without
any reduction in the level of agricultural output. Some economic
models have been developed to show how such surplus labour can aid
industrialisation, as will be discussed later.

7.2 The concept of ‘financial’ surplus

Agriculture has sometimes been regarded as the source of the

creation of‘financial’ surplus in LDCs. The idea of‘financial’ surplus
is related to the assignment of money value of the ‘physical’ surplus
generated within the agricultural sector. Thus, if Pa is the price of
agricultural goods that have been marketed (= M) then PaM = Sf,
where Sf = the financial surplus. The behaviour of financial surplus
can be postulated by looking at the following relationship.

Sf = f(Pa, Qa,U)
where Qa = the volume of agricultural production, M = aQa and U =
other variables besides Q and Pa. It is expected that

dSr dS(
>0 and >0

In other words, the higher the level of agricultural prices, the greater
• will be the supply of agricultural produce. Also, the higher the volume
of agricultural output produced, the higher will be the marketable
production, i.e. <?M/5Qa > 0.
This phenomenon emerges partly because producers’ consumption
of agricultural goods reaches an optimum level, and partly due to a
rise in the demand for industrial goods by the producers of
agricultural goods which, in turn, gives rise to an increased demand
for money. Obviously, a positive price policy for agriculture (which
implies a relative rise in prices of agricultural goods vis a vis prices of
industrial goods) and/or a rise in output or both will imply a rise in
farm income. A part of this farm income (Ya) will be consumed
(= Ca)) and the other part can be saved and invested (= Sa). Thus we
have:

Ya = Ca + Sa or Y„ Q = Sa

Government can play an important role in mobilising the savings

(= Sa) by a judicious policy of taxation or borrowing (see chapter 3,
4.1).
86 AGRICULTURE AND ECONOMIC DEVELOPMENT

In view of the above discussion, it is not difficult to understand

why economists have paid much attention recently to the problem of
mobilising surplus from agriculture for promoting economic growth.
The creation and distribution of surplus may vary according to
different systems of land tenure in different countries. We shall now
discuss some important types of land tenure in underdeveloped
agriculture.

8 Characteristics of Landownership in
Underdeveloped Agriculture
8.1 Share tenancy

Share tenancy may be defined as a land lease under which the rent
paid by the tenant is a contracted proportion of the output produced
in a given crop year. Usually, the tenant supplies labour and the
landlord supplies land. As Cheung (1969) says:

Share tenancy is thus share contracting, defined... as two or more individual

parties combining privately owned resources for the production of certain
mutually agreed outputs, the actual outputs to be shared according to certain
mutually accepted percentages as returns to the contracting parties for their
productive resources forsaken.

However, in many cases, tenancy rights are not legally well protected
and terms of contract are very complicated.

8.2 Owner-occupier

Under an owner-occupier system, the farmer owns his land and

cultivates it with either family or hired labour, or both. Here, the
cultivator is a landlord who is free to choose the amount of inputs
necessary to realise a certain objective function (e.g. maximisation of
output, revenue or profit). In most LDCs, the distribution of land is
fairly uneven among the cultivators and as such sometimes a small
proportion of farmers may control a very large proportion of land.
Such a phenomenon also accounts for the preponderance of small
farmers with small or very small plots which are not always viable.
Since land is the most important asset that farmers in LDCs have, it is
not surprising that most redistributional programmes for agriculture
involve land reform. Note that changes in legislation pertaining to
landownership would also change allocation of resources and the
distribution of wealth.
 THEORY OF RENT AND THE CONCEPT OF SURPLUS’ 87

8.3 Ordinary lease

Under a system of ordinary lease, one observes transfers of owner¬

ship rights which may be considered as special cases of land lease. Land
lease usually implies the use of another person’s agricultural land for
the purpose of cultivation by oneself by paying a rent. Sometimes, as
in Taiwan, the lessee is not, even with the consent of the lessor,
allowed to sub-lease the whole or part of the leased farm land to
another person. However, as Cheung has shown, this does not
prevent other disguised forms of sub-leasing, such as joint tenancy.
Also, the lessor can terminate the lease contract for his own
cultivation, and can grant preferential right to lease to the original
lessee; there is no constraint on the lessor’s right to revise the original
terms upon renewal of the contract (Cheung, 1969, pp. 12-13).

8.4 Landlord as the moneylender

The problems of tenants are aggravated by another feature of a semi-

feudal agriculture - the absence of a complete specialisation. Land¬
lords frequently act as creditors to their tenants and consumption
loans are usually advanced at high interest rates. Hence the tenant
leases his land from the same man to whom he is perpetually indebted
and this reduces him virtually to the state of a traditional serf (Bhaduri,
1973). Thus the ‘semi-feudal’ landlord exploits the tenant both
through usury and through his ‘property rights on land’. It should,
however, be conceded that empirical evidence has not always
confirmed the interlocking of factors in underdeveloped agriculture
(Bardhan and Rudra, 1978; Rahman, 1979; Bardhan, 1980). The
problem of usury in the agricultural sector of the LDCs will be
discussed later. (See chapter 8 for a detailed discussion.)
There are several other types of complex contracts that may exist
between the landlords and the tenants in the LDCs. Here, we shall
confine our attention to the analysis of the theory of resource
allocation under share tenancy. This discussion will closely follow the
arguments first developed by Marshall.

9 The Theory of Share Tenancy

Let us assume that competition exists for the use of the same property
resource (land) in an economy. Let us further assume share cropping
with one tenant first. Later, this assumption can be relaxed without
88 AGRICULTURE AND ECONOMIC DEVELOPMENT

loosing generality in the conclusions drawn. In Figure 4.7 let us

measure the amount of land in the horizontal axis and product per
unit of land in the vertical axis. Let S denote the total area of land
available to a landlord, h indicate the land held by the tenant, and q
be the product. The marginal product of land is then given by dq/<3h,
which falls as h rises. Let the rent (= r) for the landlord be 60 per cent
of the yield per annum. Hence we can write: r = 0.6. The marginal rent
curve is then given by (5q/5h)r. The vertical difference between
dq/5h and (dq/5h)r is the marginal tenant income (YTM). Thus:

Ytm = (dq/dh)(l - r)
Clearly, the area below (dq/5h)r is the rent for the landlord, whereas the
shaded area is the income received by the tenant [ = (<3q/<3h)(l - r)], i.e.
change in income of the tenant w.r.t. change in land under cultivation
by the tenant. If the income of the tenant is as high as, or even higher
than, his alternative source of income, he will cultivate the land as long
as the MP of land is positive. In order to maximise income, the landlord
will raise the rental percentage, (i.e. dq/Sh x r) until the income of the
tenant from land equals his income from the alternative source.
However, the landlord may divide his land into many parts for many
tenants to cultivate if such an action results in a higher total rent. When
the number of tenants rises, MP of land moves upwards in comparison
with the MP of land when there was only one tenant cultivator. Let
(dq/dhj^dq/dh)^ ... betheMPsoflandforeach tenant and (dq/dh^r,
(dq/dh2r,... be the marginal rent curves for different tenants. The
landlord will then optimise the gap between the integral of MPs of
land and the integral of the income of tenants (see Figure 4.7). Such
an action will maximise the integral of marginal contract rents. It
also implies non-farm income of the tenant will be less than his farm
income. It follows that with private property rights over land and
tenant inputs the terms in a share contract will depend upon the
proportion of rent (= r) and the ratio of non-land to land input which
are consistent with equilibrium (for a mathematical solution, see
Cheung, pp. 19-21).
For a long time, economists have discussed the efficiency of
resource use under different types of leasing agreements. Tradition¬
ally, it has been argued that share tenancy (as well as short period
leases) is inefficient compared with the alternatives of cash tenancy
and owner cultivation. The economic case against share-tenancy
rests on three premises.
(I) The incentive argument. From the point of view of the tenant,
share tenancy is inefficient because terms of the contract are such that a
 THEORY OF RENT AND THE CONCEPT OF SURPLUS’ 89

proportion of output harvested will always be collected as rent by the

landlord; hence tenants will be lacking in incentives to produce more
only if tenants share is less than the marginal opportunity cost of his
labour.
(2) The security argument. Tenants are usually allowed to cultivate
land for a short period (typically a year or so). Such a type of contract is
not helpful to create a secure atmosphere for the tenants for cultivation
of the land owned by the landlord.
(3) The investment argument. Investment under share tenancy is
likley to be low because landlords usually charge very high rents from
the tenants. Some have considered such rents as ‘exploitative’
(Bhaduri, 1973). If the tenants are required to pay such high rents, it is
obvious that they will not have either the ability or the willingness to
invest in agriculture. However, note that the ‘exploitation’ of tenants
may reflect monopoly power rather than a particular tenancy system.
In view of the above arguments, it is easy to understand why land
reform has been strongly advocated in many LDCs. Such reforms are
regarded as useful on three grounds: (a) protection of tenants from an
‘exploitative’ mode of production which implies drastic modification
90 AGRICULTURE AND ECONOMIC DEVELOPMENT

or even an abolition of tenancy; (b) promotion of a more egalitarian

society with wealth (in this case, land) distributed in favour of
tenants; (c) efficient allocation of resources since resource allocation
under share tenancy is generally inefficient. As Chen (1961) says:

Land reform in Taiwan was carried out at just the right time. The time was
opportune because by then the landlords had outlived their usefulness and
landownership has become an obstacle to further development of agriculture
as well as industry.

Cheung (1969), however, argues that different contractual arrange¬

ments do not imply different efficiencies of resource allocation as long
as these arrangements are themselves aspects of private property
rights. Resource allocation will be different and less than optimal
if the government interferes in the activities of the market for
achieving a rational allocation or if property rights have been
controlled or abolished altogether.
Cheung gives the following argument in support of his thesis: it is
irrelevant to observe whether the landlord requires the tenant to invest
more in land and charges him a lower proportion of rent or whether the
landlord invests in land himself and levies a higher proportion of rent
from the tenant; the investment will be made as long as this raises
landlord’s ‘rental annuity’. Hence, for any contract, the tenant does
not have to possess the required amount of input.

If what he has is insufficient, the tenant may raise the farming inputs by co¬
operating with the landowner, by subleasing, by hiring farm hands, by
borrowing or by joint tenancy with another family. Moreover, the different
tenant input requirements in farms of different land grades and production
functions will match tenants with different input endowments accordingly.
(Cheung, 1969, p. 26)

However, it is worth noting that all tenant families may not be equally
productive, and landlords may fail to discriminate among tenants of
different efficiencies Next, even if all land is of the same quality, the
land size per tenant may vary. With difference in production function
for the tenants, this will imply different rental percentages for
different tenant farms to enable the landlord to maximise the rental
annuity from his total landholding. Finally, if the costs of contracting
are assumed to be positive, then such costs need to be included in the
formulation of optimum contracts.
 THEORY OF RENT AND THE CONCEPT OF SURPLUS’ 91

9.1 The Marshallian theory of share tenancy

In the classical theory as expounded by Smith and Ricardo, rent is not

generally included in the cost of production. Also, the marginal
analysis has not been used by the classical economists with precision
to obtain an equilibrium. Marshall, however, used the marginal
analysis to demonstrate clearly the tax-equivalent argument under
share tenancy. The exposition is neo-classical and the following
diagram is used to explain the basic Marshallian theory.
Let the amount of tenant labour (=t) be measured along the
horizontal axis and the average and marginal product per unit of t be
measured in the vertical axis. Let us assume that t is the only input
available to produce output and dq/di is the marginal product. In
Figure 4.8, the marginal cost of hiring an extra unit of t is given by
d(W,)/dt, where Wt is the current wage level. In a competitive economy,
this wage level is supposed to be fixed and hence W = 5(Wt)/3t is a
horizontal line. Equilibrium is obtained at a point like B since at this
point the marginal product of tenant is equal to its marginal cost, i.e. at
B: dq/di — d(Wt)/dt. The number of tenants hired will be Ot2. If we
assume owner-cultivation, we derive the same conclusion ‘whether the
owner works up to t2 and works elsewhere, or whether he works to less
then t2 and hires extra labourers at W.’ Note that at t2 the rent to
landlords is equal to MBD, an area equal to that of a fixed rent
contract.
If we assume a tax-equivalent approach to share tenancy than the
marginal product minus rent will shift downward. Suppose that the
landlord gets r per cent rent on the annual yield of a crop, the tenant
will then be left with dq/di. (1 — r). Equilibrium is now obtained at a
point like A where the marginal cost of a tenant is equal to the
marginal gain, i.e. at A d(W,)/dt = {dq/di){\ -r). The employment of a
tenant is now Otx. Out of the total product ODJ t {, the landlord will get
ED J A and the tenant will receive OEA11. Note that the tenant now gets
an area AEM which is higher than his opportunity cost given by Otx
AM. Since, at A, the MP of tenant labour is higher than the marginal
cost it is concluded that share tenancy is not an efficient form of
cultivation. It can be easily confirmed from the diagram that the area
BAJ represents the size of inefficiency in resource allocation.
The implication of the Marshallian analysis has been well sum¬
marised by D. Gale Johnson (1950):

Under a crop share lease, if the landlord’s share of the crops is half, the tenant
will apply his resources in production of crops until the marginal cost of crop
 92

Product
Labour
AGRICULTURE AND ECONOMIC DEVELOPMENT

c
c
(/>

O
Figure 4.8: Share tenancy: the tax-equivalent argument
 THEORY OF RENT AND THE CONCEPT OF SURPLUS’ 93

output is equal to half the value of the marginal output. The same tenant,
however, will conduct his livestock operations, where important costs are
borne by the landlord and the receipts are not shared with him in the usual
manner. The landlord will not invest in land assets unless the value of the
marginal product is twice the marginal cost. (p. Ill; see also Sen, 1975)

Johnson, however, has not taken into consideration that contracting

parties are free only to accept or not to accept a contract and that they
‘can get by with’ only as much as the constraints of competitions
permit. Also the emergence of positive transaction costs in contracts
will have the following three effects: (a) Existence of positive transac¬
tion costs will reduce the number of transactions. This will have
adverse effect on specialisation in the process of production and
employment; (b) such costs may influence the marginal equalities in
the use of resources; and (c) the choice of contracts may also be altered
due to the presence of transaction costs.
The last point deserves some emphasis. It is possible to observe
three major types of contracts in agriculture: (i) a fixed rent contract,
(ii) a share contract, (iii)a wage contract. Given the existence of
private property rights, the contracting parties are free to choose a
specific kind of contract. The nature of contract varies from country
to country, for example, in Japan, a fixed rent contract was the usual
phenomenon in agriculture; in Taiwan, before the land reform, a
share contracts were more frequent than fixed rents (see Cheung,
1969). The choice of contract is affected by a number of economic and
institutional factors. Within the economic costs, we should include
negotiation costs, and the enforcement costs of controlling inputs
and distributing output. It seems that total transaction costs are likely
to be higher under share tenancy than a fixed rent or a wage contract.
The terms in a share contract usually incorporate, inter alia, the kinds
of crops to be grown, the ratio of non-land input to land and the rental
percentage and they are generally agreed upon by the tenant and the
landlord.
Under a wage contract and fixed rent, usually the landlord (i.e.
single party) decides how much of the other party’s resources should
be employed and the types of crops to be grown. Under a share
tenancy, output is usually shared on the basis of actual yield, and
hence the landlord must try to monitor the yield. Thus, it is clear that
under share tenancy, negotiations and enforcements are quite
complicated and more costly than are under either a wage or a fixed
rent contract.
If we assume risk-aversion (i.e. given the same expected mean
94 AGRICULTURE AND ECONOMIC DEVELOPMENT

income, a cultivator prefers a lower to a higher variance of income), it

can be stated that share tenancy is a way of sharing risk. Note that
under a wage contract, the landlord bears most of the risk whereas
under a fixed contract, the tenant bears most of the risk. Hence in a
situation of risk-aversion, share tenancy may be preferred to fixed
rent or wage contracts. The existence of fixed rent and wage contracts
can then be explained by differences in transaction costs between share
tenancy and fixed rent and wage contracts.

10 Some Extensions of the Share Tenancy Model

It has been suggested by Cheung that the theory and evidence of

inefficiency in resource allocation under share tenancy is wrong.
However, he has reached this conclusion only from the standpoint of
the landlord who is the maximising agent. Bardhan and Srinivasan
(= BS) have used a different model where they have determined the
demand side from the point of view of a maximising tenant, just as the
supply side is determined from the landlord’s maximising decision.
Also, in the BS model, the supply of labour is determined by the tenant
rather than the landlord. It has been shown in this model that share¬
croppers stop short of equalising the MP of labour to the wage rate and
hence they are not ‘efficient’. Given his assumption of a competitive
market, one would expect r to be exogenously given. It seems that, in
fact, Cheung’s model can operate only when the landlord has the
monopoly power to choose r. Further, Cheung assumes that the
landlord can sub-divide his land, leasing out each plot to a different
tenant. If this is true, then ‘there is no reason why he should ignore a
similar decision variable on the part of the tenant, i.e., the tenant can
also choose the number of landlords from each of whom he leases in a
parcel of land’ (Bardhan and Srinivasan, 1971). Indeed, if it is assumed
that there are a large number of landlords from whom the tenant leases
in a parcel of land, then, in the BS model, it has been shown that the
tenant continuously gains from sub-division. This is expected because
of the diminishing returns to scale implied by a strictly concave
production function.
The BS model explains a common observation in Indian agriculture:
when the landlord shares the costs, the rental share he is paid is much
higher than otherwise. The Farm Management Surveys in West
Bengal show that under the Bhagchasi system the landlord does not
share in the costs and receives about half of the crop produced by the
share-cropper, whereas in the alternative Krishani system, the land-
 THEORY OF RENT AND THE CONCEPT OF ‘SURPLUS’ 95

lord himself covers most of the non-labour costs and usually gets
about two-thirds of the crop produced by the share-cropper. It follows
that government may fail to implement rent-regulating legislation
and at the same time expect to induce the landlords to share more in
tenants’ fertiliser costs through sheer exhortations.
As regards the superiority of share tenancy over fixed rent and
wage contract, the empirical evidence seems to suggest a positive
correlation between the relative importance of share-cropping
(measured by the percentage of total area under share-cropping) and
some measure of production uncertainty. (See Rao (1971) and
Bardhan (1977), for India; Cheung (1969), for pre-Communist
China.) Such a correlation may, perhaps, be important to understand
the nature of risk-aversion in underdeveloped agriculture and its
implication for the rate of adoption of new technology. (This is
examined further in chapter 6.)

References

Bardhan, P. (1977), ‘Variations in forms of tenancy in a peasant

economy’, J. Dev. Econ., 4, 2, pp. 105-18.
Bardhan, P. (1980), ‘Interlocking factor markets and agrarian
development: a review of issues’, Oxford Econ. Papers, 32, 82-98.
Bardhan, P. and Rudra, A. (1978), ‘Interlinkage of land, labour and
credit relations: an analysis of village survey data in East India’,
Economic and Political Weekly, vol. 13, nos. 6 and 7, 367-84.
Bardhan, P. and Srinivasan, T.N. (1971), ‘Crop-sharing tenancy in
agriculture: a theoretical and empirical analysis’, Amer. Econ. Rev.
61, 48-64.
Bhaduri, A. (1973), ‘A study in agricultural backwardness in semi¬
feudalism’, Economic Journal, 83, 120-37.
Chen, C. (1961), Land Reform in Taiwan, China Publishing Co.,
Taipei.
Cheung, S. (1969), The Theory of Share Tenancy, University of
Chicago Press, Chicago.
Currie, J. (1981), The Economic Theory of Agricultural Land Tenure,
Cambridge University Press, Cambridge.
Fei, J.C. and Ranis, G. (1961), ‘A theory of economic development’,
Amer. Econ. Rev. 51, 533-65.
Johnson, G.D. (1950), ‘Resource allocation under share contracts’, J.
Pol. Econ. pp. 156-68.
Kuznets, S. (1959), Six Lectures on Economic Growth, Glencoe, Ill.
96 AGRICULTURE AND ECONOMIC DEVELOPMENT

Lewis, B.A. (1954), ‘Economic development with unlimited supplies

of labour’, Manchester School of Economic and Social Studies, 22,
139-91.
Nicholls, W.H. (1963), ‘An “agricultural surplus” as a factor in
economic Development’, J. Pol. Econ., 71, 1-29.
Rahman, A. (1979), ‘Agrarian structure and capital formation: a
study of Bangladesh agricultural with farm-level data’, unpublished
PhD thesis, Cambridge University.
Rao, C.H. (1971), Technological Change and Distribution of Gains in
Indian Agriculture, Macmillan, India.
Sen, A.K. (1966), ‘Peasants and dualism with or without surplus
labour’, J. Pol. Econ., 74, 425-50.
Smith, A. (1776), A Discovery of the Wealth of the Nations,
Macmillian.
5 Agriculture in Dualistic
Development Models

1 Introduction

In most theoretical models of LDCs the relationship between sectoral

rather than aggregative growth rates has received considerable
attention from the researchers in economic development. It has been
argued that while aggregative growth models as postulated by
Harrod, Domar, Solow and Kaldor (see Jones, 1976, for a lucid
analysis) are useful for analysing the growth problems of the
developed countries (DCs), such models are not very helpful in
understanding the special problems of the LDCs. In particular, it has
been pointed out that the LDCs have special ‘dualistic’ features
(Boeke, 1953). The market structure in LDCs suffers from a great
degree of imperfection - for instance, the product market is charac¬
terised by a barter system in many rural areas of Asia, Africa and
Latin America. In many cases, peasants living within a predomi¬
nantly agrarian society exchange their crops for industrial goods and
as such the extent of monetised economy is reduced. The money and
capital markets are far from being homogeneous due to the existence
of organised and unorganised credit agencies. It has already been
shown in chapter 3 that the labour market also suffers from
considerable imperfections. The problem of disguised unemployment
clearly shows the lack of a clear relationship between the marginal
productivity of labour and wages.
Given such dualism, a number of papers have been written on the
theme of dualistic economic growth. Central to the development of
such models is the assumption that the whole economy could be
diveded into two main sectors: (i) an advanced sector which is largely
associated with an organised industrial sector; and (ii) a backward

97
98 AGRICULTURE AND ECONOMIC DEVELOPMENT

sector which mainly comprises the unorganised agricultural or rural

sector. A system of interlinkage has been assumed between the two
sectors through production relationships, and then the growth of
the whole economy has been demonstrated through a growth
of the advanced and the backward sector. However, the growth of
the agricultural sector has not been discussed in most of these dual
economy models. Such models, and their policy implications, clearly
rest on a set of assumptions which need to be examined. Since work
on the dual economy models started with the pioneering work of
Lewis (1954), it is convenient to set out his model first and then to
examine later models in turn.

2 The Lewis Model

Lewis has made the following assumptions to develop his model:

(a) The economy is divided into two sectors-a backward, pre¬
dominantly rural sector, and an advanced well-developed capitalist
sector where the market operates reasonably well and exchange takes
place. The advanced sector has also been designated as the capitalist
sector.
(b) The advanced sector utilises capital stock which is reproducible,
and capitalists receive payments for such utilisation. On the other
hand, the backward sector utilises non-reproducible capital - for
example, land.
(c) The elasticity of the labour supply is infinite in the backward
sector because in most LDCs, labourers, particularly those who are
unskilled, are available in abundance. This assumption is very crucial
for the operation of the Lewis model as it implies that (i) the real wage
rate in the subsistence sector is constant; and (ii) the marginal
productivity of labour in excess supply must be approximately equal
to zero. In other words, if some workers are removed from the paddy
or wheat fields of LDCs total output will remain the same. (Such a
situation, which has already been discussed in detail in chapter 3, has
been termed ‘disguised’ unemployment.)
However, Lewis has acknowledged the problem of scarcity of
Labour, particularly of male labour, in parts of Africa and Latin
America.
(d) Due to the differences in the production technologies, it has been
postulated that the output per capita is higher in the advanced sector
compared with the backward sector. Given the empirical evidence
available so far, such an assumption is quite reasonable.
 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 99

(e) Although the marginal productivity of labour in the subsistence

sector is zero or near zero, wages are not related to marginal product
in the backward sector. In fact, wages could consistently be above the
MPLa. In the modern, industrial sector, wages are related to
marginal productivity. The economy has very little capital. With an
unlimited supply of unskilled labour, it is very difficult to promote
economic growth, as the skill constraint could be very severe.
However, such a constraint can be overcome via education and
training (i.e. investment in human capital). In co-operation with
whatever capital is available, economic growth can take place
through a transfer of‘surplus’ labour from agriculture to industry. It
is argued by Lewis that, at the outset, a rise in the demand for labour
in the industrial sector does not raise wages because the supply of
labour from the backward sector is infinitely elastic with respect to
wages (see, assumption (c) above).
The transfer of‘surplus’ labour from the subsistence to the modern
sector should be beneficial to both. After the transfer, the backward
agricultural sector experiences an improved land: labour ratio; and
the modern, industrial sector obtains extra labourers which it needs
to increase output. Clearly, the amount of labour that can be
transferred will depend upon the amount of capital stock that is
available and the numbers of ‘surplus’ labour. The rate of transfer
will depend upon the rate of growth of profits (or ‘surplus’) within the
capitalist sector. Lewis argues that profits or surplus generated in the
industrial sector are usually invested by the capitalists. This may not
be always true. In fairness to Lewis though, it should be mentioned
that he does recognise the possibility of some ‘leakages’ from
profits, but these, like savings by workers, are supposed to be very
small.
Fig. 5.1 shows the operation of the Lewis model. The horizontal
axis measures the amount of industrial labour while the vertical axis
measures the marginal productivity and wages. The MP curve
measures the marginal productivity of labour. The fixed subsistence
wage is given by OWs. The industrial wage is equal to OWi where
OWi > OWs. It is necessary to have OWi — OWs > 0 to offer
inducement to the agricultural labourers to be transferred to the
industrial sector. Such an argument is, however, open to question.
The industrial sector employs labour up to the point where wages
are equal to MPL. Hence P will be the initial point of equilibrium and
OM amount of labour employed. The size of profit or surplus within
the advanced sector is given by the difference between ONPM and
OWiPM (= the total wage bill). Hence profit is equal to W^P.
100 AGRICULTURE AND ECONOMIC DEVELOPMENT

Assuming that capitalists will reinvest all their profits, MPL will shift
toMjPj in the next phase. Employment now goes up toOMj and the
size of profit rises to WjMjP,.. (Note that the real wage has remained
fixed at OW;.)
The process will continue to operate until all the surplus agricultural
labour is absorbed into the industrial sector. In Figure 5.1, the point
of exhaustion of ‘surplus’ labour is reached at P0 with employment at
OM0. After that point, the level of wages will start to rise, indicated by
the dashed line. This implies that after P0, the supply of labour from
the agricultural to the industrial sector will be less than perfectly
elastic and the agricultural sector will be competing with the
industrial sector for more labour. Lewis describes this phase of
economic development as the phase of commercialisation of agricul¬
ture. It is important to point out that such commercialisation takes
place because of the rise of profits in the capitalist sector, and the level
of real wages tends to increase.

Figure 5.1: The Lewis model

 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 101

2.1 Limitations of the Lewis model

Despite the fact that the Lewis model is one of the pioneering efforts
in the development of dual economy models, many criticisms can be
levelled against it.
First, the process of exhaustion of surplus labour may come to
an abrupt end because of relative change in the distribution of factor
shares. For example, the level of real wages may rise instead of
remaining constant before the economy has reached the commer¬
cialisation point (= P0) because of the operation of minimum wages
laws, governmental intervention and/or trade union activities.
Sometimes real wages may rise in the subsistence sector due to a rise
in productivity. Indeed, ‘surplus’ labour can be found not only in the
agricultural sector, but also in the industrial towns and cities of many
LDCs. ‘Surplus’ labour of the cities then, has to be absorbed within
the industrial sector and this may effect adversely the process of
labour transfer from agriculture to industry. However, Lewis might
reply that he never intended to model the rural - urban labour
transfer process as such. He specifically stipulates that the backward
sector is not confined to agriculture but also includes unorganised
urban activities, like domestic service.
Secondly, many have pointed out that it is difficult to accept the
view that the marginal productivity of labour is always near or equal
to zero in subsistence agriculture. Many empirical studies tend to
suggest that the MPLa > 0. (See chapter 3 in this volume; and Mehra,
1966, and Desai and Mazundar, 1970.) If this is true then there will be
positive opportunity costs due to a transfer of labour from agricul¬
ture to industry since such a transfer will tend to reduce the level of
agricultural output. However, some empirical studies have suggested
that a fall in labour supply has led to a fall in agricultural output and
such studies tend to confirm the Lewis-type hypothesis of a ‘surplus’
agricultural labour (Schultz, 1964). On the other hand, others have
questioned the conclusions of the study of Schultz because (a) the
study is not comprehensive enough; (b) MPLa could be positive in
some farms and negative in others; (c) MPLa could be positive or
negative in different seasons of agricultural crop production; (d)
random elements (e.g. the weather) could probably account for such
a fall in output when labour has been displaced (for an elaboration
of some of these points, see chapter 3 in this volume; see also Mehra,
1966 and Desai and Mazundar, 1970.
Theoretically, it is possible to argue that the rate of surplus labour
absorption will crucially depend upon the reinvestible surplus. Figure
 102

MPL

$
AGRICULTURE AND ECONOMIC DEVELOPMENT

Figure 5.2: Shifts in the MPL and the size of the surplus
 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 103

5.1 shows that capitalists reinvest their profit or surplus, i.e. NW,P
and the marginal productivity curve shifts from MPto M:P It is easy
to show that the size of the surplus will be smaller if the shift of the
MP curve is not parallel. In Figure 5.2 we show how this occurs. The
smaller size of the reinvestible surplus will inevitably slow down the
process of labour absorption. Such a situation arises if we assume
that the propensity to save and invest by the capitalists is less than
unity.
It is also important to demonstrate that the Lewis model operates
in a specific way which allows employment to grow with the
reinvestment of surplus or profit. But suppose the capitalists decide to
adopt capital rather than labour-intensive techniques. Such a situ¬
ation is not difficult to envisage in large parts of the underdeveloped
sector where landlords may try to introduce technical progress by
mechanisation (or ‘tractorisation’) of agriculture. An increase in
capital intensity in underdeveloped agriculture will raise the MP per
capita and this will shift the MP curve. But such a shift may fail to
raise the level of employment: Figure 5.3 illustrates this.
104 AGRICULTURE AND ECONOMIC DEVELOPMENT

In the vertical axis in Figure 5.3 we measure the MP of labour and

in the horizontal axis employment is measured. The MP! indicates
the labour productivity at the margin before the reinvestment of
surplus. Equilibrium is given by E where wages (assumed as fixed) line
WW, intersects the MP, and hence employment is given by OM.
The amount of reinvestible surplus is, as usual, given by the area
NEW. The capitalist may now decide to introduce relatively more
capital-intensive techniques which shift the labour demand curve
(MP) to MP,. The size of the reinvestible surplus now rises to N2EW.
But the level of employment remains unchanged at M. Quite clearly,
if the chosen techniques of production for raising profit are highly
capital-intensive - a phenomenon which is not rare in LDCs - then
the degree of labour absorption will be quite limited. Indeed, instead
of employment creation, there could be a labour displacement which
would aggravate problems of unemployment.
There are other factors which may raise the level of real wages.
Suppose that the level of real wages has been raised either because of
the minimum wages law or because of pressures from the trade
unions. Such a rise has been shown in Figure 5.3 by a change in the
real wages level from OW to OW,. Notice that the result is a fall in
the size of profit from N2EW to N^jWj as well as a fall in the
level of employment by MM,. This assumes MP2 is the demand
curve and not MP, (the initial position). Such a state may be
described as growth in the national income due to a rise in capital
accumulation without real development!
In the light of the above discussion, it is not difficult to see why the
Lewis model has sometimes failed in its predictive ability. In the case
of Egypt it has been reported that the Lewis model did not perform
satisfactorily for the following reasons:
(a) The choice of factor proportions in the Egyptian and Taiwanese
industries was much more capital-intensive than envisaged by Lewis
for a smooth labour absorption. As such, the level of unemployment
in Egypt and Taiwan did not show a significant fall in the 1950s and
1960s (Mabro, 1967, Ho, 1972).
(b) Lewis did not consider the important impact of a rapid rise in
population growth on labour supply. Many LDCs have in fact
experienced a population explosion between 1950 and 1975 because
of a significant fall in the mortality rate without a corresponding fall
in the birth rate (Mabro, 1967, Ho, 1972). Note that a high rate of
population growth has other effects on consumption, saving, invest¬
ment and growth.
In view of these criticisms others have tried to modify or extend the
 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 105

basic Lewis model in different ways. One of the major contributions

came from Fei and Ranis (FR), which is the subject of our discussion
in the next section.

3 The Fei-Ranis (FR) Model of a Dual Economy

The main aim of the dual economy model of Fei and Ranis is to
demonstrate that by transferring ‘surplus’ labour from the agricul¬
tural to the industrial sector, an economy can be fully commercialised
and developed.
At the outset, Fei and Ranis assume a Lewis-type economy which is
characterised by the presence of‘surplus’ labour. The level of wages
in the agricultural sector is assumed to be fixed by institutional factors.
The supply curve of labour in the industrial sector is infinitely elastic
at the beginning since the opportunity costs of displacing labour is
zero or very small. In the FR model, it has been rightly pointed out
that Lewis did not pay much attention to the role of agriculture in
promoting industrial and economic growth. Further, FR argue that a
labour transfer from agriculture to industry should be preceded by a
rise in agricultural productivity.
The FR model of dualistic economic development consists of three
stages of economic growth. In the first it is assumed that the labour
supply is infinitely elastic and, thus, the supply of labour is invariant
with respect to changes in the level of wages. In this stage there is
very little difference in the operations of the FR and the Lewis model.
‘Surplus’ labour with zero or near zero marginal productivity is
transferred from the agricultural to the industrial sector. Figure 5.4
illustrates. Figure 5.4 shows that the elasticity of labour supply is
infinity, and real wages and marginal product, measured in the
vertical axis, remain fixed at a constant institutional wage (CIW). The
derivation of the MP of labour is shown by the slope of the
production function in Figure 5.4(b). As the slope of the production
function is flat between A and D, MP is equal to zero. Figure 5.4 (c)
also depicts zero marginal product between O and C. In Figure 5.4 (c)
the constant institutional wage is given by OW whereas the average
product is given by WVQ. Growth takes place through reinvestment
of profit by the capitalist (area d*st in Figure 5.4(a)) and the transfer
of surplus labour is OC = AD. This transfer is completed when the
surplus labour is exhausted and the marginal product begins to rise.
This is the second stage of economic development. (See DP in
Figure 5.4 (b) and CH in Figure 5.4. (c).) But wages remain higher
106 AGRICULTURE AND ECONOMIC DEVELOPMENT

than both average (AP) and marginal product. Since the marginal
product in agriculture is positive there would be positive opportunity
costs in labour transfer from agriculture to industry (because of
forgone output in agriculture) and hence beyond the point t in Figure
5.4. (a) the supply curve of labour ceases to be completely elastic
as it begins to rise.
The economy will be fully commercialised at the point where the
MP = CIW. Such an equality is shown at the point E in Figure 5.4. (c)
where CIW = MP,. In Figure 5.4 (b) the corresponding point is
indicated by the tangent to the producton function at the point R.
Hence, sections DP in Figure 5.4 (b) and CH in Figure 5.4 (c) mark
the completion of the second stage of economic development.
From the point of view of the landlord, the logic of such a situation
is fairly straightforward. As long as the labour supply exceeds OL2, it
is in his interest to allow the labourer to leave since such a departure
will imply a fall in crop yield by an amount less than the level of wages
(see Figure 5.4. (a).) but below OL2 it is in the interest of the landlord
to keep the labourer on his farm. In this sense, in the FR model, the
points like E, R and L2 signify points of commercialisation.
It is possible to imagine that, at the outset, the whole population
^engaged in agricultural activities. The total output is then given by
WL, where W = AP and L = number of labourers. So long as the
number of labour taken off the land does not exceed OL0, total
wages fall in proportion to the labour transfer. However, total output
does not fall as the MPL is zero (indicated by the flat section of
the production function in Figure 5.4(b)). The surplus is WL, and
this amount is proportional to the size of the labour force in urban
areas. If we assume along with FR that landlords will sell all the
surplus on the market, then _the per capita food for the industrial
worker remains the same at W, which is the agricultural subsistence
wage. In other words, each agricultural labour goes to industry with
the same amount of food for his consumption as before. Hence, as
Dixit (1970) argues, a constant industrial product wage and constant
terms of trade are a consistent outcome of trade in food.
In Figure 5.4 (a), as the supply of labour in industry is greater
than OL0, a further transfer of labour from agriculture to industry
will result in a fall of agricultural output. Hence, beyond this point,
‘surplus’ for every industrial worker begins to decrease. This will result
in a rise in the relative price for food and it is now clear why the
supply curve of labour in the industrial sector will begin to rise. In
other words, the ‘classical’ Lewis-type model is valid only up to L0;
after that, such a model no longer functions for the industrial sector
 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 107

Marginal output
Average output

Figure 5.4: The Fei-Rams model of labour transfer from the agricultural
to the industrial sector
108 AGRICULTURE AND ECONOMIC DEVELOPMENT

since ‘unlimited’ labour supply at zero opportunity costs ceases to be

available. Note, that at the point L2 both sectors have been
‘commercialised’ as the values of MPX in both sectors have been
equalised and wages in both sectors will be equal to the MPt.
In the industrial sector a positive surplus or profit will permit capital
formation and raise the MPC. This will shift the demand curve for
labour to the right. Technical progress in industry which is
neutral/capital/labour augmenting will expedite the whole process
and the MPC in industry will be such as to exceed the first turning-
point at Lq. As wages in the industrial sector begin to rise beyond the
point L0, investment and the rate of economic growth will tend to fall.
Even if it is assumed with FR that landlords will use a substantial
proportion of their profits (from the sale of agricultural goods to the
urban sector) for investment in industries, it may not be adequate to
take the economy to a point like L2, namely the point of full
commercialisation. At this stage, it is important to introduce
innovation and technical progress in agriculture because such an act
will shift the production function in agriculture upwards. The fall
in agricultural surplus will thus be offset by the generation of more
output and price of agricultural goods will remain unchanged.
However, landlords will be induced to introduce innovation in
agriculture if the price of agricultural goods is allowed to rise initially.
Subsidisation of input costs may provide further incentives to
landlords to introduce technical progress.
The critical role of the landlord in the whole mechanism of
economic development has been well summarised by Dixit (1970).

Successful development of this economy hinges altogether too crucially on

the role played by the landlord. Heshould be eager to save. He should sell his
surplus to industry, and should transfer his savings to industrial entrepre¬
neurs. He should be eager to innovate, and thereby improve the technology in
agriculture, (p. 342)

Quite clearly, it is not always easy to envisage the portrait of such a

dynamic landlord in a typical underdeveloped agrarian economy.
In the FR model the speed with which surplus labour can be
transferred from the agricultural to the industrial sector will depend
upon (a) the rate of population growth; (b) the nature of technical
progress in agriculture; and (c) the growth of industrial capital stock,
which is determined by the growth rate of profits in industries and the
growth of surplus generated within the agricultural sector.
In contrast to the Lewis model, Fei and Ranis argue that a balanced
 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 109

growth of both the industrial and the agricultural sector is necessary

to avoid stagnation of the rate of economic growth. Hence at the
policy level it follows that the growth of agriculture and industry
should be balanced since the growth of agriculture is as important as
the growth of industry; and the rate at which labour is transferred
from the agricultural to the industrial sector must be higher than the
overall rate of growth of population.

3.1. Limitations of the FR model

Among many comments that have been made on the assumptions

and operation of the FR model, we shall discuss the following major
points.
First, it can be argued that FR have not made much attempt to
discuss the nature and causes of stagnation that characterise the
economies of LDCs. A clear understanding of some important
reasons of agricultural backwardness would have been very useful for
policy formulation. Second, FR have not made any distinction
between wage-based labour and family-based labour. Likewise, a
comprehensive analysis of self-sustaining growth in open economies
where changes in terms of trade play a crucial role would have been
illuminating. Also, the role of money and prices has been neglected. It
has been argued that money is not a simple subsititute for physical
capital in an aggregate production function. There are reasons to
believe that the relationship between money and physical capital
could be complementary to one another at some stage of economic
development, to the extent that credit policies could play an
important part in easing bottlenecks on the growth of agriculture
and industry. (See e.g. McKinnon, 1973; Shaw, 1973; Ghatak, 1981.)
Note that the investment functions has not been specified in the FR
model.
Third, FR assume along with Lewis that MPL in agriculture is zero
in the initial phase of economic development. Such an assumption
may be called in question in view of the empirical evidence obtained
so far (see chapter 3 in this volume; see also Mehra, 1966; Desai and
Mazunder, 1970; Sen, 1975). Furthermore, theoretically it is possible
to show that, in the first phase of the FR model, the level of
agricultural output will fall except in three special cases.
(a) there is leisure satiation, i.e. people are completely satisfied with
leisure and as such additional leisure consumption does not enhance
the utility of the individual;
110 AGRICULTURE AND ECONOMIC DEVELOPMENT

(b) leisure is an inferior good and as such people would like to give it
up even when income rises;
(c) food and leisure are fully substitutable with one another and as
such the marginal rate of substitution of food for leisure will always
remain the same.

Such a fixed marginal rate of substitution will imply that the

indifferences curves depicting various combinations of income and
leisure are no longer convex to the origin (indicating diminishing
marginal rate of substitution), but they are linear as shown in the
Figure 5.5. (See Berry and Soligo, 1968 for details).
 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 111

Now if we assume that the MPL in agriculture is greater than zero,

then case (a), which states that there is leisure satiation, is violated.
Alternatively, if it is assumed that the MPL = 0, then a perfect
substitution between food and leisure is no longer a sufficient
assumption for holding the total output constant. The only sufficient
condition for a fixed level of output is that leisure is an inferior
commodity regardless of the value of the MPL because under such a
condition a transfer of labour will always imply that those who are
left on the farm will intensify their work effort (since we have assumed
that leisure is an inferior good) to hold output at the same level as
before. Figure 5.5 illustrates this point. Let food be measured on the
vertical axis and leisure on the horizontal axis. The difference curves
show the possible combinations of leisure and food (i.e. I0, Il5...). The
subsistence level of consumption is equal to OS. Since OG measures
the maximum leisure hours, a move from G to a point like X,
indicates a rise in the supply of labour. The transformation of leisure
into food is given by the curve SAG which retains its normal
properties. At a point like A, the marginal rate of transformation
between food and leisure is nil, i.e. the curve is ‘flat’. As the
indifference curve I0 is tangent to the curve SAG at the same point, it
also indicates a zero marginal rate of substitution between food and
leisure. Such a flat section of the indifference curve clearly implies
leisure satiation as farmers are willing to give up infinite amounts of
leisure for an extra unit of food.
If the wage rate is fixed, the transformation curve is GH. If we now
assume that one unit of labourer has been transferred from the
agricultural to the non-agricultural sector and his land is equally
distributed among those who remain on the farm, then the curve
SAG shifts to GTR. Now to obtain the same level of output the
indifferences curve at point T must be flat, which will imply that either
leisure is a really inferior commodity or complete leisure satiation
rules. But without these extreme assumptions, total output will fall.
(For a detailed discussion of these points, see chapter 3.)
There are other criticisms of the FR model. It may be argued that
FR assume a closed economy model which denies the possibility of
importing food and raw materials. Also, if factors like seasonalities
are taken into account, then the MPL in agriculture may not always
be equal to zero. Alternatively, ‘seasonal unemployment due to
seasonality of labour demand may be mistaken for permanent
redundancy.’ Moreover, FR did not analyse the role of capital in
agricultural production - an omission, which many who believe in a
significant role of capital will find very difficult to overlook. In fact,
112 AGRICULTURE AND ECONOMIC DEVELOPMENT

some may even argue that the slow growth rate of industries can be
explained by slow growth rate in agricultural productivity in LDCs.
Under such circumstances, it is very important to raise agricultural
productivity and surplus in the first place. As Jorgenson points out,
for the expansion of the industrial sector in developing countries it is
essential not only to generate an agricultural surplus but also to
maintain it through technical progress (Jorgenson, 1961; 1967). The
other pitfall of the FR model is that it has completely neglected the
role of the service sector in the absorption of surplus labour. In many
LDCs, the service sector has played a significant role in providing
employment.

4 The Jorgenson Model

The Jorgenson model has been considered by some as the application

of the neo-classical theory of growth in LDCs. Actually, Jorgenson’s
model has elements of both the classical and the neo-classical theory.
On the one hand, Jorgenson assumes, along with Lewis and FR, that
‘surplus’ labour may exist in LDCs but not in the sense of zero MPL in
agriculture. The framework of analysis is still a ‘dualistic’ model
which consists of an industrial and an agricultural sector. Total
(= Qa) in the agricultural sector is given by land (= L) and labour
(= N). Such a production function is subject to the operation of the
law of diminishing returns. It is also assumed that technical progress
in agriculture is neither labour nor capital-using, that is, it is neutral.
Let this production function to be given by the following equation:

Qa = eatL/,N1-^... (1)

where eal measures a change in output due to technical progress. If we

assume with Jorgenson that the supply of land is given then equation
(1) can be written as follows:

Qa = eatN1“/>... (2)
Let per capita output be given by y = Q/N

y = § = eatN-p ... (3)

N

If we now differentiate (3) with respect to time (= t) and divide the

 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 113

equation by y, we get the following equation:

y/y = oc-0^... (4)

If N/N = rj (i.e. rate of growth of labour) then equation (4) can be

written as follows:

y/y = a - jfhf... (5)

y = (a - prj) y... (6)

The time path of output growth is then given by the following

equation:
y(t) = e(«~/J'i)ty(0)... (7)

It is clear that to get a positive growth of output, we should have:

a — fit] > 0

In many LDCs, it is reasonable to assume that the value of p will

remain constant as the output elasticity with respect to changes in
labour is relatively stationary. Hence, it is important either to
accelerate the rate of technical progress (i.e. the value of a), or to
reduce the rate of growth of population (i.e. the value of rj). If
possible, measures can be taken to take care of both problems at
the same time to make a significant impact on poverty and
underdevelopment. Note, that as long as a > rj a positive growth of
agricultural output per capita will be ensured. But when a — r\ the
economy will stagnate at a low level equilibrium trap.

4.1 Evaluation of the Jorgenson model

A number of criticisms have been levelled at the Jorgenson model. We

have already noticed Dixit’s criticism about the nature of Japanese
evidence that Jorgenson has cited in support of his hypothesis. The
other important point that has been made is that when Jorgenson
claims superiority for the neo-classical model, since he believes that
its predictions have been broadly confirmed by the Japanese data for
the nineteenth century, such a claim is difficult to accept since short-
run predictions of the classical model have been compared with the
asymptotic results of the neo-classical theory. Further, Jorgenson, in
his description of the agricultural production function, has not
included capital as an argument. On the other hand, empirical studies
of Nakamura for Japan, Shukla for India and Hansen for Egypt
 114 AGRICULTURE AND ECONOMIC DEVELOPMENT

indicate that capital has played a significant role in agricultural

development. Also, true to the neo-classical tradition, Jorgenson has
emphasised the role of supply factors in economic development.
Some argue that demand factors can also play quite an important role
in economic development. Finally, both the ‘classical’ (Lewis and
Ranis-Fei) and the ‘neo-classical’ (Jorgenson) writers have
overlooked the important role that service sector plays in promoting
overall as well as sectoral growth of both agriculture and industry.
Without proper transport and credit facilities, marketing, financial,
communication, educational and maintenance services, growth rates
in both sectors could be significantly reduced.

5 Kelley, Williamson, Cheetham (KWC) Model

The KWC model embodies a number of features. At the outset it is

assumed that the economy is closed. Such an assumption excludes the
possibility of analysing the effect of international trade and expan¬
sion of intermediate demand on industrial patterns. The economy is
‘dual’ - it consists of an agricultural and industrial sector each of
which produces a single homogeneous commodity. While the
industrial good can be consumed or invested, the agricultural output
will always be consumed. Thus the KWC model is different from the
Uzawa-type two-sector model where the industrial sector produces
capital goods and the agricultural sector producers consumer goods.
The production technology in each sector has been described by a
continuous twice-differentiable, single-valued function. Let us as¬
sume that the industrial sector is sector 1 and the agricultural sector is
sector 2. Returns to scale remain constant; joint products or
externalities are ruled out. Producers in both sectors wish to minimise
cost and maximise profit. Hence, we have

5.1 Production technology

Qi(.) = F'x(t)Ki(t), y(t)Ls(t)...

(i = 1,2)

Note that K,(t) > 0, Lf(t) > 0, which shows the amount of capital (K)
and labour (L) currently employed in the ith sector; also x(t) > 0 and
y(t)>0 are the respective variables of technical progress. Hence,
x(t) Kj(t) and y(t) L;(t) may be regarded as efficiency capital and
 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 115

efficiency labour, respectively. Since it has been observed that the

industrial sector in LDCs is more capital-intensive than agriculture,
the following constraints have been imposed on the elasticities of
factor substitution (0^ and <r2) in industry and agriculture

0 < fx1(t) < 1

and
1 < <r2(t) < 00

The elasticity of factor substitution is usually given by

^i(t) = F[F[/FiF;L

where

F» - gF‘
k <5[(x(t)K,(t)]

F* - ^ -
L 3[y(t)Lj(t)] ’

pi ^
kL 8 [x(t)Kj(t) ] d[y(t)L1(t) ]

These F's are really MP coefficients.

The above constraints suggest that the substitution of efficiency
labour for efficiency capital is less than 1 in industry and greater than
unity in agriculture.
It is important to point out that in contrast to the Jorgenson model,
capital has been included in the agricultural production function in
the KWC model.

5.2 Technical progress

It is assumed that rates of technical progress are exogenously given,

Ak and , respectively and they are the same in the two sectors. Note
that such an assumption is rather restrictive. Hence:

x(t) = x(o)e"k'
y(t) - y(o)e'*L'
It is possible to set x(o) = 1 = y(o). The rate of technical progress in
each sector is a weighted mean of the rate of factor substitution, the
weights being equal to output elasticities of capital and labour. The
116 AGRICULTURE AND ECONOMIC DEVELOPMENT

current rate of technical progress in the i sector is given by

^Q.(t) 1
R,(t)
at Qi(t)
This can be written as

R1(t) = 4ai(t) + 4[l-«1(t)]

where
F‘x(t)K,(t)
‘ Q,(t)
i.e. the present elasticity of output w.r.t. capital in the ith sector.
Investment equation: Assume that current net investment (Kt) is
equal to total investment (I(t)) net of replacement, which is propor¬
tional to the aggregate capital stock [<5K(t)], Thus
R(t) = l(t) - <5fC(t)

Assume that capitalists behave in a similar way in both sectors,

allocate capital goods to both sectors, and P(t) is the present price of
industrial goods in terms of agricultural goods and r^t) is the present
rental rate of efficiency capital in the ith sector. Thus we have

P(t)I(t) = x(t)[r1(t)K1(t) + r2(t)K2(t)]

Alternatively, for aggregate investment, we have

x
1(0 r(t)K
P(t)

5.3 Factor demand conditions

Assume that W(t) denote efficiency wages for labour (L) and r(t)
denote efficiency rental for capital. Factor demand is then given by
the familiar neo-classical conditions of equality between factor prices
(W(t) and r(t)) and marginal productivities of inputs (i.e. FL and Fk).
Hence we write:

W1(t) = P(t)Fi
W2(t) = Fl
ri(t) = P(t)F^
r2(t) = F l
where 1 = industrial sector and 2 = agricultural sector.
 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 117

5.4 Factor supply relation

Assume that K(t) and L(t) are the respective currently available, total
stocks of capital and labour. The full employment condition will then
be given by:

K(t) = K1(t) + K2(t)

L(t) = Lj(t) + L2(t)
The labour force (L) in ith sector is supposed to grow at a given rate,
n, > 0. Suppose that u(t) = Lj(t)/L(t), i.e. proportion of the industrial
labour force in the total labour force, then

n(t) = n1u(t) + n2[l -u(t)]

5L(o ,J_
d(t) L(t)
It is assumed that n2 > n2. Since n(t) = L(t)/L(t), the above equation
can also be written as

fyt) = {n2u(t) + n2[l — u(t)]}L(t)

5.5 Commodity markets

A novel and interesting feature of the KWC model is the introduction

of the role of demand relationships. The commodity demand system
is described by a set of equations:

P./t)
L/t) = ^t)[y(t)Wj(t)~y]

P2j(0
L/t)
= /^2jy(t)wj(t) + (1 - p2i)y
Where Dj/t) = the total amount of the ith good consumed by the
labour force in the jth sector, and are fixed parameters. Note that
7 = minimum subsistence level of per capita consumption of agricul¬
tural output. Such demand equations are based on the Stone-Geary
system of linear expenditure model which aggregates perfectly over
individuals in a group, if each group has the same utility function. In
the Stone-Geary system, the utility function for a labour in the jth
sector (Uj(t)) is given by:

Pij(t)
U,(‘) = I slog -y,i j = 1,2
i= 1 . L/t)
118 AGRICULTURE AND ECONOMIC DEVELOPMENT

The above demand relationships (i.e. D^tJ/L^t)) state that given

wage income per capita in the jth sector, ywj(t) and the community
price ratio, (Pt), each labour unit consumes at least y and then he is left
with his income which equals ywj(t) — y. This he allocates among
goods in the proportions of (ltj.
The general form of the Stone-Geary model can be stated as
follows:

Pii(t)
L/t)

Ml jr^y(t)Wj(t) + [i - /?2j]y2 - P$2 Ml

Lj(t) Pi(t)
Assume yx =0, y2 = y, P2 = 1, and we obtain the demand system
which has been obtained by KWC.

5.6 Market balancing equations

The overall market balancing equations may now be given in the

following sets of equations:
Q1(t) = D11(t) + D12(t) + I(t)
Q2(t) = D21(t) + D22(t)
wx(t) = w2(t) = w(t)
ri(t) = r2(t) = r
Note that w(t) and r(t) are equilibrium wage and rental rates. The
above equilibrium conditions require that excess demand in any
market is zero and market prices are not negative. If it is assumed that
there is at least one positive value for the terms of trade that will
satisfy the equilibrium conditions then it can be shown that
equilibrium in the two factor markets and in either commodity
market implies equilibrium in the remaining commodity market.
Hence, one of the two goods market equations may be neglected; the
model then consists of fourteen variables and fourteen equations.
Once the equilibrium factor price conditions have been used, the
basic model comprises fourteen endogenous variables and four
exogenous variables: L(t) = L, K(t) = K, x(t) = x, y(t) = y.
Although the KWC model is an extension of the Jorgenson model,
there are some differences between them. First, in contrast with the
Jorgenson model, KWC assume that both the industrial and the
agricultural sector use capital and labour. Second, labour supply in
 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 119

the Jorgenson model is exogenously given whereas the KWC,

although the labour supply function is exogenous in nj parameters,
endogenously determines the rate of population growth as a weighted
function of the differential rates of population growth in the rural and
urban sectors. However, with economic growth urban-rural migra¬
tion reduces the total rate of population growth.
There are other differences in assumptions on parameter values.
Whereas Jorgenson assumes a Cobb-Douglas production function
for both industry and agriculture, KWC assume that the substitution
elasticity between K and L, (a) in agriculture will be ^ 1 whereas
in industry a < 1. Third, Jorgenson assumes demand parameters as
the same for both sectors. In the KWC model, such parameters are
different in the two sectors. Fourth, in the Jorgensen model, wage rates
in the ‘traditional’ and the ‘modern’ sectors may not be equal.
Income of the peasant in the ‘traditional’ sector as an owner-operator
comprises the per capita product given by returns to both land
and labour. Such an income need not be equal to the income in the
modern sector which consists of a return from capital and labour. In
the Jorgenson model, the urban wage and farmer’s opportunity costs
are equal. But in the KWC model, capital returns and labour returns
are both equalised between sectors in depreciation equilibrium. Fifth,
in the KWC model, the rate of depreciation of capital is positive,
whereas in the Jorgenson model it is nil (in the industrial sector,
presumably, since A sector does not employ capital).
Thus the KWC model has been extended to analyse dualism not
only in production technology, but also in demand behaviour and
population growth patterns and as such, it clearly merits serious
attention. On the other hand, the model does not discuss in depth
different theories of population growth. The overall population
growth rate in the KWC model is assumed to be exogenous. Also,
enough attention has not been paid to the different patterns of income
distribution and their impact on consumption, growth and employ¬
ment (see Mirrlees, 1975). Finally, the role of institutional factors has
not been mentioned as such in the model.

6 Some concluding Remarks on The Dual

Economy Models
The advocates of the dual economy models have rightly emphasised
the need to examine the role of both the agricultural and the industrial
sector to promote economic growth in LDCs. Unfortunately, some
120 AGRICULTURE AND ECONOMIC DEVELOPMENT

have neglected the vital role that a developed agriculture can play in
the promotion of such economic growth. Lewis, despite his seminal
contribution, has not paid much attention to the mechanism of
development within the agrarian sector. In addition, some of the
assumptions of Lewis and FR model have not been sustained in
practice. It is not difficult to understand why some of the predictions
of these dual economy models have gone wrong. The crucial role of
population growth has not always been duly emphasised (with the
exception of the Jorgenson type of model). Once we include the role
of population growth in LDCs, it is easy to understand why the pace
of labour absorption has taken place at a very slow rate in most
LDCs. The process of capital accumulation has been closely related
to the choice of techniques in both industry and agriculture. If a
technical1 progress is capital-augmenting (i.e.jcapital is used relatively
more than labour) then the rate of growth of employment in industry
will be slower than under a neutral or labour-augmenting technical
progress. Similarly, if agricultural development largely means in¬
creased mechanisation (‘tractorisation’) of agriculture in LDCs, then
despite some improvement in agriculture (measured by output or
productivity), some labour displacement is inevitable. This, again,
raises the total supply of labour, unemployment, malnutrition and
abject poverty. It is important to emphasise that the failure to relate
capital accumulation to the choice to techniques has significantly
diminished the predictive power of the dual economy models.
The neglect of the analysis of the development of the agricultural
sector has already been mentioned. It has indeed been argued that
without rapid technical progress in agriculture, phase 3 of the Lewis,
Fei-Ranis-type model would be characterised by hyperinflation
rather than self-sustained growth (Watanabe, 1965).
It is worth emphasising that the assumption made by Lewis and Fei
and Ranis, that an agricultural surplus already exists to be mobilised,
is not always tenable, particularly in the least developed countries. It
is, perhaps, more useful to postulate (as Jorgenson does) that the
surplus has to be created within the agricultural sector by productiv¬
ity improvement first.
Further, most writers of the dual economy models postulate a
closed-economy model. Such an assumption has some justification in
view of the minor role that has been played by trade in the economic
development of LDCs. On the other hand, it is possible to argue that
some LDCs may actually gain from trade. More specifically, trade in
agricultural commodities could play an important role at the initial
phase of economic development. In a later stage, apart from trade in
agriculture, LDCs may export manufactured goods quite success-
 AGRICULTURE IN DUALISTIC DEVELOPMENT MODELS 121

fully due to the comparative cost they enjoy in the production of

relatively more labour-intensive commodities.
The important role that the service sector can play in economic
development of LDCs within the dual economy model has never been
sufficiently emphasised; nor from the point of view of industrial
development, has due allowance for leakage from the wages fund to
cover the cost of marketing and distribution been made. It is widely
known that the growth of such credit, marketing and distribution
facilities aids economic growth considerably.
More generally, most writers of dual economy models do not pay
much attention to the crucial problems of extracting the agricultural
surplus. A more comprehensive investigation of such problems
should include, interalia, the analysis of terms of trade, inter-sectoral
resource flows, taxes and subsidies, the demand functions (as
emphasised by Kelley and Williamson and Cheetham) and perhaps
the money demand functions. It is no surprise why some writers have
taken such a critical view of the dual economy model. Griffin, for
instance, argues that the dualistic models are not very helpful because
the assumptions on which they are generally based are not always
plausible and the predictions of the theory wrong. On the utility of the
FR model Griffin states, ‘in its country of origin [i.e. Pakistan], it
remains pathetically irrelevant’ (Griffin, 1969).
As regards the growth of the service sector, it seems sensible to
argue that it should increase at the same proportionate rate as the
modern sector to avoid bottlenecks in the overall rate of growth. As
Dixit (1973) rightly points out: ‘If we wish to emphasise the
commercialisation aspects of the division, we should classify in the
modern sector those services which are available on the market and
leave domestic services in the traditional sector’ (p. 326).

References
Berry, A. and Soligo, R. (1968), ‘Rural-urban migration, agricultural
output and the supply price of labour in a labour surplus economy’,
Oxford Economic Papers, 20, 2, 230-49.
Desai, M. and Mazumdar, D. (1970), ‘A test of the hypothesis of
disguised unemployment’, Economica, 37, 39-53.
Dixit, A. (1971), ‘Theories of the dual economy: a survey’, in Mirrlees,
J. and Stern, N. (eds), Models of Economic Growth, Macmillan.
Fei, J.C. and Ranis, G. (1964), Development of the Labour-Supply
Economy: Theory and Policy, Homewood, Ill., Irwin.
Ghatak, S. (1981), Monetary Economics in Developing Countries,
Macmillan.
122 AGRICULTURE AND ECONOMIC DEVELOPMENT

Griffin, K. (1969), Underdevelopment in Spanish America, George

Allen & Unwin, London.
Hansen, B. (1968), ‘The distribution share in Egyptian agriculture,
1897-1961’, International Economic Review, 9, 275-94.
Ho, Y.-M. (1972), ‘Development with surplus-labour population -
the case study of Taiwan’, Economic development and Cultural
Change, 20, 210-34.
Jorgenson, D. (1961), ‘The development of a dual economy’,
Economic Journal, 71, 309-34.
Jorgenson, D. (1967), ‘Surplus agricultural labour and the develop¬
ment of dual economy’, Oxford Economic Papers, 19, 3, 288-312.
Kelly, A.C., Williamson, G.G. and Cheetham, R.J. (1972), Dualistic
Economic Development: Theory and History, University of Chicago
Press, Chicago.
Lewis, W.A. (1954), ‘Economic development with unlimited supplies
of labour’, Manchester School, 22, 139-91.
Mabro, R. (1967), ‘Industrial growth, agricultural under¬
employment and the Lewis model: the Egyptian case, 1937-1965’, J.
Dev. Studies, 3, 4, 322-51.
Mackinnon, R.I. (1973), Money and Capital in Economic develop¬
ment, Brookings Institution, Washington, DC.
Mehra, S. (1966), ‘Surplus labour in Indian agriculture’, in Chaud-
huri, P. (ed.) (1972), Readings in Indian Agricultural Development,
George Allen & Unwin, London.
Nakamura, J.I. (1965), Growth of Japanese agriculture 1875-1920,
in Lockwood, W. (ed.), The State and Economic Enterprise in Japan,
Princetown University Press, N.J.
Schultz, T.W. (1964), Transforming Traditional Agriculture, Yale
University Press, New Haven.-
Sen, A.K. (1975), Employment, Technology, Development, Clarendon
Press, Oxford.
Shaw, E.S. (1973), Financial Deepening in Economic Development,
Oxford University Press, New York.
Shukla, T. (1965), Capital formation in Indian Agriculture, Vora,
Bombay.
Stone, R. (1954), ‘Linear expenditure systems and demand analysis:
an application to the pattern of British demand’, Economic Journal,
64, 511-27.
Watanabe. T. (1965), ‘Economic aspects of dualism in the industrial
development of Japan’, Economic Development and cultural Change,
14, 293-312.
 6 Resource Use Efficiency
and Technical Change
in Peasant Agriculture

In this chapter we examine the scope for increasing agricultural

output and incomes in LDCs from three aspects. First, we consider
how efficient peasant farmers are in utilising their existing resources
and the feasibility of agricultural progress and development without
technical change. Second, we consider how agricultural technical
change is generated and how it affects agricultural output and
incomes, including the distributional effects. Finally, we consider
policy implications, particularly for agricultural research and the
spread of information, and for farm mechanisation.

1 Efficiency of Resource Utilisation

In this section we attempt to clarify the meaning of various concepts

of efficiency, and how these relate to each other, before proceeding to
a critique of the hypothesis that peasant farmers are ‘poor but
efficient’. We also consider how efficiency relates to farm size
specifically in the context of LDC agriculture.

1.1 Concepts of efficiency

The term ‘efficient’ is often used ambiguously, even by economists.

So, before proceeding to examine the efficiency of peasant agricul¬
ture, we remind readers of how technical efficiency differs from
allocative (or price) efficiency, and of how economic efficiency
combines these concepts (Farrell, 1957; Jones, 1977-8).

123
124 AGRICULTURE AND ECONOMIC DEVELOPMENT

Consider a simple production function in which land and labour

are combined to produce a single homogeneous product. Thus in
Figure 6.1, isoquant SS' is an outer-bound production function, i.e.
given the current state of knowledge, it is not technically feasible to
produce the level of output represented by SS' on an alternative
isoquant which is closer to the origin, 0. Thus, any farm lying on SS',
such as Q* or Qls is technically efficient: but any farm in the space
above SS', such as Q2 or Q3, is not technically efficient.
Allocative efficiency demands that factors be combined in the same
ratios as their relative prices. The land: labour factor price ratio is
given by the slope of the isocost curve AA'. Thus Q* is allocatively
efficient, whereas Qj is not: Q2 and Q3 may or may not be allocatively
efficient, depending on whether the slopes of the (technically
inefficient) isoquants at those points (not shown in the figure) equate
with the slope of AA'.
In summary, then, Qx is technically efficient, but allocatively
inefficient. Q2 and Q3 are both technically inefficient, and may be
allocatively inefficient as well. Only Q* is both technically and
allocatively efficient. Defining economic efficiency as the combination
of technical efficiency with allocative efficiency, only Q* is economi¬
cally efficient.

Land

Figure 6.1: Allocative, technical and economic efficiency

 PEASANT AGRICULTURE 125

1.2 Efficiency of peasant agriculture

The ‘poor and efficient’ hypothesis postulates that peasant farmers

are poor, not because they utilise their resources inefficiently, but
because of restrictions in the kinds and quantities of the resources
they command. It is claimed that the results of a celebrated study of
traditional Indian agriculture support this hypothesis (Hopper, 1965;
Schultz, 1964).
Hopper fitted a Cobb-Douglas production function to empirical
data derived from 43 peasant farms in a village in northern India. The
data pertained to four major inputs and four production alternatives
(crops).1 Recall that fitting a C-D function ‘imposes’ constant
elasticities of production on the input variables: also that

^ _
MPPX *i

where ex. is factor x's elasticity of production. Thus, with knowledge

of APPX derived from the data, and ex from the production function,
an estimate of MPPX| is obtained. By choosing an appropriate
numeraire, MPPX can be transformed into MVPX. Using this
method. Hopper derived the MVPs of the four production inputs
from the data and production function parameters, together with the
implicit prices of the four products (estimated at the mean levels of
input and expected output). The results so obtained were then tested
against the norms of neo-classical profit-maximising behaviour.
The neo-classical model postulates 3 profit maximising rules. First,
that the MVP of each factor must equate with its price, i.e.

MVPXi = PXi... (1)

Secondly, that factors must be combined in the least cost factor
combination, i.e.

MPP
_2J_ _
Px. *i
(2)
MPPX Xj
P X)
' "

Thirdly, that products must be combined in the highest profit product

combination, i.e.

MVPx.yi = M VPX. y2 = ... MVPX. yt... (3)

Hopper claimed that, allowing for estimation errors, his empirical

estimates of the MVPs of the four production inputs, the prices of the
126 AGRICULTURE AND ECONOMIC DEVELOPMENT

four products, and the relationships between them, were broadly

consistent with the model. He went on to make the further inference
that, because they appeared to be maximising profits, the case study
farmers were allocatively efficient.
The results of another much larger empirical study of Indian
agriculture broadly support the Schultz-Hopper hypothesis that,
even in traditional agriculture, farmers allocate resources according
to the rules of neo-classical profit maximisation. The study covered
nearly 3000 holdings, spread over six agricultural regions and was
based on accounting records covering three consecutive years
(Yotopoulos and Nugent, 1976, ch. 6). Again, a Cobb-Douglas type
model was used to test the profit maximisation hypothesis, but in this
case, a clear distinction is made in the model between (1) price
efficiency, concerned with the allocation of variable factors, and (2)
technical efficiency, as determined by fixed factors or factor endow¬
ments. Inter-farm differences in factor productivities can be expected
to occur because of differences in factor endowments, such as the
qualities of land, labour and entrepreneurship. In fact, Yotopoulos
and Nugent found that more than 80 per cent of the observed
differences in factor: output ratios amongst farms in the study sample
were due to such ‘systematic variation’. Less than 20 per cent was left
to be explained by ‘mistakes in maximisation’. It was thus concluded
that ‘Indian farmers... seem to be remarkably price efficient’
(Yotopoulos and Nugent, 1976, p. 93). The results of this more com¬
prehensive study are more convincing than Hopper’s, not only because
of the much larger size of the survey sample (and consequent greater
representativeness of Indian agriculture as a whole) but also because
of the clear distinction made between allocative and technical
efficiency. Yet even this evidence is subject to one important caveat,
namely, that no alternative resource allocating decision rule to neo¬
classical profit maximisation was tested. More specifically, the
possibility that resource allocation in agriculture is influenced by risk-
aversion was not investigated.
Before proceeding to criticise the ‘poor but efficient’ hypothesis
and the empirical evidence offered in its support, we refer to its policy
implications. Schultz and Hopper infer that because peasant farmers
are ‘allocatively efficient’ they cannot materially increase their output
(or, by implication, their income) except through technical innovation.
They are doing the best they possibly can with their existing
resources. Thus the only feasible means of enabling them to do better
is by equipping them with better quality (i.e. more productive)
resources. A ‘technical revolution’ is essential. More formally, and
 PEASANT AGRICULTURE 127

referring to Figure 6.1, the outer-bound production function must be

shifted towards the origin.

1.2.1 Critique of the ‘Poor but Efficient’ Hypothesis and

Supporting Evidence
Recall that the neo-classical model is static and implicitly assumes
that producers are endowed with perfect knowledge. These limita¬
tions, and others, have been emphasised by critics of the Schultz-
Hopper hypothesis that peasant farmers allocate resources efficiently,
in the neo-classical mode. The empirical evidence offered in support
of the hypothesis has been criticised on the same grounds.
The main criticisms of the ‘poor but efficient’ hypothesis, and its
policy implications, concern: (i) the choice of a neo-classical model to
represent the behaviour of peasant farmers; and (ii) the distinction
between allocative efficiency and economic efficiency.
Behaviour of peasant farmers under risk and uncertainty. The essence
of the first criticism is that the restrictive assumptions of perfect
competition underlying the neo-classical model are inapplicable to
peasant agriculture. Far from enjoying perfect knowledge and
freedom to combine resources and adopt the practices needed to
maximise long-run profit, peasant farmers are in fact confronted
both by considerable uncertainty and by numerous institutional and
cultural restraints. Moreover, being poor, they tend to be risk-averse.
Recall that risk-aversion contrasts with risk-neutrality. A risk-neutral
decision-maker is indifferent between accepting: (i) a certain gain of a
fixed value; and (ii) an equi-probable gain or loss of the same average
or expected monetary value (EMV) as the certain gain.
Given the same choice, a risk-averse decision-maker will express
indifference between the EMV of the uncertain option and its lower
valued certainly equivalent (CE). In the jargon of decision theory,
CE = EMV with respect to a gamble signifies risk neutrality, whereas
CE < EMV signifies risk-aversion. A simple numerical example will
serve to illustrate the difference.
Suppose the choice lies between accepting (a) a certain gain of
£400, and (b) an uncertain outcome, with an EMV of£500, where the
equi-probable actual outcomes are a £2000 gain/£ 1000 loss; i.e.
(2000)(0.5) -F (— 1000)(0.5) = 500. If the decision-maker is risk-
neutral he must prefer (b), since £500 > £400. But if he is risk averse
he will opt for (a), unless the CE of (b) exceeds £400.
If peasants are risk-averse, then their predominant goal is economic
survival. Adequate stability of output and income, and the avoidance
of major short-run losses, take precedence over profit-maximisation
128 AGRICULTURE AND ECONOMIC DEVELOPMENT

(Lipton, 1968). The results of an empirical study of smallholder

agriculture in Kenya support the hypothesis that peasant farmers do
trade-off lower risk against higher profits. Or, in other words, they
maximise utility rather than income, where utility is a function of the
variance of income, as well as its average level (Wolgin, 1975).
Referring again to the three profit-maximising conditions of the
neo-classical model, it was postulated that if farmers lacked perfect
knowledge, and were also risk-averse, condition (2) (least cost factor
combination) would still hold, but not conditions (1) or (3). Rather,
given uncertainty and risk aversion, condition (1) is revised to:

MVPX > PXj... (la)

The size of the ‘safety margin’, represented by the positive difference

between the MVP of the factor and the factor price, depends upon the
degree of risk: the larger the risk the higher the safety margin.
Similarly, inequality enters the highest profit enterprise combination
condition, since exposure to risk is determined in part by the choice of
products and the proportions in which they are combined. Thus, if
products yx, y2... yn are ranked in descending order of risk (i.e. y x is
the most risky) condition (3) is revised to:
MVP Xj,yi > MVP
x Xj,y2
> • • >MVP„Xj>yn
v ... (3a)
v 7

Like Hopper, Wolgin estimated the MVPs of factors (land, labour,

capital and purchased inputs) by fitting Cobb-Douglas production
functions to farm survey data on alternative crops, classified by
ecological zone. The most salient results were:

(1) The MVPs of factors were generally higher than the correspond¬
ing market prices of factors thus verifying condition (la). As
expected the safety margin was smaller for subsistence crops than
for cash crops, and generally higher for purchased inputs, such as
fertiliser, than for non-purchased inputs, such as family labour.
(2) Across crops, the observed ratios of MPPs of factors were
broadly consistent with the corresponding factor price ratios,
thus verifying condition (2).
(3) The observed ranking of MVPs by crop was found to be closely
and directly correlated with the ‘marginal increment to risk’, a
direct function of crop income variance. This observation verifies
condition (3a).

A further observation, consistent with the hypothesis of risk

aversion, is that most of the 1500 farmers in the survey sample in fact
practised multi-cropping. In sum, the results of this study point firmly
 PEASANT AGRICULTURE 129

to the conclusion that if farmers are risk averse, and are obliged to
make decisions in risky situations, the neo-classical model of profit
maximisation mis-specifies their allocative efficiency.
Subject to the constraints imposed by fixed resource endowments
(e.g. farm size), the level and variance of farm income tend to be
positively correlated. The enterprises with the highest expected
profits tend also to be the most risky enterprises. The risk-averse
individual trades off the utility of increasing income against the
disutility of increasing risk. The expectations-variance (E-V) frontier
formalises this trade-off, as shown in Figure 6.2 where farm income
variance, var(Y), is plotted against the expected income level, E(Y).
The E-V frontiers of representative farms with ‘small’, ‘medium-sized’
and ‘large’ fixed resource endowments are shown by f1} f2 and f3.
Thus, for example, it is technically feasible for a medium-sized farm
to be operated either at any point on f2, or at any point in the space
to the left of f2. But, due to the fixed resource constraint, operation at
any point to the right of f2 is strictly infeasible. A prior condition of
reaching the E-V frontier is that the farmer actually operates on the
farm’s outer-bound production function, i.e. that he is technically
efficient.2 However, his preferred position on the frontier will depend
on his risk preference. The more risk he is willing to accept, in order to
gain more income, the higher up the curve he will be prepared to go:
but the greater his risk-aversion the greater the attraction of a
position at the lower south-western extremity of the frontier. A neo¬
classical producer with perfect knowledge is ipso facto immune from
risk. Such a producer must automatically opt for the highest
attainable level of income subject only to the constraint of fixed
resources. Thus, in terms of this model, neo-classical economic
efficiency demands not only that the producer is at his E-V frontier
but also that he operates at its north-eastern extremity.
A variant of the E-V frontier model was used to study the choice of
cropping patterns, relative to risk, amongst a sample of peasant
farmers in a district of India (Schluter and Mount, 1974). The initial
step was to estimate, for each farm in the survey, the particular
combination of crops likely to minimise the mean deviation of annual
net income over six years, subject to various constraints. The most
important constraints were annual income level, and available land
(classified by productivity). A parametric linear programming model
was used to derive these estimates, with the income level para-
meterised from zero to the maximum attainable level. The second
step was to compare each farmer’s actual position on the E-M map
(with E-M used as a proxy for E-V), as determined by his actual
130 AGRICULTURE AND ECONOMIC DEVELOPMENT

ro

Figure 6.2: E- V frontier model

VAR(Y)
 PEASANT AGRICULTURE 131

choice of crops, in order to assess: (i) his proximity to the E-M

frontier (i.e. his technical efficiency); and (ii) his position relative to its
economically efficient NE extremity and its risk minimising SW
extremity.
The empirical results were that although most of the cropping
patterns actually observed were quite near to their E-M frontiers (i.e.
the farmers scored relatively well on technical efficiency) the observed
incomes were generally much lower than the maximum correspond¬
ing with economic efficiency. Thus it appeared that most of the
farmers had deliberately chosen to accept a substantial reduction in
income in order to lower their exposure to the risk of income
fluctuations. In terms of the simplified model of Figure 6.2, the
observed positions of the farmers relative to the E-M frontier were
typically as signified by the asterisks. Thus the results of this study
provide further evidence that because of risk and uncertainty,
peasant farmers in LDCs aim to maximise utility and not profit in the
neo-classical mode.
If farmers were not exposed to risk (an unrealistic assumption), it
would be rational for them to maximise income subject only to the
constraint of fixed resources: a farmer who is risk-neutral can
rationally adopt the same objective function. But for farmers who are
risk-averse (due to poverty, for example) it is rational to add a
minimum profit constraint. Such behaviour can be inferred from the
results of the study by Schluter and Mount, since opting for a low
income variance is tantamount to choosing to avoid large reductions
in the level of income below EMV.
An alternative approach to modelling risk-averse decision-making
behaviour directs the primary focus of attention away from maximis¬
ing EMV to maximising the minimum profit at some predetermined
probability or confidence level. Two variants of this principle are
termed the ‘safety-fixed’ and ‘safety-first’ rules (Roumasset, 1974).
The rationale of this approach is that, corresponding with each
feasible farm production plan there is a frequency distribution of
possible profits. The farmer’s overriding concern is to avoid low
profits. In the presence of risk no level of profit can be attained with
certainty. But, given the distribution of profits, the probability of
falling below some minimum profit level can be determined (from the
cumulative frequency distribution of profit levels).
Under the safety-fixed rule, the objective is to maximise d subject to:

Pr(fl < d) ^ a
132 AGRICULTURE AND ECONOMIC DEVELOPMENT

where II = profit, d = minimum profit and a = probability (confi¬

dence) limit. The cumulative frequency distribution of profit is given
by:

Fn= j P(x)dx
X = — oo

where x = level of profit and n = a particular level of profit.

The cumulative frequency distributions of two hypothetical pro¬
duction alternatives, A and B, are depicted in Figure 6.3. In choosing
between the alternatives, A maximises d at probability limit a (since
d2 > dj even though E(II)B > E(II)A (as signified by the difference in
the magnitudes of the two shaded areas between the curves).
Under the safety-first rule, the objective is to maximise n subject
to:

Pr(II < d) ^ a

where d = a specified ‘disaster level’ of profit. Suppose that in

Figure 6.3 d = d2. Under production alternative B:

Pr(n < d) = a

Since a > a, production alternative A maximises E(FI) subject to

the minimum profit constraint d at the probability limit a. Thus,
again, because of risk-aversion, A is preferred to B even though
E(n)B > E(ii)A.
The parameter a is determined subjectively by the decision-maker,
but might be set at 0.05 or 0.10. The safety-fixed model was used, with
others, in analysing fertiliser application rates, and the use of other
cash-intensive techniques, by peasant rice farmers in the Philippines.
In this case, the S-F principle did not withstand the test of empirical
verification particularly well for two reasons. First, it was found that,
generally speaking, yield and profit distributions were such that the
goals of profit (EMV) maximisation and ‘safety’ were not in
competition. That is, there was no strong positive correlation
between average profit (EMV) and risk. There was consequently no
strong a priori reason for expecting these farmers to show risk-
aversion, and, in fact, a risk-neutral EMV maximising alternative
model fitted the data at least as well, or better, than the S-F model.
The second reason why the S-F model failed to carry conviction in
 PEASANT AGRICULTURE 133

TT

Figure 6.3: Safety fixed model

this case was that it was found that, despite the inherent risks of losses
due to crop failures or low prices, some poor farmers were prepared
to gamble on the use of cash intensive practices in the hope that a
favourable pay-off would enable them to get out of debt (Roumasset,
1976, pp. 213-14). Hence there is empirical justification for rejecting
the facile assumption that all poor farmers are necessarily risk-averse.
Where risk-aversion is characteristic of the behaviour of peasant
farmers, the actual existence of major risks and uncertainties must
constrain agricultural output and incomes in LDCs. Assuming that it
is in the national interest to relax that constraint, the obvious policy
implication is that measures to reduce the riskiness of farming
are needed. So, for example, minimum price guarantees, or better
market outlets, or readier access to agricultural credit might be
introduced.
Technical efficiency of peasant agriculture. We turn now to the second
major criticism of the Schultz-Hopper view of the efficiency of
peasant agriculture, namely, that by neglecting the distinction
between allocative and economic efficiency, it takes technical effici¬
ency for granted. Recall from the beginning of this chapter that
134 AGRICULTURE AND ECONOMIC DEVELOPMENT

whereas economic efficiency is conditional upon allocative and

technical efficiency, allocative efficiency is not conditional upon
technical efficiency, or vice versa. So, economic efficiency can be
proved only by providing explicit evidence of both allocative and
technical efficiency. It is not adequate to use explicit evidence of
allocative efficiency as implicit evidence that the same firm, or class
of firms, is also technically efficient. The allegation that Schultz-
Hopper committed this error forms the nub of the argument in
criticism of their attempt to prove the poor but efficient hypothesis.
Consider a priori reasons for expecting firms to be efficient. A
common assumption is that competition will compel efficiency,
otherwise firms risk being driven out of business. Another assump¬
tion of western capitalist economics is that producers are, above all,
motivated by self-interest and the desire to maximise their private
profits. However valid these assumptions may be in the West, it can
be questioned whether they are equally valid in LDCs, particularly in
peasant agriculture (Shapiro, 1977). One argument is that, particularly
in a mainly subsistence-type agriculture, competitive forces are likely
to be relatively weak. If a farmer gets into economic difficulty his
neighbours are more likely to help than to drive him out of business.
The same reasoning may apply even to the moneylender who may
prefer an extension of credit to foreclosure. A second and more
familiar argument is that the maximisation of utility and money
profit coincide only so long as the marginal utility of money income
exceeds the MU of other satisfactions. In reality, money income
competes with other satisfactions such as leisure, fulfilment of social
obligations, and the protection of social status. Providing that a
farmer or landowner has a ‘satisfactory’ income, by the standards of
his own community, there is no obvious reason why acquiring more
income should take precedence over adding to other satisfactions.
Bearing in mind that peasant agriculture is labour-intensive, that
physical labour is arduous (particularly in the tropics) and possibly
socially demeaning also, the notion that farmers in LDCs do not
typically strive to maximise profits possesses a considerable common-
sense appeal. This is so, quite apart from the risks of profit
maximisation as previously discussed in this chapter. But what does
choosing not to maximise profits imply for efficiency? Clearly, non¬
maximising farmers cannot achieve economic efficiency but can they
be allocatively or technically efficient? The reply to this question is
that, given the full information they need on MVPs and factor prices,
they can achieve allocative efficiency. A farmer allocating his
resources efficiently at some level of production or income below the
 PEASANT AGRICULTURE 135

outer-bound production function is effectively maximising profit at a

sub-optimum level of production (or work-effort intensity) either by
choice or due to incomplete knowledge of available production
alternatives. This, then, is a rationalisation of allocative efficiency
combined with technical inefficiency. The reverse combination -
allocative inefficiency combined with technical efficiency is less
plausible. A farmer with a full knowledge of technical production
possibilities is unlikely to lack the knolwedge and ability to equate
real marginal productivities with real factor prices. A study reported
by Shapiro (1976) attempted to subject the technical efficiency of
Tanzanian cotton farmers to objective assessment.
For the purpose of this study an outer-bound maximum technical
efficiency production function was derived by linear programming
form the application of the ‘best’ farm practices in use amongst
farmers in the sample. These best practices, relating to fertiliser rates,
numbers of sprayings, and similar matters, also accorded with
agricultural extension service recommendations. A distribution of
technical efficiency scores, ranging from 0-1 (maximum score) was
derived by comparing each farmer’s actual output with his predicted
output on the outer-bound production function. The results showed
only a small minority of farms with scores at or very near the
maximum, signifying that the majority were not operating at the
production frontier. The average score was only 0.66 and it was
estimated that, had all the farmers modified their practices so as to
move to the outer-bound production function, the aggregate output
would have been 51 per cent higher. This points to the conclusion
that, contrary to the implications of the poor but efficient hypothesis,
scope does exist in peasant agriculture for improving farm productivity
and incomes by persuading farmers to make simple improvements
in crop management practices. This conclusion is supported by other
evidence (World Bank, 1978, pp. 39-40). But it is facile to suppose
that merely by improving the quantity and quality of agricultural
extension and farmer education, the performance of every farmer can
be raised to the level of the ‘best’. Thus the authors of another recent
World Bank publication observe:

Best traditional practice is visible to all farmers and to the extent that they are
able and willing to emulate the example, they can be expected to do so
without extension advice. This suggests that differences in natural ability and
assiduity are responsible for part of the observed differences in yields, and
differences in land quality, structure of the family labour force, etc., for
another part. The part due to information gaps is probably relatively small.
(World Bank, 1981, p. 75, fn. 39)
136 AGRICULTURE AND ECONOMIC DEVELOPMENT

There is an obvious danger in attempting to estimate the produc¬

tion function relating to an individual farm or farmer from a blue¬
print based on aggregate data or ‘average practice’. In agriculture,
more than in most other industries, each enterprise is in some sense
unique. Farms differ in their physical characteristics and farmers
certainly differ in their capacity for farm work and their innate
entrepreneurial ability. Although most farmers may be capable of
improving their management through learning by experience, there is
nevertheless a sense in which each individual is uniquely constrained
by his own limitations and those of the farm. Thus, in analysing
farming efficiency, there is much to be said for adopting a disag¬
gregated approach which does not implicitly assume either that all
farms are capable of being operated at the same absolute level of
technical efficiency or that all farmers are capable of reaching the
same absolute standard of efficiency. The pitfalls of attempting to
apply standardised efficiency norms in agriculture are exemplified by
evidence on fertiliser application rates in the Philippines (Roumasset,
1976, ch. 7). The finding that, on average, actual application rates
were only about one third of the amounts recommended by the
extension service raised the question of why farmers applied fertiliser
at less than the recommended rate. The explanation of risk-aversion
was ruled out by the poor explanatory power of the S-F model, as
already discussed. An obvious alternative is the ‘learning lag’
explanation. It is unrealistic to suppose that all farmers possess full
knowledge of current recommendations, either by contact with the
extension service or through ‘learning by experience’, particularly in
LDCs where barriers to the rapid dissemination of information exist.
However, the study authors considered that the primary explanation
was that the recommended fertiliser rates were appropriate only for
farmers operating under the most favourable physical and economic
conditions. They were not appropriate for the majority of farmers in
the study who operated in less favourable conditions. In other words,
fertiliser recommendations were based on ‘idealised’ production
functions ascribing to farmers higher levels of production intensity
and technical efficiency than their fixed resource endowments would
permit. A common error is to base extension recommendations on
the results of field trials which, for a mixture of technical and
economic reasons, commercial farmers are unable to match.

1.2.2 Efficiency and Farm Size

Strictly speaking, the theory of returns to scale is concerned with the
relationship between the firm’s level of output and its long-run
 PEASANT AGRICULTURE 137

average costs, when all factors of production are varied in the same
proportion. In agriculture, scale theory has limited relevance because
of factor rigidities and indivisibilities, even in the long run. In seeking
to gain profit or utility by changing their level of output, farmers
typically vary the proportions in which factors are combined. For
example, varying the fertiliser rate alters the fertiliser: land ratio. An
important potential source of cost saving in agriculture is the fuller
utilisation of‘spare capacity’ embodied in indivisible factors such as
machinery and buildings. The term ‘economy of size’ embraces not
only economies of scale, in the strict sense of constant factor
proportions, but also economies derived from using indivisible
factors more efficiently.
Indivisibility applies mainly to capital inputs. However, the
principal inputs of traditional agriculture are land and labour.
Generally these are relatively easy to divide, though we recognise that
individual farmers may experience difficulty in gaining access to
additional land, and that some farms are too small to provide full¬
time employment. Thus, a priori, traditional agriculture might be
expected to show constant returns to scale, i.e. differences in output
between farms are broadly proportional to the corresponding
differences in the inputs of both land and labour. However, the
constant returns hypothesis is subject to two important assumptions:
first, the assumption of factor divisibility, as already discussed; and
second, the assumption of common factor prices, regardless of farm
size. The second assumption requires some elaboration.
If, in fact, the real price of land declines with increasing farm size,
whereas the price of labour increases, the land: labour ratio will tend
to vary directly with farm size, i.e., ceteris paribus, labour will be used
most intensively on the smallest farms. Under a labour-intensive
system of agriculture, where output per unit of land area is closely
and directly related to labour input, the productivity of land is largely
a reflection of the degree of labour intensity. In this situation the
productivity of land, as reflected by crop yields, for example, will tend
to be inversely related to farm size. If land is scarcer than labour,
which is often the case in LDCs, this argument carries the important
policy implication that smaller farms may utilise resources more
efficiently than larger farms. Moreover, there may be a direct link
between a country’s farm size structure and its level of aggregate
agricultural production. In an LDC context, a large farm structure
may not be consistent either with efficient resource utilisation or with
achieving the optimum level of aggregate production.
Due to measurement problems, empirical evidence on differences
138 AGRICULTURE AND ECONOMIC DEVELOPMENT

in the prices of land and labour between large and small farms in
LDCs is virtually impossible to obtain. However, there are sound a
priori reasons for expecting labour to be relatively cheap and land to
be relatively dear to the operators of small farms, with the obvious
implication that the reverse situation of dear labour and cheap land
applies to large farm operators. The rationale of the wage differential
has already been explained, under labour market dualism, in chapter
2, pp. 8-10. The expected land price (and rent) differential is also a
consequence of inequality within the agricultural sector. By behaving
monopsonistically, large-scale farmers tend to have preferential
access to the land market where they buy or rent land relatively
cheaply. But small-scale farmers are obliged to pay a higher price or
rent because of their relatively weak bargaining position in the
market. This is also discussed at greater length in chapter 2,
pp. 20-1.
The results of two major empirical studies support the hypothesis
that in traditional agriculture agricultural output per unit of land
area does tend to be inversely correlated with farm size. In an
extension of the production function study referred to earlier in this
chapter, Yotopoulos and Nugent (1976, ch. 6) compared the relative
economic efficiencies of ‘small’ and ‘large’ farms in six agricultural
regions of India. Their results appeared to justify the rejection of the
null hypothesis of equal efficiency between the two groups of farms
and the sign of the farm size dummy variable indicated that, subject
to fixed factor constraints, small farms were more profitable (i.e.
efficient) than large ones. It was deduced that the observed difference
in economic efficiency in favour of the small farms was due to their
superior technical efficiency, since, as already remarked, all the farms
covered by the study, regardless of size, appeared to be price efficient.
Thus the small farms produced more per unit of scarce resources. But
note, following our earlier discussion of the technical efficiency of
peasant agriculture, that the results of this study are concerned only
with the relative efficiency of large and small farms: they throw no
light on how the actual performance in either group compared with
any absolute standard of technical efficiency, based on complete
knowledge of available production alternatives and the ability to
select and apply the one corresponding with the outer-bound
production function.
The second source of empirical evidence on the relationship
between the intensity of land use and farm size in traditional
agriculture is a set of cross-sectional regression studies of land
productivity across farm size groups in six LDCs conducted by Berry
 PEASANT AGRICULTURE 139

and Cline (1979, ch. 4). The countries were Brazil, Colombia, The
Philippines, West Pakistan, India and Malaysia. Allowance was
made for inter-farm variations in the quality of land, either by using
the price of land as a deflator or by the separation of irrigated from
non-irrigated land. In all six countries the statistical results appeared
to confirm the existence of a negative relationship between farm size
and output per unit area of constant quality land. Output per land
area unit is only a partial productivity measure. However, where land
and labour are the dominant factors of production, as in traditional
agriculture, the productivity of land may serve as a reasonable proxy
for total factor productivity, especially if the marginal opportunity
cost of labour is low. Parallel estimates of total factor productivity
across farm sizes in the same group of countries, at estimated social
prices of factors, were broadly consistent with this expectation,
especially with a zero social cost imputed to labour.3 As emphasised
by the authors of this study, there are at least three components of
any change in agricultural output per unit of land area. These are
crop yield, cropping intensity and product mix. Thus, in principle,
output per unit of land can be changed either by changing the yields
of specific crops, or by changing the frequency of cropping (as
reflected by the ‘cropping index’ which expresses the total area of
crops actually grown per unit time period as a proportion of the total
area available for cropping), or by re-allocating the available land
amongst crops with different unit area output values. So, evidence on
comparative crop yields alone provides an insufficient basis for
measuring the relationships between land use intensity and farm size.
Indeed, available evidence suggests that the observed difference in
land productivity between large and small farms in countries with a
traditional agriculture derives principally from differences in crop¬
ping intensity and product mix, rather than from crop yield disparities
as such (Berry and Cline, 1979, pp. 13-16).
The empirical observation that, under the conditions of traditional
agriculture, output per unit of constant quality land is inversely
correlated with farm size appears to remove one of the stock
objections to land redistribution, or land reform, in LDCs, namely,
the objection that, because small farms are allegedly less efficient than
large, land reform is likely to reduce the aggregate level of agricultural
production. However, even in LDCs, a policy of adopting the small
farm as an efficiency norm is limited by two considerations. First,
since the superior efficiency of small farms is confined to labour
surplus economies, that superiority is destined to disappear as labour
becomes scarcer and its opportunity cost rises. Second, and even
140 AGRICULTURE AND ECONOMIC DEVELOPMENT

more importantly, there are sound a priori reasons for doubting

whether the superior efficiency of the small farm can be sustained in
the face of agricultural technical change in LDCs. Virtually all the
evidence on the superior efficiency and profitability of small farms,
such as that provided by Yotopoulos and Nugent and Berry and
Cline, pertains to traditional agriculture and pre-dates the ‘Green
Revolution’. This represents a major caveat to the findings of these
studies.
Although there is some justification in the argument that, being
readily divisible, some modern farm inputs are ‘scale neutral’, this
claim cannot be made with respect to other inputs associated with the
first group. So, for example, improved (i.e. higher yielding) seeds and
chemical fertilisers contrast with irrigation equipment and field
machinery such as tractors and mechanical harvesters. Even if large-
and small-scale farmers have equal access to ‘modern’ agricultural
inputs (which is doubtful), yields on large farms can be expected to be
at least as high as yields on small farms. But, due to the indivisibility
of fixed capital inputs which are an integral part of the new
technology, unit costs must tend to be lower, and profits higher on the
larger farms, save where hiring the services of indivisible or ‘lumpy’
inputs such as field machinery or cooperative machinery ownership
are feasible options for small-scale farmers. To the extent that a
higher profit per unit of output reflects superior efficiency, large farms
may well become more efficient than small as a consequence of the
modernisation of traditional agriculture. Unfortunately, there is a
dearth of empirical evidence on the relative efficiency of large and
small farms in LDCs in areas where modern agricultural technology
has been adopted. So, it is not possible to ‘prove’ that small farms have
lost (or are losing) their traditional advantage in terms of ‘efficiency’
where the scarcest resource is land. But the a priori argument for
expecting such a trend in all LDCs affected by the Green Revolution
is compelling.
Returning to the issue of farm size policy in LDCs, a dynamic view
anticipating technical change might lead to the rejection of a policy of
uniform farm size in favour of one of size diversity. There are at least
two arguments in favour of diversity (Bachman and Christensen,
1967, p. 255). First, since larger farmers generally lead in adopting
new methods, it might encourage the more rapid adoption of
technological changes in traditional agriculture. Second, because
types of agricultural enterprises vary in the scope they afford for
economies of size (e.g. contrast plantation type crops produced
primarily for export with domestic food crops), an adequate degree of
 PEASANT AGRICULTURE 141

flexibility in farm size policy is desirable in any case.

Even though production activities in traditional agriculture may
tend to show constant returns to scale, due to the absence of major
capital inputs, the constant returns hypothesis lacks credibility with
respect to the provision of external credit. Because a good many
categories of marketing cost are fixed and indivisible (e.g. the
depreciation of transport, storage and processing equipment), the
relationship between total marketing costs per sales unit and total
sales volume must be inverse, all other things being equal. Thus,
larger farmers with more to sell must enjoy a marketing cost
advantage compared with smaller producers with less output and
low sales. The economies of large-scale marketing operations tend
to be especially marked with respect to tropical export crops such
as coffee, tea, sugar and bananas (Bachman and Christensen, 1967,
p. 250). A similar argument applies to the purchase of farm requi¬
sites, such as seeds, fertilisers and pesticides. Suppliers are often
prepared to reduce the price for larger-sized deliveries, quite apart
from the possibility that larger farmers can exert more bargaining
power. The argument can also be extended to the procurement of
agricultural credit. Not only do larger-scale farmers tend to enjoy
preferential access to credit, particularly from institutional sources,
they are also in a better position to bargain for or otherwise obtain
loans on preferential terms.
There are two obvious policy approaches to ameliorating or
removing the small farmer’s external services cost handicap. The first
approach, favoured in socialist countries, is ‘collectivisation’, under
which large numbers of formerly independent small farms are merged
to form very large-scale production and marketing units with
centralised administration and control. The rank-and-file members
of the collective lose virtually all of their former independence with
respect to both production and marketing. The second approach is
‘co-operation’ for the procurement of specific agricultural services
such as marketing and credit. Under this approach farmers’ resources
are pooled and their independence compromised only in relation to
the provision of specific common services which are expected to yield
economies of scale, or even enable small producers to exert some
group bargaining power. In this way, small farmers aim to procure
services of the standard enjoyed by large farmers at a similar cost. Co¬
operative mambership may be either voluntary or compulsory,
provided that government is prepared to legislate the necessary
powers of enforcement. Whereas voluntary co-operatives tend to be
more actively supported by their members, compulsory co-operatives
142 AGRICULTURE AND ECONOMIC DEVELOPMENT

or, as they are more usually termed, marketing boards, are generally
able to exert great bargaining power in negotiating prices. A
marketing board may even enjoy monopoly trading rights.
Collective farming, the principles of co-operative marketing and
credit, and the activities of agricultural marketing boards are all
discussed at more length in chapter 8. The Chinese system of
collective farming is referred to in chapter 11.

2 Technological Change in Agriculture

We open this section by examining the character of agricultural

technical change, with special reference to HYV technology. We then
consider the generation of technological change in agriculture,
including the theory of induced technical and institutional change.
Next, we review the origins, spread and economic consequences of
HYV technology in LDCs, including both direct and indirect effects
together with relevant empirical evidence. This leads on to consider¬
ing the distributional consequences of factor biases embodied in new
agricultural technology, both in theory and specifically with respect
to the effects of farm mechanisation on agricultural employment in
practice. Finally, we consider policy implications relative to the
discovery and selection of new agricultural technologies that are
appropriate to the economic circumstances of LDCs, emphasising
the role of government.

2.1 Nature of technological progress in agriculture

A technological improvement possesses two general properties. First,

a new production function is created such that any given quantity of
resources yields a larger product, at least above some threshold level
of resource input. Second, the proportions in which resources are
combined to produce a given output at least cost are generally
changed. We now proceed to examine the first property in more
detail, but defer further discussion of the second property to
section 2.6 below.
Technological progress in commonly conceptualised as a ‘shift
variable’ which, in graphical terms, shifts the production function
positively or ‘vertically’. In Figure 6.4(a) and (b), X represents any
variable input (with all complementary inputs assumed fixed) and Y
output. The functional relationship between X and Y depends upon
the choice of technology, with t2 representing an improvement on t}.
 CM
PEASANT AGRICULTURE

o
Figure 6.4: Technological progress
143
144 AGRICULTURE AND ECONOMIC DEVELOPMENT

Thus, supposing X is the fertiliser application rate and Y is the yield

of wheat, tt might represent a traditional wheat variety and t2 a new,
higher yielding variety. Note that in this example the technological
improvement is not ‘embodied’ in the variable input (fertiliser) but in
one of the fixed inputs (the seed). The new, higher yielding variety is
more responsive to fertiliser than the more traditional variety in two
senses:

(1) the yield per unit of fertiliser is higher above the threshold
application rate (0 in 6.4(a), and a in 6.4(b)); and
(2) the economic response to fertiliser extends to a higher rate of
application: tangents p and p' represent a (constant) fertiliser:
wheat price ratio, and the points of tangency between p(or p') and
tx(or t2) signify the optimum application rates (subject to the
usual neo-classical assumptions), i.e. profit is maximised where

<5y/<5x = Px/Py.

There is no reason, either in principle or in practice, why the

technological improvement should not be embodied in the variable
input (rather than in one of the fixed inputs). Thus, to return to our
example, the improvement might be embodied in the fertiliser instead
of in the seed. Any innovation increasing the available crop nutrients
per unit weight of fertiliser (including filler and other impurities)
would have this effect. Empirically, it may be difficult to identify
precisely sources of technological improvement. Production func¬
tions are estimated and are observed to shift dynamically. But the
reasons for the shift are unknown. Technological improvement
effectively becomes ‘disembodied’. However, this is merely an expe¬
dient to conceal the analyst’s ignorance of the technological reasons
for the observed shift and, in this sense, the concept of ‘disembodied
technical change’ lacks credibility.
With some reservations, technological progress is beneficial both
to individuals and to society as a whole. Producers adopting
improved techniques of production benefit, at least in the short run,
from an increase in profits. Consumers, and the nation, stand to gain
from increased aggregate supplies, either through the relief of actual
physical scarcity, or lower prices, or both. In the present state of the
world, there can be few LDCs where, ceteris paribus, any increase in
domestic agricultural production is not socially beneficial through its
effects on supplies available for domestic consumption, real food
prices and the balance of overseas payments. Through the favourable
 PEASANT AGRICULTURE 145

effect on crop yields, even subsistence farmers stand to gain from

adopting HYV varieties. They are able to produce more for their own
consumption with the same effort (or the same amount with less
effort). Commercial farmer benefits are largely confined to an increase
in producer surplus, which may be eroded as time passes. If the rate of
growth in aggregate food supplies should eventually excceed the rate
of growth in aggregate food demand, agricultural product prices
must decline in real terms to the detriment of commercial farmers and
the benefit of consumers. Whereas commercial farmers as a group
suffer a loss of producer surplus, there is a gain in the surplus accruing
to consumers. As the sole or principal consumers of their own output,
subsistence farmers also benefit from any gain in consumer surplus.
But, selling little or nothing, they are unaffected by losses (or gains) in
producer surplus (Hayami and Herdt, 1977).
But there tend to be losers as well as gainers from technological
progress-hence the reservations. As far as the modernisation of
agriculture is concerned, the possible losers include farmers who are
unable or unwilling to adopt better methods. They may also include
landless labourers, and even some classes of domestic food con¬
sumers. We discuss the contrasting effects of land-augmenting and
labour-displacing agricultural technologies on output, employment
and income distribution in section 4 of this chapter. Possible barriers
to the adoption of improved agricultural technology in LDCs are
fully discussed in chapter 2 section 1.3.2. As shown there, vulnera¬
bility to such barriers is particularly marked in the case of large
numbers of peasant farmers operating on a very small scale, due to
the combination of inadequate farm size, poverty, risk-aversion,
imperfect knowledge and other market failures resulting in discrimina¬
tion against small producers. Non-adopters are unable to share the
benefits of increased physical production and productivity (especially
of land) that are enjoyed by adopters. If they produce a marketable
surplus, their lower physical productivity (e.g. in crop yields) is
particularly costly in terms of income forgone if, due to a shift in the
overall balance of aggregate supply and demand, there is a real
decline in the market price. Although, ceteris paribus, adopters and
non-adopters alike suffer an absolute loss of income because of the
price decline, adopters remain relatively better off due to the buffer
provided by their larger output and sales volume. The non-adopters
lack this buffer.
It is also recognised that the benefits of technological progress are
more certain for landowners than for tenant farmers. The latter,
whether they are potential adopters of the improved technology, or
146 AGRICULTURE AND ECONOMIC DEVELOPMENT

not, risk dispossession by their landlord. The increased farm profits

afforded by technological advances give landlords an added induce¬
ment to re-possess tenanted land in order to farm it themselves. Even
without repossession, a major proportion of the gains from the
adoption of new technology is likely to accrue to landlords through
increased rents. Rent equilibriates the demand for rented land with
the available supply. The demand for rented land function shifts to
the right as the expected level of farm profit rises. But the supply is
inelastic for two reasons. First, the area of land beyond the existing
extensive margin of cultivation is extremely limited, especially in
overpopulated countries. Second, landlords have a greater incentive
to retain, or even repossess land for their own cultivation. Hence most
of the excess demand for agricultural land resulting from techno¬
logical progress in farming (and the consequent rise in farm profit
expectations) is removed, not by any substantial increase in the
supply of land suitable for cultivation, but by the escalation of rent to
a higher equilibrium level. Tenant farmers must pay a higher rent to
secure (or retain) tenancies. Landlords benefit from an increase in
wealth as well as from a higher rental income. Since, theoretically, the
value of land is merely its capitalised rent, a landowner’s wealth must
rise as a consequence of each rent increment. A principal argument
for land reform, under which the ownership of agricultural land is
transferred from large landlords to their former tenants, is to achieve
a more equitable distribution of the benefits of technological progress
in agriculture. However, even after an equalising land reform,
technological progress still drives up the price of land to those
wishing to enter agriculture by farm purchase, including landless
labourers. With land wholly or mainly under private ownership,
controlling the price or taxing windfall gains is fraught with many
difficulties, including the undesirable consequences of discouraging
landowners from offering land for sale.
Despite reservations concerning the distribution of its benefits,
there can be little serious doubt that LDCs are considerably better-off
with technical advances in agriculture than without them. In terms of
the relief of hunger alone, the Green Revolution must have been
beneficial in the countries affected by it (Lipton, 1978). Under
favourable conditions, nearly all categories of farmers in LDCs,
regardless of farm size, tenure, access to credit, attitude to risk, and
other constraints, can benefit from adopting HYV technology.
Despite a tendency for smaller farms to lag behind larger farms in the
adoption process they eventually catch up (Feder et al., 1981). As
 PEASANT AGRICULTURE 147

discussed in chapter 2, the consequences of technological stagnation

are continuing low labour productivity on farms and, for most
farmers, very low and stagnant incomes. Because of the scarcity of
certain agricultural resources - particularly land - the rate of aggre¬
gate agricultural growth and development in most LDCs is critically
dependent on the rate of technological progress. Moreover, due to
agriculture’s critical role in fostering overall development in LDCs,
as discussed in chapter 3, the importance of an adequate rate of
technological progress in farming extends beyond agriculture itself to
the national economy as a whole.

3 Generation of New Agricultural Technology

Technological innovation is a two-stage process of invention or

discovery, and adoption of the improved input or method of
production by producers. Although an innovation can have no
economic impact unless and until it is adopted by producers,
adoption is logically preceded by invention. Despite the recognition
by economists that scientific discoveries and inventions can explain
faster economic growth and development (or their absence explain
economic stagnation) the traditional approach to modelling
economic growth has been to treat invention as an exogenous shift
variable. In other words, traditionally, there has been no economic
theory of invention or technological improvement.
An alternative to the exogenous shift variable approach is to
assume that a stock of unused inventions is always at the disposal of
producers, i.e. adoption never catches up with invention, so that
invention per se can never impede growth. Such an assumption
underlies Boserup’s contra-Malthusian theory that, even in poor
countries, the response of food supplies to population growth is
elastic (see chapter 9). Boserup believes that in primitive agriculture
farmers in the aggregate do not actually adopt more productive
technologies until forced to do so by population pressure. However,
the credibility of this theory is dependent on what many critics regard
as dubious assumptions.
As far as agriculture is concerned, there is a more credible
alternative to the traditional ‘manna from heaven’ approach to the
generation of new technological discoveries. This is the theory of
induced technical and institutional change (Hayami and Ruttan, 1971,
148 AGRICULTURE AND ECONOMIC DEVELOPMENT

ch. 3; Ruttan, 1974). The crux of this theory is that the research and
investment which necessarily precedes new discoveries leading to
technical progress is induced by market forces. In agriculture,
changes in the relative scarcities of resources (as expressed by price
changes), especially land and labour, induce a derived demand for
technological innovations to facilitate the substitution of relatively
less scarce and cheap factors for more scarce and expensive ones. For
example, in a labour-scarce economy there is a tendency for capital in
the form of labour-saving machinery to be substituted for human
labour. But, in a land-scarce economy, yield-increasing and land¬
saving inputs such as fertilisers, irrigation and HYVs are substituted
for land.

3.1 Induced technical change

Hayami and Ruttan have evolved a meta-production function

hypothesis to explain how induced technical change increases the
elasticity of response to factor price changes. Consider first varying
the amount of a single input factor, such as fertiliser, in response to a
change in the factor: product price ratio. Although farmers can
normally be expected to increase fertiliser inputs in response to a
decline in the fertiliser: crop price ratio, the amount of the fertiliser
increment and crop yield increment corresponding with it may both
be only comparatively small unless new crop varieties are developed
which are more responsive than traditional varieties to fertiliser
application.
In Figure 6.5 the fertiliser application rate is measured along the
horizontal axis and the crop yield along the vertical axis. All other
factor inputs, including land, are assumed to be fixed: u0 and u, are
fertiliser response curves, respectively, representing traditional and
improved, higher-yielding crop varieties. The tangents P0 and Pj
represent different fertiliser: crop price ratios. With the farmer’s choice
of variety confined to u0, a decline in the fertiliser: product price
ratio from P0 and P, would justify only a relatively small increase in
the rate of fertiliser application, from Of, to Of2, corresponding with
a rise in crop yield from Oy, to Oy2. Now suppose in response to the
same price change, the higher-yielding crop variety u, is evolved.
Farmers are now able to switch from u0 to u, as well as varying the
fertiliser rate. Compared with the original equilibrium at Of,, Oy,,
the new profit-maximising equilibrium is at Of3, Oy3. The fertiliser
increment has increased from Of2 - Of, to Of3 - Of„ and the yield
increment from Oy2 - Oy, to Oy3 - Oy,. It is concluded that with
Crop
Yield
PEASANT AGRICULTURE

Figure 6.5: Induced technical change

149
150 AGRICULTURE AND ECONOMIC DEVELOPMENT

the discovery or development of more fertiliser-responsive crop

varieties, any given decline in the price of fertiliser will have a greater
impact on both fertiliser usage and crop yield.
Conceptually, u0 and u, are a pair of short-term fertiliser response
curves drawn from a much larger ‘family’. The long-term relationship
between the rate of fertiliser application and crop yield, assuming the
dynamic development of even more fertiliser-responsive crop
varieties, is shown in Figure 6.5 by the envelope curve U. This is
Hayami and Ruttan’s meta-production function. In effect, this
function represents the very long-run when all conceivable technical
alternatives might be discovered.
The behavioural rationale of this model is that a decline in the price
of a single variable factor (fertiliser in the example) provides
producers (or their industry representatives) with an economic
incentive to press the R & D ‘industry’ to discover and develop new
technology possessing the property of making output more input
responsive. That is, the economic rate of input usage at the new lower
price level would be substantially increased compared with the rate
justified by existing technology. Similarly, it .can be argued that
changes in factor prices give guidance to the discoverers of improved
technology regarding types of technological advance with the best
‘market prospects’. However, this version of the model over-
simplifies^ reality because, in the real world of input factor substi¬
tution, th*e incentive for promoting technological change may equally
well derive from changes in factor/factor price ratios.
In Figure 6.6 units of fertiliser are measured on the horizontal axis
and units of land on the vertical axis above the origin. Below the
origin the vertical axis measures research input directed to the
discovery of more fertiliser-responsive crop varieties. In the upper
quadrant the short run isoquants u0 and ut represent the same level of
output, but different crop varieties, with u0 being less fertiliser-
responsive than ut. Given an initial fertiliser: land price ratio as
represented by the tangent P0, the least-cost factor combination is
Oli land and Oft fertiliser. The corresponding research input is Ott.
Suppose that a fall in the fertiliser: land price ratio to the level
represented by tangent is matched by an increase in research input
to Ot2. The research pay-off is shown by movement down the scale¬
line in the lower quadrant of the diagram from Otx, Ofx to Ot2, Of2.
Fertiliser level Of2 corresponds with the discovery of crop variety ut.
With that discovery LCFC equilibrium shifts from 01l5 Ofx on u0, to
012, Of2 on u2. This shift entails a substantial absolute increase in the
use of the cheaper input (fertiliser) and a large reduction in the use of
 PEASANT AGRICULTURE 151

the dearer input (land). This substitution of factors may be regarded

as taking place along the envelope curve U', or the long-term isoquant
corresponding with the short-term isoquants u0 and ut. In this model
the meta-production function is the complete ‘family’ of long-term
isoquants representing all possible levels of production.
This model emphasises the link between the technological or
research input and the discovery of innovations which broaden the
scope for factor substitution in response to price change. It can
readily be extended to all factors and from biological to mechanical
technology. Advances in mechanical technology are generally
labour-saving. When the price of labour rises relative to the price of
land, farmers are consequently induced to press agricultural en-
152 AGRICULTURE AND ECONOMIC DEVELOPMENT

gineers to develop new machinery with an enhanced capacity for

substituting land for labour, i.e. bigger and more powerful machines
with the capacity for enabling each worker to handle a larger land
area are demanded. We shall return to capital — labour substitution,
under the heading of the distributional consequences of adopting
factor-biased technology, in a subsequent section of this chapter.
The meta-production function hypothesis has been shown to give
a statistically acceptable explanation of contrasting past patterns of
agricultural growth in the USA and Japan (Hayami and Ruttan,
1971, ch. 6; Ruttan, 1974). The historical record shows that both
countries achieved similar rates of agricultural growth over the
period 1880-1960, though with different technologies and factor
mixes. Whereas growth in the output of US agriculture was primarily
based on improvements in mechanical technology (reflecting labour
scarcity), in Japan improvements in yield-increasing biological
technology (reflecting land scarcity) were dominant. Variations in
factor proportions - measured by land: labour, machine - power:
labour and fertiliser: land ratios - were satisfactorily explained in
both countries by changes in factor price ratios. Elasticities of factor
substitution were estimated and used as indicators of the influence ot
induced technical innovations on the rate of growth. It was consi¬
dered a priori that fixed technology would afford comparatively little
scope for substitution. Thus, the fact that empirically the elasticity
coefficients proved to be quite large was interpreted as evidence in
favour of the induced technical innovation hypothesis. The study
concluded that

Development of a continuous stream of new technology, which altered the

production surface to conform to long-term trends in factor prices, was the
key to success in agricultural growth in the US and Japan. (Hayami and
Ruttan, 1971, p. 135)

3.2 Induced institutional change

In an LDC context an important element of Hayami and Ruttan’s

theory is that induced technical change tends to be impeded by
institutional barriers. In particular, LDCs generally lack adequate
agricultural research institutions to foster the discovery and appli¬
cation of new scientific and technical knowledge. Institutional
innovation is consequently, needed to break this bottleneck. In
other words, technical innovation and institutional innovation are
complementary.
 PEASANT AGRICULTURE 153

Knowledge of how to create effective research institutions in LDCs

is only fragmentary. Though true, the observation that it involves the
recruitment of suitable personnel, training and finance amounts to
little more than a statement of the obvious. However, the main
responsibility for organising and financing research is likely to fall on
government for two reasons. First, the prospective commercial pay¬
off from agricultural research in LDCs may be too small or uncertain
to attract private investors. The manufacture and distribution of
agricultural machinery, chemicals and other modern inputs is
increasingly dominated by large international companies. Although
such companies spend heavily on R & D they are primarily interested
in developing new products for mass markets, principally in the
developed countries. The local problems and research needs of small-
scale peasant farmers in poor countries tend to be neglected. Because
the private sector is unlikely to cater adequately for its research needs,
peasant agriculture must perforce rely principally on government for
a research programme tailored to its particular requirements.
The second reason for arguing that government must assume the
main responsibility for agricultural research in LDCs is that, for
reasons of self-interest, research conducted in the private sector is
likely to be biased in favour of capital intensive and labour-saving
mechanical innovations. These are generally not well-suited to the
requirements of peasant farmers, for whom capital is scarce and
dear whereas labour is plentiful and cheap. Rather their need is for
labour-intensive, biological innovations such as HYVs. The bias
towards mechanical technology stems from the partitioning of the
benefits of innovations between their suppliers and their users, i.e.
between manufacturers and farmers. The distribution of benefits is
principally determined by the extent to which the discoverer of an im¬
proved product is subsequently able to monopolise its supply. With
competitors free to copy and market his design, most of the benefits
rapidly pass to them and ultimately to users who get the new product
at a competitive price. Whereas machinery manufacturers can
protect new designs by taking out patent rights, which may even be
respected internationally, the fruits of biological research, (such as
the breeding of improved crop varieties) are much harder to protect.
Because they are inconspicuous, some biological materials - such as
seeds - are easily stolen and smuggled across international frontiers.
All such materials are readily producible in the hands of competitors.
Whereas the research programmes of firms in the private sector are
necessarily based on self-interest, research in the public sector
is directed towards discoveries which promote social interests.
154 AGRICULTURE AND ECONOMIC DEVELOPMENT

Government is responsible for national agricultural policy and the

only way of ensuring that a country’s agricultural research pro¬
gramme is compatible with its policy objectives is for government
to oversee the research programme and provide many of the
resources needed for its implementation. By this means the in¬
stitutional barriers to induced technical change may be removed.

3.3 International agricultural research institutions

Much of the agricultural research which has benefited LDCs in

recent years has been conducted by international research insti¬
tutions, such as the International Maize and Wheat Improvement
Centre (better known as CIMMYT, the initial letters of its Spanish
title) located in Mexico, and the International Rice Research Institute
(IRRI) situated in the Philippines. These institutions, originally
financed by the Rockefeller Foundation and other large charitable
foundations in the developed countries, but more recently by an
international consortium of aid agencies and financial institutions -
the Consultative Group on International Agricultural Research
(CGIAR)-have conducted disinterested agricultural research on
behalf of the LDCs as a contribution to their agricultural deve¬
lopment. The main emphasis of this research has been the dis¬
covery of biological innovations to raise crop yields. The fruits of
this research include the high yielding varieties (HYVs), particularly
of wheat and rice, which have revolutionised agriculture in some
LDCs since the mid-1960s. The work of the international research
institutions is complementary with that of national institutions in the
LDCs. An important function of national research organisations is to
conduct local trials to select imported HYVs for adaptation to local
conditions.

3.4 Evaluation

Although the theory of induced technical and institutional change

represents an undoubted advance on treating technical progress as
‘manna from heaven’, it is vulnerable to the criticism that whereas its
underlying behavioural assumptions are neo-classical, these do not fit
the conditions of peasant agriculture in LDCs. Thus, it is assumed
that farmers are profit-maximisers and that their attitude to technical
change is strictly market oriented. Moreover, it is taken for granted
that farmer possess full information on factor substitution rates and
factor price ratios. Whist these assumptions may be acceptable as an
 PEASANT AGRICULTURE 155

approximation to reality as far as farmers in developed capitalist

countries are concerned, their validity for LDCs is much more
questionable. The meta-production function model explains the
contrasting patterns of past agricultural development in the USA
and Japan quite well, but these are both developed countries. For
reasons discussed in chapter 2, as well as in section 1 above, the neo¬
classical paradigm may not adequately explain the behaviour of
peasant farmers. The barriers to technical change in peasant agricul¬
ture may be virtually insurmountable in the short term, for cultural
and social reasons, as well as because of economic uncertainty, risk-
aversion and lack of knowledge. Thus, where agriculture is most
backward (by western capitalist standards) the theory of induced
technical change may have little immediate relevance. Under such
conditions, it is not a question of agricultural research institutions
responding to market forces initiated by profit-oriented farmers. On
the contrary, assuming that technical progress in agriculture is a valid
policy goal, the initiative must come from the research institutions
and government to persuade farmers to adopt new technology in
their own and the national interest. The rate at which farmers can be
won over to accepting new technology is likely to depend in part on
the success of measures to reduce the number and severity of
economic uncertainties confronting them. It may be possible to
reduce uncertainty relatively quickly by means such as minimum
price guarantees, but to the extent that it is also necessary to break
down social and cultural barriers to technical change, progress in
achieving the goal of a rapid rate of technical advance in agriculture
may be much slower.
We have discussed reasons for technological stagnation in LDC
agriculture and the nature of barriers to technical change in chapter 2
section 1.3.2. We shall not repeat the points made in that discussion
here but merely remark that they are also relevant to the subject-
matter of this chapter.

4 Factor-Biased Technological Change and its

Distributional Consequences
Economic theorists distinguish between three types of technological
change based on factor input ratios as follows:

(1) Hicks-neutral technological change under which despite the

change in output, the capital - labour (K/L) ratio remains
constant.
156 AGRICULTURE AND ECONOMIC DEVELOPMENT

(2) Harrod-neutral technological change under which the capital-

output (K/O) ratio remains constant, so that the factor propor¬
tions are biased in favour of saving labour, i.e. the K/L ratio
increases.
(3) Solow-neutral technological change under which the labour-
output (L/O) ratio remains constant, so that factor proportions
are biased in favour of saving capital, i.e. the K/L ratio
diminishes.

Thus Hicks-neutrality marks the dividing line between labour-saving

Harrod-neutrality, and capital-saving Solow-neutrality.
Although the concept of a constant K/L ratio possesses some
theoretical interest as a reference datum, in practice technological
change usually involves a change in factor proportions. Moreover,
the change in factor proportions must imply some change in the total
levels at which factors are employed. However, the adoption of
factor-biased new technology does not necessarily imply that, in
total, less of the ‘disfavoured’ factor will be employed. So, for
example, far from diminishing the total demand for labour, the
adoption of labour-saving Harrod-neutral technology may initiate
a process of economic adjustment which actually creates more
jobs or lengthens working hours. A 2-factor isoquant diagram
will serve to clarify the analytical argument (Donaldson and
Mclnerney, 1973). In Figure 6.7, isoquants P0 and Pt represent
identical levels of output but different techniques of production, i.e.
they exist on different production surfaces. But the technique
underlying P0 is more labour-intensive (or less capital-intensive) than
the technique underlying Pt. The higher level isoquant Pt belongs to
the same family as Pt. The slope of the isocost line Ox represents the
K/L price ratio, and similarly with the isocost lines 02 and 03,
representing higher levels of total outlay. Suppose that initially a
producer is in profit maximising equilibrium where P0 is tangent to
02, giving employment to 01t labour and OKt capital. Factor
proportions are signified by the slope of the ray L/K. Suppose now
that the producer switches to a more capital intensive technique of
production without adjusting the level of output. At the new equili¬
brium where Pt is tangent to Ol5 a lower labour input of 012 is
combined with a higher capital input of Ok2. The difference in slope
between the rays L/K and L'/K' signifies the capital-using bias of the
shift to the new technique.
But, assuming constant prices, P, is not the profit-maximising level
of output. Compared with the initial equilibrium on P0, total costs
 PEASANT AGRICULTURE 157

Figure 6.7: Factor-biased technological change

have been reduced from level 02 to level Thus, assuming that

factor proportions are now fixed at L'/K', the new profit-maximising
level of output must be further from the origin than 012, Ok2.
Expanding output along any scale-line, such as L'/K', necessarily
entails employing more of both factors. Suppose that, following the
switch to the more capital-intensive production technique, the profit-
maximising level of output increases to Pt. Then, the factor input
levels are correspondingly increased to 013 labour and Ok3 capital.
At 013 the employment of labour is in fact at a higher level than the
original profit-maximising equilibrium level of 01t on P0.
Depending upon the direction of the change, a simultaneous
change in the factor-price ratio may either reinforce or offset the effect
of a change in the choice of production technology on total factor
employment levels. Thus, the effect of any decline in the K/L price
ratio will be to ‘flatten’ the isocost lines shown in Figure 6.7. This gives
producers the inducement to adopt an even more capital-intensive
alternative technology than before, i.e. the ray L'/K' is also flattened,
signifying an even larger change in factor proportions compared with
the initial choice of technology. Contrarily, if the emergence of the
more efficient capital-using technology happened to coincide with a
permanent increase in the K/L price ratio (implying cheaper labour
158 AGRICULTURE AND ECONOMIC DEVELOPMENT

and dearer capital) a ‘steeper’ L'/K' would reduce the change in factor
proportions induced by the new technology. But whatever the change
in the factor price ratio, and its effect upon the choice of technology
and associated factor proportions, the conclusion regarding the effect
of technological choice on factor employment levels remains unchan¬
ged. This is that changes in factor input requirements are jointly
determined by the change in factor proportions and the change in the
profit maximising level of output. For this reason, it is impossible to
determine a priori whether, for example, the adoption of labour
saving technology will really displace labour or actually create more
jobs.

4.1 Land-augmenting and labour-displacing technical change

Although the adoption of new technology with a labour-saving bias

does not necessarily displace labour, it may do so, depending on the
relative strengths of the ‘changed factor proportions’ and ‘output
adjustment’ effects. From the standpoint of policy, the risk of labour
displacement may be reduced by discouraging technological change
with a capital bias. Indeed, if increasing agricultural employment is a
prime policy objective, encouraging producers to adopt labour-using
technologies is an appropriate means of achieving that goal. In this
context, and particularly in the literature on farm mechanisation
policy in LDCs, a distinction may be made between land-augmenting
and labour-displacing technical change (e.g. Yudelman et al., 1973,
chapter in.)
Land-augmenting technical change induces an increase in total
output, either through enlarging the cultivated area or through
higher crop yields - ceteris paribus, higher output will tend to
increase labour requirements. Labour-displacing technical change
saves labour but does not direcly affect output. Its economic rationale
is to substitute some other input for labour in order to reduce unit
costs. But, as readers of this chapter have already been reminded, cost
reduction usually has a disequilibriating effect upon the level of
output which maximises profits. Thus, in theory at least, technical
change which is ostensibly labour-displacing must also exert some
countervailing pressure on the demand for labour through the
indirect effect on output. Whether this occurs in practice is an
empirical question.
The most obvious examples of LATC are biological and chemical
innovations which increase crop yields. But mechanical innovations
cannot be excluded, ex hypothesi, because they also may be output
 PEASANT AGRICULTURE 159

increasing, either through a crop area or a crop yield effect. If a

farmer has unlimited land at his disposal the maximum area which he
can feasibly cultivate during any time period is constrained by the
available supplies of human labour and mechanical power, parti¬
cularly during critical periods of the year such as sowing and
harvesting seasons. Thus, if a mechanical innovation is used to break
a seasonal bottleneck in the supply of farm labour its effect can be to
increase output by bringing more land into cultivation.
Employment remains unchanged, or may even increase during
other periods of the year, due to the enlargement of the cultivated
area. A mechanical innovation can still be land-augmenting, despite
an inelastic land supply, if it is the means of increasing the intensity of
cropping. So, for example, given a fixed land area, the feasibility of
growing two crops a year rather than only one may hinge on the use
of mechanical power to accomplish essential field operations within a
critical time period. The conditions which are least conducive to
LATC are the combination of a plentiful labour supply with very
inelastic land and a cropping system which is already highly intensive.
Under such conditions, it is difficult to see how farm mechanisation
could be land-augmenting, unless mechanisation possessed some
yield increasing property. This is possible provided that the applica¬
tion of mechanical power enables some yield-augmenting task to be
accomplished which is beyond the capacity of unaided human
labour. So, for example, if there is a direct cause- and-effect relation¬
ship between depth of cultivation and crop yield, and the optimum
depth is attainable only with the assistance of machinery, this
condition is satisfied. Seminal work on the economic forces underly¬
ing agricultural mechanisation in LDCs, with special reference to the
adoption of tractors in Brazil, is reported in Binswanger and Ruttan
(1978), chapter 10.
Our discussion of land-augmenting and labour-displacing techni¬
cal change points to the conclusion that they are ambiguous
concepts. LATC is supposed to increase output whilst maintaining,
or even increasing, the demand for labour. LDTC is supposed to
reduce costs by saving labour without varying output. Although it is
tempting to think of LATC and LDTC as policy alternatives, they
may in fact occur at sequential stages of technological progress. Thus,
in the first stage, output and the demand for labour both increase due
to the adoption of LATC. At first, the supply of labour is elastic, but
becomes increasingly inelastic as agricultural output and the demand
for labour both continue to grow. Eventually, seasonal bottlenecks in
labour supply emerge and, at this stage, LATC leads to LDTC.
160 AGRICULTURE AND ECONOMIC DEVELOPMENT

Although the ostensible motivation for LDTC may be cost saving,

the relief of an acute seasonal labour scarcity may be a more
fundamental but hidden reason. Once LDTC has been adopted for
one purpose, it tends to be extended to other tasks at the same or a
different period of the year. One reason for this is that, other things
being equal, it pays to intensify the use of a capital asset in order
to spread its fixed costs over a larger output.
A further reason for the ambiguity of the supposed distinction
between LATC and LDTC is that some new agricultural inputs
embodying technical change can either be land-augmenting or
labour-displacing, depending on how they are applied. Thus, for
example, a farmer may in principle acquire a tractor either to save
labour (by substitution) or to increase output (by extending his
cultivated area). But these opposite motivations are difficult to
observe in practice and, in any case, having acquired the tractor for
one purpose, the farmer may change his mind and use it for the other
one, or even for both. The near impossibility of identifying inputs which
are invariably labour-displacing has important policy implications,
especially for mechanisation policy. But before proceeding to policy
conclusions in the final section of the chapter, we briefly consider
some empirical evidence on how agricultural technical change,
including farm mechanisation, has affected employment in agricul¬
ture in certain LDCs.

5 Agricultural Technical Change and Agricultural

Employment: Empirical Evidence

We remarked in chapter 3 that because of employment linkages

between agriculture and other industries, agricultural growth creates
additional jobs outside agriculture, as well as in farming itself. In the
same way, agricultural technical change may affect employment, not
only in agriculture itself (as discussed in the previous section) but also
in all industries linked with agriculture in the input-output matrix.
The first of these may be termed the direct employment effects, the
second the indirect effects.

5.1 Direct employment effects

An empirical study of the direct employment effects of agricultural

technical change in the Indian Punjab from 1968-9 to 1973-4
 PEASANT AGRICULTURE 161

(Krishna, 1975) was based on a theoretical model attributing the

technologically-induced change in agricultural employment between
two periods to the following three causal variables:
(1) a pure ‘technology effect’, reflecting factor bias;
(2) a ‘crop-mix effect’, reflecting the change in the allocation of land
from a given area amongst different crops (with different labour
requirements) associated with the change in technology; and
(3) an ‘intensity effect’, reflecting change in the index of cropping
intensity (100 = one crop per annum).

The principal crops were wheat and rice. The empirical findings
showed an overall decline of 64 man hours per hectare per annum in
the total labour input for both crops over five years. But the crop mix
and intensity effects were both positive (totalling a gain of approxi¬
mately 45 man hours per hectare). Thus the net loss in direct
employment was entirely due to a large negative technology effect
(approximately 109 man hours per hectare).
The results of a study of introducing tractors to farms, also in the
Punjab, confirmed Krishna’s finding that more intensive cropping is
favourable to on-farm employment, but not the finding of an overall
net loss of employment due to technical change (Roy and Blase,
1978). In order to find out how the introduction of tractors affects
cropping intensity, crop yields, and on-farm employment, a cross-
sectional study of farms with tractors (TFs) and farms without
tractors (NTFs) was undertaken. The results indicated that the
principal effects of tractorisation had been more intensive cropping
and higher crop yields. These gains were attributed in part to the
better ‘quality’ of tractor ploughing. But despite some differences
between TFs and NTFs in.the proportions of hired labour and family
labour, as well as in the relative numbers of permanent and casual
workers, there was no significant difference between them in total
labour use per unit of net crop area.
An on-farm sample survey of farm tractors supplied to India under
the British overseas aid programme reached a similar conclusion,
namely, that although tractors had affected cropping and increased
farm profits, they had not displaced human labour (ODM, 1976). But
there is also contrary evidence on the effects of mechanisation
supporting Krishna’s more general finding that technical change in
agriculture tends to be labour-displacing. So, for example, the results
of a survey of cross-sectional studies similar to the one conducted by
Roy and Blase, mainly in India, indicated that large-scale mechanisa¬
tion is generally labour-displacing (Yudelman, et ai, 1971, ch.iv). A
162 AGRICULTURE AND ECONOMIC DEVELOPMENT

review of studies in East Africa concluded that although tractor

mechanisation is often profitable, especially on large farms, it is
usually labour-displacing. But, exceptionally, a labour supply
constraint may be binding on farm output and profit, especially on
farms in the vicinity of large towns where farmers must compete with
urban employers for labour: under such conditions tractorisation
may not be labour-displacing (Clayton, 1972).
In summary, Krishna’s study of farm labour displacement due to
agricultural technical change in the Punjab is outstanding for its
generality and analytical rigour. Its somewhat ‘pessimistic’ findings
therefore command close attention. Apart from this, most empirical
evidence on how farm technical change affects the employment of
labour relates specifically to mechanisation, particularly with trac¬
tors. The evidence on whether tractors displace labour is mixed. This
is not surprising since, as we have agreed a priori, the direct
substitution of machine power for the power of human labour is not
the sole motive farmers have for adopting machinery. Moreover,
mechanisation tends to be a gradual and piecemeal process taking a
long time to complete. At any given time the process is likely to be
more advanced in some places, be they individual farms, regions or
whole countries, than in others. Thus, the evidence provided by
contemporaneous but geographically separate studies may cover
many different stages of the mechanisation process. It would not be
surprising if the results of ‘earlier stage’ studies indicated little or no
labour displacement, whereas those of ‘later stage’ studies indicated
the opposite.
The hypothesis that the labour-displacing effect of farm mechanisa¬
tion in LDCs is progressive needs rigorous empirical verification.
However, positive verification would strengthen the case for govern¬
ment intervention to control the rate of mechanisation. In the final
section of this chapter we consider whether ‘selective mechanisation’
is a feasible policy instrument.

5.2 Indirect employment effects

Krishna describes a comparative static input-output model which he

used to measure the aggregate output and employment effects of
technological innovation in the agricultural sector. Readers are
reminded that relative to a given change in the final demand for the
product of a particular sector, input-output analysis measures not
only the consequential changes in that sector’s output but also the
additional output which is induced in other sectors. The size of the
 PEASANT AGRICULTURE 163

total induced demand is determined by the intensity of inter-industry

transactions and the size of the relevant demand multipliers.
Empirically, a 77 x 77 input-output table of the Indian economy
was condensed into a 2-sector table by grouping sectors into a
composite farm sector (FS) and a composite non-farm sector (NFS).
Appropriate assumptions were made concerning the values of the
parameters of the model’s exogenous variables. These included the
aggregate labour force growth rate, the agricultural growth rate,
technologically-induced rates of change in agricultural input co¬
efficients, and rates of change in marginal propensities to spend on
FS goods, and NFS goods. Alternative parameter values were used in
some cases (e.g. agricultural growth rate) to suit different study
objectives.
The model was used to project the effects of an assumed rate of
technical progress in the FS on aggregate (i.e. FS + NFS) output and
employment. Three basic variants of the model were used to separate
the effects of technical change per se from technical change with
growth. The variants were (i) technical change without growth in the
FS; (ii) 5 per cent per annum growth without technical change in the
FS; and (iii) 5 per cent per annum growth with technical change in
the FS. For each of these variants the projected rates of growth in
output and employment are shown in Table 6.1.
As expected, variant 1 actually displaces labour from the FS and,
despite positive employment growth in the NFS (induced by farm
technical change) aggregate employment also declines. An interesting

Table 6.1: Projected growth and employment effects of farm technological change in
India (Base year: 1964/65)

Variant FS NFS FS + NFS

X L X L X L

Growth rate (per cent per annum)

1 0.0 -4.0 1.0 1.0 0.7 -2.6

2 5.0 5.0 4.0 4.0 0.4 4.7
3 5.0 0.7 5.1 5.1 0.5 2.0

Note: X = sectoral output; L = sectoral employment.

Source: Adapted from Krishna (1975), Table 11.7.
164 AGRICULTURE AND ECONOMIC DEVELOPMENT

feature of variant 2 results is that, without farm technical change,

employment grows more slowly in the NFS than in the FS. As its
inherent assumptions are probably closer to reality, variant 3 results
are the most interesting of all. The combination of growth and
technical progress in the FS boosts employment and growth in the
NFS to the high level of more than 5 per cent. But technologically-
induced labour displacement reduced FS employment growth to less
than 1 per cent per annum. The growth rate of aggregate (FS + NFS)
employment is consequently reduced to only 2 per cent per annum. In
terms of absolute values, and the context in which we discuss it, the
results of this study are of no more than passing interest. Krishna
himself warns against attaching too much credence to projections
based on a particular set of assumptions. However, this reservation in
no way detracts from the value of the study in demonstrating the
difference between viewing the consequences of farm technical
change for factor employment in the FS alone, and the economy as a
whole. The choice of approach must obviously depend on the
investigator’s objectives. But it is quite clear that in many contexts the
partial or single-sector approach can be unduly myopic and therefore
very misleading.

6 Agricultural Resources and Technical Change in

LDCs: Policy Conclusions

Our policy conclusions fall principally under the headings ‘research

and information’ and ‘mechanisation’.

6.1 Research and information

Agricultural research is needed to hasten socially beneficial farm

technical change in LDCs. For institutional reasons, the initiation
and finance of agricultural research in LDCs is primarily the
responsibility of government. If government neglects this responsi¬
bility the likely consequence is either a research vacuum or a research
programme financed by the private sector and biased in favour of
private vested interests. In many LDCs agricultural progress is
currently retarded by an unfavourable labour:land ratio. This points
to a research programme emphasising biological innovations to
promote land-augmenting technical change. The search for mechan¬
ical innovations to save labour will often be better postponed until
 PEASANT AGRICULTURE 165

agricultural labour becomes scarcer at a later stage of development

(this point is elaborated in chapter 3 under labour surplus develop¬
ment models).
Substantial scope exists for international co-operation in the
promotion of agricultural research to benefit LDCs. The fruits of the
existing CIMMYT and IRRI research programmes, as well as those
of many less publicised bilateral arrangements between DCs and
LDCs (often under technical aid programmes), show the benefits of
such co-operation. A sensible division of labour is for DCs to
concentrate on solving more basic research problems with a broad
spread of possible applications. LDCs can then concentrate on
adopting what has been discovered in the DCs to suit their particular
natural environment and economic conditions.
Advancing agricultural research is not synonymous with agricul¬
tural technical change. The actual adoption of new technology in
agriculture depends on the spread of information, and the removal of
barriers to adoption. Responsibility for bridging the gap between the
research worker and the farmer rests primarily with the agricultural
extension service, particularly with respect to the dissemination of
information. The removal of barriers to adoption is, for reasons
discussed in chapter 2, a much harder nut to crack, but there is
frequently a case for local research specifically to identify the major
barriers to adoption.
Local farm trials and demonstration plots are important means of
disseminating knowledge of new discoveries to farmers, including
their net economic benefits. But the extension service function is not
confined to the dissemination of new knowledge it also includes
showing imperfectly informed farmers how to manage their existing
resources more efficiently to yield a larger profit or greater utility.
Economic efficiency combines allocative with technical efficiency. As
far as peasant agriculture is concerned, the evidence that farmers are
allocatively efficient is more convincing than the evidence that they
are technically efficient. Thus, with better knowledge of existing farm
practices, many farmers may be able to move on to a higher
production function without waiting for new discoveries.
In chapter 2 we referred to the argument that, because of their
different resource endowments and technical skills, western-style
technology is ‘inappropriate’ for adoption by the LDCs. This
argument has the obvious policy implication that part of the LDC
agricultural research effort must be directed to the development of
more appropriate technologies to accelerate the economic betterment
of peasant farmers. A central aim of such research must be to raise
166 AGRICULTURE AND ECONOMIC DEVELOPMENT

agricultural labour productivity at a low capital cost and without

undue labour displacement.

6.2 Mechanisation

Here we consider whether it is desirable and feasible for LDCs to

pursue ‘selective’ mechanisation as a policy objective. We also
consider how the private costs and social consequences of farm
mechanisation in LDCs may be affected by non-specific government
policies.
Selective mechanisation means ‘restricting mechanisation to where
it contributes to increasing employment or is necessary to break a
seasonal bottleneck’ (FAO, 1973, ch. 3). This is a very general
definition and little attempt has been made to make it sufficiently
specific for use as a working principle or rule. We do not know of
anywhere where selective mechanisation has been adopted as a
specific policy goal, although a few examples of its ‘feasibility’ have
appeared in the literature. But, before considering the feasibility of
selective mechanisation, we first discuss the conditions of its ‘desira¬
bility’ as a policy objective. There are two aspects of this. First, there
is the familiar trade-off between short-term gains and long-run losses
(or vice versa). Whereas the long-term social interest requires capital
to be substitued for labour in order to raise agricultural labour
productivity and farm incomes, the short-term interest requires that
jobs be preserved and living standards maintained for the many. In
practice, there has to be a compromise between the pursuit of short-
and long-term policy goals. Selective mechanisation is biased to¬
wards the short-term, and its desirability as a policy goal has to be
judged within a broader policy framework reflecting government’s
views on the relative importance of long- and short-run policy
objectives.
The second reason for doubting the desirability of selective
mechanisation is that the favourable concealed or indirect effects of
farm mechanisation on employment may out-weigh the direct effects.
We cited earlier Krishna’s projection based on Indian data, in which
the additional work-time created outside agriculture by farm techni¬
cal change exceeded the worker time lost in agriculture itself. Thus,
taking a wider view than that of employment in agriculture per se,
selective mechanisation could actually be inimical to the maximisa¬
tion of aggregate employment, even in the short term.
It might be argued that the indirect employment benefits of farm
technical change can be discounted if the relevant inputs are imported,
 PEASANT AGRICULTURE 167

so that the extra non-farm employment generated by their purchase

leaks abroad. However, this argument is open to the criticism that,
because of trade reciprocity, it may be necessary to admit more
imports of machinery (as well as imports of other goods) in order to
expand exports. Thus, indirectly, curtailing machinery imports may
also limit employment in export industries.
Taken together, these two aspects point strongly to the conclusion
that, even if selective mechanisation’s objective of preventing dis¬
placement of farm labour is feasible, it is not necessarily a desirable
policy goal.
Turning from the desirability to the feasibility of selective mech¬
anisation, most published examples have merely stipulated the
technical conditions of feasibility, rather than demonstrating that it is
economically feasible in practice (e.g. Yudelman, et al., 1971, ch. iv).
In one example, involving the mechanisation of rice production in
Thailand, the displacement of labour could be avoided only if the rice
was transplanted (rather than the seed being broadcast), the use of a
tractor was restricted to the first of two ploughings, and the rice
acreage was expanded to absorb the consequent ‘saving’ of labour. In
another example, the feasibility of double cropping wheat with
cotton in Pakistan depended on introducing a stationary threshing
machine (to replace threshing by hand) to break a labour bottleneck
created by an overlap of the wheat harvesting and cotton planting
seasons. If mechanisation had been extended to non-threshing
operations labour would have been displaced; but this could be the
more profitable option for the farmer.
These two examples illustrate the very restrictive conditions
governing the successful attainment of the objectives of selective
mechanisation. It is also apparent that selective mechanisation
implies a substantial degree of external control over the types of
machinery that are permitted to be used on farms, and even more
important, the operations that the machines admitted are allowed to
be used for. Even if it were possible to identify systems of selective
mechanisation which, if widely adopted by farmers, would be socially
beneficial in a particular place at a particular time, the administration
of such a scheme would scarcely be feasible under private machine
ownership. Even if private machine ownership is banned in favour of
a government monopoly, with farmers being restricted to hiring
machine services from the government, the administrative burden of
control, to ensure that machines are restricted to authorised uses, is
still considerable. There is also the danger of monopoly breeding
inefficiency due to the absence of competition. Even in LDCs where
168 AGRICULTURE AND ECONOMIC DEVELOPMENT

private agricultural contractors have been allowed to compete with

machinery services supplied by the government, the latter have often
operated at a substantial financial loss. Thus, even if the objective of
selective mechanisation — to prevent or retard the displacement of
labour from agriculture — is desirable in itself, there are serious
doubts concerning its feasibility, particularly in market-type eco¬
nomies based on decentralised decision-making and producers
freedom of choice. Binswanger and Ruttam (1978), chapter 10,
contains a good discussion of farm mechanisation policy, including
the feasibility of selective mechanisation, with specific reference to
Brazil.
Despite the question mark against the feasibility of selective
mechanisation, there is no doubt that government can influence the
rate of mechanisation unselectively, by controlling the distribution of
agricultural machinery and the terms on which it is acquired by
farmers. In many LDCs virtually all modern agricultural machinery
is imported due to the lack of domestic manufacturing capacity. The
variety of instruments government can use to regulate imports is
considerable. Imports can be reduced or curtailed by means of import
duties and quotas. Alternatively they can be encouraged by removing
duties or increasing quotas, or by granting positive import subsidies.
Inducements to import may or may not be specific to a particular
category of import, depending on whether government operates an
import licensing system. The currencies of LDCs frequently trade at a
discount in the open market. Ceteris paribus, over-valuation in¬
creases the volume of imports (including machinery imports) because
of its effect on the terms of trade. A dual exchange rate system may
operate to regulate the balance of payments. Importers who are
licensed to import at the market rate of exchange, rather than at the
higher ‘official’ rate, enjoy a concealed import subsidy. Countries
pursuing a policy of economic self-sufficiency (or autarky) erect high
barriers against virtually all imports. But, apart from these, govern¬
ments of LDCs have probably tended to encourage rather than
discourage imports of agricultural machinery. In some cases the
encouragement is explicit. Import licences are liberally granted to
would-be importers, with or without an import subsidy, in pursuance
of a policy- of speeding the ‘modernisation’ of agriculture. Many
countries have also admitted substantial quantities of agricultural
machinery from abroad under overseas aid programmes. But
elsewhere the encouragement has only been implicit - or even
unintentional - notably in the case of countries with chronically
over-valued currencies.
 PEASANT AGRICULTURE 169

Regardless of whether agricultural machinery is imported or

domestically produced, there are numerous ways in which govern¬
ment may indirectly encourage the substitution of machinery for
human labour. For example, advancing credit to farmers at a
subsidised interest rate stimulates all forms of investment, including
machinery. Similarly, a fuel subsidy encourages the acquisition of
machines such as tractors and self-propelled combine harvesters. The
encouragement to mechanise is not necessarily intentional. It may
well be the unintentional effect of policies designed to achieve other
objectives. So, for example, subsidised credit may be intended to
encourage farmers to purchase non-capital inputs, such as fertilisers
and pesticides; the stimulus to machinery investment is unforeseen.
This review of farm mechanisation policy in LDCs points to four
major conclusions. First, although the substitution of capital for
labour is inevitable in the long run, the optimum rate of mechanisa¬
tion is for public decision. Secondly, government’s ability to control
the rate of farm labour displacement, by pursuing a policy of selective
mechanisation, is strictly limited. Thirdly, government can choose to
encourage or discourage mechanisation unselectively, by applying
specific taxes or subsidies. Fourthly, because numerous non-specific
government policies, such as the exchange rate, influence the
purchase of farm machinery, such policies need frequent review and
possible revision to avoid unintentional effects on the rate of
mechanisation.

References

Bachman, K.L. and Christensen, R.P. (1967), ‘The Economics of

Farm Size’, in Southworth, H.M. and Johnston, B.F. (eds). Agricul¬
tural Development and Economic Growth, Cornell University Press.
Ithaca.
Berry, D.A. and Cline, W.R. (1979), Agrarian Structure and Pro¬
ductivity in Developing Countries, Johns Hopkins University Press,
Baltimore.
Binswanger, H.P. and Ruttan, V.W. (1978), Induced Innovation,
Technology, Institutions and Development, Johns Hopkins University
Press, Baltimore.
Clayton, E.S. (1972), ‘Mechanisation and employment in East
African agriculture’, Int. Labour Rev., 105(4).
Donaldson, G.F. and Mclnerney, J.P. (1973), ‘Changing machinery,
170 AGRICULTURE AND ECONOMIC DEVELOPMENT

technology and agricultural adjustment’, American J. Agric.

Economics, 55(5).
Eckart, J.B. (1977), ‘Farmer response to high-yielding wheat in
Pakistan’s Punjab’, in Stevens, R.D..(ed.) Tradition and Dynamics in
Small-Farm Agriculture, Iowa State Univ. Press, Ames, Iowa.
FAO (1973), The State of Food & Agriculture, Food and Agriculture
Organisation of the United Nations, Rome.
Farrell, M.J. (1957), ‘The measurement of productive efficiency’, J.
Royal Stat. Soc., Series A, vol. 120, Part pp. 254-6.
Feder, G., Just, R. and Silberman, D. (1981), Adoption of Agricul¬
tural Innovations in Developing Countries: A Survey, World Bank
Staff Working Paper no. 444, The World Bank, Washington DC.
Hayami, Y. and Herdt, R.W. (1977), ‘Market price effects of
technological change on income distribution in semi-subsistence
agriculture, American J. Agric. Econ. 59(2).
Hayami, Y. and Ruttan, V.W. (1971), Agricultural Development: An
International Perspective, Johns Hopkins University Press,
Baltimore.
Hopper, W.D. (1965), ‘Allocation efficiency in a traditional Indian
Agriculture’, J. Farm Economics, 47(3).
Jones, William O. (1977-8), ‘Turnips, the Seventh Day Adventist
principle and management bias’, Food Res. Inst. Studies, XVI(3).
Junankar, P.N. (1980), ‘Do Indian farmers maximise profits?’
J. Dev. Studies 17(1).
Krishna, R. (1975), ‘Measurement of the direct and indirect employ¬
ment effects of agricultural growth with technical change’, in
Reynolds, L.G. (ed.), Agriculture in Development Theory, Yale
University Press, New Haven.
Lipton, M. (1968), ‘The theory'of the optimising peasant’, J. Devel.
Studies, 4(3).
Lipton, M. (1978), ‘Inter-farm, inter-regional and farm-non-farm
income distribution: the impact of the new cereal varieties’, World
Development, 6(3).
Newbery, D.M.G. and Stiglitz, J.E. (1981), The Theory of Commodity
Price Stabilization, Oxford University Press.
ODM (1976), British Aid Tractors in India: An Ex-Post Evaluation,
Ministry of Overseas Development, London.
Rask, N. (1977), ‘Factors limiting change on traditional small farms
in southern Brazil, in Stevens, R.D (ed.) Tradition and Dynamics in
Small-Farm Agriculture, Iowa State Univ. Press, Ames, Iowa.
Roumasset, J. (1976), Rice and Risk, North-Holland, Amsterdam.
Roy, S. and Blase, M.G. (1978), ‘Farm tractorisation, productivity
and labour employment: a case study of Indian Punjab’, J. Devel
Studies, 14(2).
 PEASANT AGRICULTURE 171

Ruttan, V.W. (1974), ‘Indeed technical and institutional change and

the future of agriculture’ in Hunt, K.E. (ed.). The Future of
Agriculture, Institute of Agricultural Economics, Oxford.
Schulter, M. and Mount, T. (1974), Management Objectives of the
Peasant Farmer: An Analysis of Risk Aversion in the Choice of
Cropping Pattern, Surat District, India, Occasional Paper no. 78,
Cornell University, Department of Agricultural Economics.
Schultz, T.W. (1964), Transforming Traditional Agriculture, Yale
U.P., New Haven.
Shapiro, K. (1977), ‘Efficiency differentials in peasant agriculture and
their implications for development policies’, in Contributed Papers
Read at the 16th International Conference of Agricultural Economists,
Institute of Agricultural Economics, Oxford.
Wolgin, J.M. (1975), ‘Resource allocation and risk: a case study of
smallholder agriculture in Kenya’, Amer. J. Agric. Econ., 54 (4).
World Bank (1978), World Development Report, World Bank,
Washington DC.
World Bank (1981), Accelerated Development in Sub-Saharan
Africa, World Bank, Washington DC.
Yotopoulos, P.A. and Nugent, J.B. (1976), Economics of Develop¬
ment, Harper & Row, New York.
Yudelman, M., Butler, G. and Banerji, R. (1971), Technological
Change in Agriculture and Employment in Developing Countries,
Organisation of Economic Co-operation and Development, Paris.

Notes
1. Specifically, the inputs were land area, bullock time (hours), human
labour (hours), and irrigation water (volume): the production alternatives
were barley, wheat, peas and gram, all measured by weight.
2. Technically, the individual producer will choose the point on the
production frontier which is tangent to his indifference curve (see
Newbery and Stiglitz, 1981, pp. 170-1).
3. Factor productivity has physical attributes affected by resource endow¬
ments. A poor farmer with very little land and an excess of labour cannot
be efficient in terms of total factor productivity without access to
additional land. Berry and Cline (1979) overlook this aspect of the
relationship between factor productivity and farm size.
7 Supply Response

1 Introduction

In economics ‘supply response’ in underdeveloped agriculture gener¬

ally means the variation of agricultural output and acreage mainly
due to a variation in price. Let Qa be the volume of agricultural
output and P indicate the price level, Wt weather condition (e.g.
rainfall), A acreage, and t any time period. A typical simple supply
response function can then be written as follows:

Qa = f(Pt-o A„ Wt, Ut) ••• (1)

where Pt_, is really a proxy for the expected price and Ut a statistical
error term. Such a supply response functon indicates that the volume
of agricultural output will depend upon past price level of such
output, the acreage under cultivation at present and the level of
rainfall in the current year plus a catch-all variable which is not
specifically known, i.e. Yt - the yield variation - is implicit in Ut. The
exact nature of such a function can vary due to difference in the nature
of the economies. But, at the outset, the reason behind formulating
such a function is quite clear. It is to be expected that:

di di di
>0; >0; >0
dP t-i 8A, 3Wt <
This means that output is expected to vary positively with past price
and land under cultivation but it could either rise or fall with changes
in rainfall depending upon whether or not we have a normal rainfall
or flood and drought.
Sometimes, acreage under cultivation is taken as a proxy for
output due to a very close and direct relationship between acreage and
output, i.e.

Qat = ai + a2At a2 > 0 ••• (2)

172
 SUPPLY RESPONSE 173

Hence in many cases, supply response has been assumed to be

equivalent to response of acreage under cultivation to changes in
economic and non-economic factors. Thus, we have:

At = f(Pt-ls Wt, ••• ,Ut) ••• (3)

Sometimes, equation (3) has been modified to take into account the
impact of relative rather than absolute prices. Assume that Pat stands

Prices

Acreoge

Output

Figure 7.1: Output and acreage response to price

 174 AGRICULTURE AND ECONOMIC DEVELOPMENT

for the price of alternative crop that can be raised from the same plot
of land, whereas Pot is the own price of the crop. Hence, we obtain

At = f(Pot-1/Pat_1,Wt,...,Ut)- (4)
The mechanism of supply response can be illustrated with the help of
a simple figure. In Figure 7.1, we measure acreage (=A) on the
horizontal axis and price (relative or absolute) on the vertical axis.
The curve OA represents the supply response of acreage to changes in
the level of prices. Such a curve is clearly upward sloping. In the
bottom section of the figure we show the relationship between output
and acreage. As it has been already suggested, such a relationship is
positive and this has been approximated by the curve OC. It can be
easily shown from Figure 7.1 that when the price is OPl5 the amount
of acreage under cultivation is OAt and output is OQt. If the price
level rises to OP2, acreage under cultivation rises to OA2 and output
increases to OQ2.
However, it is possible to postulate a direct relationship between
price of agricultural output and its supply. Here we can write:

Q.,=f(p.t-i) •• (5)

(6)
Such a supply response model is generally known as the Cobweb
model. Note that Pa = own Price and P0 = Price of other crops.

2 The Cobweb Model: An Illustration

The Cobweb model is based on the theory of lagged adjustment.

Given the annual cycle of agricultural crops, markets in agriculture
operate period by period and such a phenomenon is also known as a
‘hog cycle’ in the USA because production there is cyclical. If this is
true, then the continuous adjustment principle should be replaced by
a discrete one in the following way:

Pt ~ Pt-1 = f[Dp(Pt_,) — S(Pt_,)] • • • (7)

If the value of f were found to be very high, then the price amplitude
of fluctuations, after an initial disturbance, could be ever-increasing
around the equilibrium price and this could render the model
‘explosive’ (i.e. once the equilibrium price has been disturbed, there
will be a tendency to move away from the point of equilibrium). Let
 SUPPLY RESPONSE 175

us define an equilibrium as the point where demand (= Dt) is equal to

supply (= St). Hence we have:

(8)
St is not instantaneous supply curve but lagged output function. Let
the demand function be written as:

(9)
This simply means that demand is inversely related to price, and as
price goes up, demand goes down (and vice versa). The form of the
demand curve is assumed to be linear to facilitate exposition.
The supply curve can be written as a function of the price in the last
period (Pt_,) since it is assumed that current supply decisions are
taken solely on the basis of past prices. Thus we have on the basis of
our rather naive assumption:

St — b0 + t>2Pt—[ ••• (10)

Such a model is ‘recursive’ as it is argued that current year’s supply is
given by previous year’s price and such a supply determines current
price, given the current demand and market-clearing conditions.
Once we obtain current year’s price, we can determine supply in the
next year. Such a situation is shown in Figure 7.2.

Price
(= P)

S,

P.

P*

Pi

0 0, 0, 0, Quantity demanded
ond supplied

Figure 7.2: A ‘convergent' Cobweb model

176 AGRICULTURE AND ECONOMIC DEVELOPMENT

It is obvious from the diagram that the equilibrium price will be

given at Pe price and output will be OQ0. Let us assume that a drought
leads to a crop failure and we have OQx amount of output. Given an
excess demand the price level rises to OP^ Such an increase in price
level encourages farmers to produce OQ2. But then we have an excess
supply, price level falls to OP2 which, in its turn, leads to a reduction
of supply to OQ3. At this output we have an excess demand situation
which results in a rise in price level. The oscillation process ends when
the equilibrium price and output are obtained at OPe and OQ0.
It is necessary to point out that both price and output can
overshoot (above the equilibrium values) or undershoot (below the
equilibrium values) due to the inherent mechanism of a Cobweb
theorem. Notice further that in Figure 7.2, the amplitude of fluctu¬
ations in price and output gradually falls and eventually convergence
(i.e. movement towards the equilibrium) occurs at the equilibrium.
This is due to the elasticities of the demand and supply curves. There
is no reason to assume that a Cobweb cycle will always converge.
Indeed, it can diverge, in which case there will be a movement away
from the equilibrium values. This is demonstrated in Figure 7.3
simply by changing the slopes of the D and S curves. Readers can
check for themselves that a Cobweb cycle may also be continuous,
which implies no movement towards equilibrium or away from the
initial cycle (see Figure. 7.4).
 SUPPLY RESPONSE 177

If the Cobweb movements are divergent, then the market is

unstable. If it is convergent, then the market is stable.

2.1 Stability of the equilibrium in a Cobweb cycle

A close look at Figures 7.2 and 7.3 suggests why some markets for
agricultural products are stable whereas others are unstable. The
degree of stability in Figure 7.3 depends critically upon the elasticity
of the supply curve. The more elastic the supply curve, the less stable
the market. In other words, if the response of the suppliers to a small
change in the expected price is drastic and large, then the market will
tend to be unstable. Check also that the greater the inelasticity of
demand, the more a sure change in supply will alter prices. As a
general rule, it is possible to formulate the following proposition:

2.1.1 The Proposition

If the absolute elasticity of supply is greater than the absolute
elasticity of demand, then the equilibrium is unstable. This is easily
verified by drawing diagrams.
 178 AGRICULTURE AND ECONOMIC DEVELOPMENT

An algebraic proof. Let the demand and supply curves be linear

and let demand be inversely related to the current price whereas
supply be given by past period’s price. Hence, we have equations
(9) and (10):

Dt = a0-a1Pt • (9)

St — b0 — b1Pt_1 ••• (10)

Thus, from the market-clearing equation (8):

a0-a1Pt = b0Tb1Pt_I... (11)

Equation (11) is useful to evaluate the equilibrium price (= Pe). Then
if the initial price (P0) 4= Pe, we want to know the time trajectory
of price that eventually goes back to equilibrium price.
In the equilibrium, we have:

Pe = Pt = Pt_, •• (12)
We can now rewrite equation (11) as:

a0 — atPe = b0 + biP6 ••• (13)

The solution for Pe is then given by:

a0-b0 = b1Pe + a1Pe- (14)

or
ao ~ b0 = (ai -I- bJP6 ••• (15)
or

pe _ ao ~ b
(16)
aj + b

To find out the line trajectory of prices, beginning with P = P0, we

must find a solution for the first-order difference equation given by
equation (11). Here, we consider price in t period as a function of
price in the initial period, P0. This yields equation (17) for Pt:

= Pe + (18)

In Figure 7.2, given the slopes of the supply and demand curves (e.g.
bi is positive as the supply curve slopes up and so on) and P0 > Pe, the
 SUPPLY RESPONSE 179

change in the market in the first period will be given as follows:

(19)

(20)

to 0 as t approaches infinity. This simply means that the market will

be stable only if the elasticity of demand is greater than the elasticity
of supply. Fortunately, and perhaps a bit surprisingly, markets for
agricultural goods are usually stable (see any good textbook, e.g.
Layard and Watters, 1978).

3 ‘Perverse Supply Response’ in Backward Agriculture

In the previous sections we have discussed the problem of supply

response in underdeveloped agriculture under the implicit assump¬
tion that farmers in LDCs behave in a normal way. This means that
the higher the price level, the greater will be the level of output per
acre. In other words, the slope of the supply curve is positive.
Under such a condition, the elasticity of supply with respect to
price change is greater than zero. Suppose P denotes the price level
and S denotes supply and 5P and <9S denote change in price and
change in supply respectively. Therefore, the elasticity of supply r\s is:

dS/S dS P dS P
ns~dP/P~~S dP ~dPS (22^

The normal supply response occurs when rjs > 0.

However, two other behavioural postulates are possible. These are
that the response to a price increase is either zero or negative. In the
first case, supply of output may not respond to price changes at all.
For instance, where farmers produce only for subsistence and not for
the market, positive price changes may not induce them to produce
more. In such cases, the supply elasticity will be equal to zero, i.e.
ris = 0. On the other hand, a price rise may actually induce the farmers
180 AGRICULTURE AND ECONOMIC DEVELOPMENT

to supply less output. This is generally known as a ‘perverse supply’

response. Here the supply curve slopes downwards whereas in the
‘normal’ case the curve will slope upwards. In such a case, the
elasticity of supply with respect to price will be negative, i.e. < 0.
Although perverse supply responses appears to be economically
irrational it has merited some attention in the literature (see Mathur
and Ezekiel, 1961; Khatkhate, 1962).
The case of the ‘perverse supply’ response crucially depend upon
the assumption of a unit elasticity of the demand for money income
so that when price falls by a certain proportion, farmers raise the
marketed output by the same proportion to retain the same amount
of money income. Hence, as price falls, marketed output rises, market¬
ed output being defined as the amount of production net of con¬
sumption by the farmers. Others have obtained the same result by
assuming that the elasticity of the money demand function of the
subsistence farmers is less than unity (see Mathur and Ezekiel, 1961).
Demand for money of the farmers is generally given by the
requirements to make payments in cash (e.g. taxes, rent, demand for
non-agricultural, say, industrial goods, etc). Although it is obvious
that such assumptions will generate a perverse supply response curve
(i.e. as price falls, supply rises), it is doubtful whether these
assumptions are realistic enough to produce a perverse behaviour in
practice. For one thing, the assumption that a fall in price of
agricultural goods wll have no substitution effect is not very realistic.
The substitution effect usually increases the demand for a product
due to a fall in its relative price. For another, it is assumed that the
income elasticity of the demand for non-agricultural goods is zero
in these models as farmers actually reduce the sale of crops when
price rises due to a fixed demand for money. This could only happen
when farmers do not have any extra need for non-agricultural goods
given an additional increase in their money income. Once again,
such a situation seems implausible.
A perverse supply response may also be observed due to the impact
of the surplus foreign food disposal programme. Under the U.S.
Public Law 480, food export from the U.S. A. at times of acute food
scarcity in LDCs has no doubt prevented a mass starvation and
death. However, some have argued that when foreign supply
competes directly with domestic production the market price at home
will be depressed. It has been pointed out that although India
succeeded in avoiding mass starvation during the severe drought-
years of 1965-67 and subsequent crop-failures, nevertheless, the
domestic supply of foodgrains failed to increase as prices of
 SUPPLY RESPONSE 181

foodgrains were kept at an artificially low level due to the American

export policy. Notice that had America been exporting to another
poor country to prevent acute scarcity and famine, indirect compe¬
tition (via arbitrage) with foreign supplies would have hindered prices
to attain their ‘normal’ equilibrium values. The actual shortfall of
domestic production will depend upon the values of elasticities.
When foreign food is imported in the Indian market, then with a
given Indian food supply, the degree of a fall in domestic price will be
given by the value of the demand elasticity. As we have seen in the
Cobweb model, the increase in supply in the next period will be
determined by the value of the elasticity of supply with respect to
prices. If we assume that the supply elasticity is normal i.e. positive,
the import of food in a LDC will reduce increases in domestic supply
(Schultz, 1960). Indeed, the higher the values of elasticity, the greater
will be the reduction in supply. However, if we assume, along with
Dantwala, that farmers do not respond at all to changes in
agricultural prices, then the analysis of the effects of imports of
foodgrains on domestic production is not applicable (Dantwala
1963).
Food imports from abroad can help to set up buffer-stocks in
developing countries. Such buffer-stocks are supposed to iron out
sharp fluctuations in prices. If it is assumed that producers of
agricultural goods are usually risk averse, then reduction of price
fluctuations via imports may actually raise the production of
agricultural goods at home. This will be the case even when the
buffer-stock price is lower than the average of the oscillating price
that would prevail in the absence of stabilization programmes
(Fisher, 1963; see also chapter 10 of this volume).

4 A Simple Supply Response Model

It is possible to develop a simple supply response model to analyse the

impact of price on output. Such a study usually involves the
specification of output equation with reference to only two variables
in the first instance. Other variables may be introduced at a later stage
of the analysis (see e.g. Yotopoulos and Nugent, 1976). Let us assume
that total output (= Q of any crop) is given by the product of acreage
(= A) and yield (= Y). Thus we have

Q = AY ••• (23)

Let us also assume that both acreage under cultivation and yield
182 AGRICULTURE AND ECONOMIC DEVELOPMENT

respond to price (— P) changes. Totally differentiating, we obtain,

8Q d\ dY
Y (24)
dP dP + AdP

Assuming constant returns to scale (i.e. an increase in inputs will raise

output by the same proportion) and dividing the above equation by
Q/P, we get,

dQ/dP Q/A-dA/dP Q/Y-dY/dP

(25)
Q/P Q/P + Q/P

SA/8P dY/dP
(26)
/qp A/P + Y/P

<,P = 4P + ^yp •" (27)

where, /ap = the elasticity of acreage with respect to price, t = the

elasticity of yield with respect to price, and <fqp = the elasticity of
output with respect to price.
It is generally assumed that /yp is non-negative. To obtain a lower
bound estimate of ^qp, it is only necessary' to regress acreage on price.
In empirical studies some have assumed that the price elasticity of
yield is negligible (e.g. Falcon, 1964; Behrman, 1968). Others have
emphasised the possibility of important yield responses to prices in
countries like the USA where a substantial increase in yields have
substituted for a limit on acreage (Nerlove, 1958).
The idea of regressing acreage rather than output on price is not
very difficult to understand. Acreage is really a proxy variable since
planned output (the variable- to be explained) cannot be directly
observed. It is, however, possible to use realised output as a proxy for
planned output. But realised output could differ significantly from
planned output due to the impact of random factors (e.g. weather) on
agricultural output. However, acreage is generally under the direct
control of the cultivator and hence it is usually accepted as the proxy
variable for output.
As regards yield response and technical change, it should be
mentioned that the substitution of yield for crop area in the USA has
been substantially due to adoption of higher yielding crop varieties
(as well as to heavier fertiliser application) and the same could be said
of some LDCs during 1960s and 1970s due to the Green Revolution.
But, in the short run, without technical change, yield response is
likely to be confined to variation in the intensity of harvesting, etc.
 SUPPLY RESPONSE 183

For example, if the price at harvest is very low (i.e. below the marginal
cost of harvesting) farmers may not bother to harvest at all. But such
varieties in short-term yield response may not be picked up empiri¬
cally by using only annual data.

4.1 Role of expected price in supply response

It is important to bear in mind that acreage planted could vary

with expected price of the output. Hence it may be necessary to
develop a model which includes expected price as one of the
explanatory variables. The price expectation models have been
developed by Cagan (1956) and Nerlove (1958), inter alia, in the
economic literature. Following Cagan, elasticity of expectation (<f )
can be defined as the percentage change in expected future price
divided by percentage change in present prices.

pt+l- p, pt-pt-,
(28)
pt pt-.

Hence a unit elasticity of expectation will imply that the proportion of

expected price change to present price change will be the same.
However, in Cagan’s model, expectations are adaptive, which means
people change their expectations in proportion to the error related
with previous levels of expectation. Expected price (P*) will thus be
the sum of past expected price (= P* ,) plus a proportion (= 0) of the
difference between actual past price (= Pt_,) and the past expected
price. Hence, we can write:

pt* = P.*-, + ^[P.-, - P,*,]

(29)
O<0^1...
where 0 = coefficient of price expectation (associated with price
uncertainty). It can be shown that:

(30)
t=o

Since:

P*=(l-/0P*_I+/?Pl_1... (31)
= /?pt-. + (1 - MO - 0P,*-2 + J••• (32)
and so on.
The above equation shows that anticipated prices form a geometri¬
cally falling lag structure as a function of all past prices (but cannot be
184 AGRICULTURE AND ECONOMIC DEVELOPMENT

estimated empirically because we are unable to observe P*_pPt*-2*

etc). .
When output is a function of anticipated prices, we have:
Qt = a + bPt*+ut... (33)
It is clear from the last three equations that we have:
Q, = a + bSfll - jSyPt-.-i + ut... (34)
In the Cobweb model Equation (34) reduces to Qt = a + bPt_,.
To estimate the above equation, it is necessary to follow an
interactive procedure to iterate P in order to obtain the highest
proportion of explained variation (by the independent variables) to
total variation.
In Nerlove’s model, anticipated prices determine equilibrium
output (=Q*) and, in each production period, output is partially
altered in proportion to the difference between last periods actual
output and the long-run equilibrium output. Hence, we have the
following basic equations:
Qt* = a + bPt*... (35)
Pt*=/?Pt-i +(1 —/?)P*-i • • • (36)
Qt = yQ,*+(i-y)Qt-.--- (37)
where y = the rate of adjustment associated with technical and
institutional rigidity.
Manipulation to eliminate non-observable variables gives:
Qt = 0ya + /JybPt_, + [(1 - P) + (1 - y)]Q,_,
-(i-ffl(i-r)Q,_2... (38)
Equation (38) can be rewritten as:
Qt = d + ePt_, +fQt-t — gQt-2" (39)
where e and e/1 -f are the short and long-run supply elasticity
parameters.
(Note: Since P and y cannot be separately identified, Equation (39)
can be solved empirically only by assuming either that p = 1 or that
y = 1 (which is obviously restrictive)). This assumption enables the lag 2
taken in Qt to be ignored.
The Nerlove’s supply response model is obviously more general
than the simple supply models which we have discussed at the outset
(see equation (10) above):
Qt — b2 + b3Pt_,... (10')
 SUPPLY RESPONSE 185

A close look at equations (10) and (10') reveal that the dependent
variable (i.e. the variable to be explained) in (10') is Qt rather than St.
The reason for such a change lies in the distinction between supply
response in general and marketed supply or ‘surplus’ response in
particular. This distinction is important and deserves special con¬
sideration. But before we discuss the problem of marketed surplus, it
is necessary to clarify some issues related to the supply response in the
agricultural labour market in LDCs.

5 Supply Response in the Underdeveloped

Agricultural Labour Market

The term supply response in the agricultural labour market in the

LDCs implies changes in the supply of labour in the agricultural
sector in LDCs with respect to changes in the level of wages or
income. It is sometimes assumed that the supply curve of labour in
underdeveloped agriculture may be backward bending. The point
can be further illustrated with the aid of figures. Consider a ‘normal’
supply curve of labour which is expected to rise upwards when
wages/income rise (see Figure 7.5 and the curve S').
186 AGRICULTURE AND ECONOMIC DEVELOPMENT

A ‘backward’ bending supply curve is shown as OS". Notice that up

to the point E, the supply curve is ‘normal’ whereas after that point
the curve bends backwards as less labour hours will be forthcoming
despite a rise in the level of wages. Such a phenomenon could occur
due to high leisure preferences when income rises. The perverse or
backward bending supply curve of labour can also be illustrated with
the help of a figure in which income/wages are measured on the
vertical axis, and leisure (rather than labour) is measured in the
horizontal axis. In Figure 7.6 the maximum amount of leisure is
shown by the line OjLj. It is possible to measure labour hours if we
move from right to left with as the point of origin. Thus, with Ox
as the point of start, a movement from right to left will indicate more
labour and less leisure. Assume that there is a family of indifference
curves between leisure and income/wages, and these are shown by
curves like i0, il5 i2, i3,... in figure 7.6(a). Farmers have the choice not
to do anything and enjoy leisure, in which case they consume of
leisure but do not earn much income. If they decide to work all the
time they can initially get OWj level of wages. In other words, the line
OjWj is the ‘budget’ line of a farmer who has the opportunity to
trade his leisure against a maximum income of OW,. The
leisure-income choice is given by the indifference curve which shows
various combinations of leisure and income to which farmers are
indifferent. Given the principles of indifference curve analysis, the
initial equilibrium income and leisure are given by OWj income and
OLx leisure at point E on the indifference curve i0 in Figure 7.6(a). In
Figure 7.6(b), such an equilibrium combination is given by the level
of wage at OWj with OL^ amount of labour supplied by a farmer.
Let us now assume that the farmer faces a new budget line as the
level of income goes up from O^Wi to OtW2. The new equilibrium is,
once again, attained where the budget line is tangent to the highest
indifference curve i.e. Consumption of leisure falls from OLi to
OL2 and wages rise from OWj to OWj. In Figure 7.6(b) this
movement is indicated by an increased supply of labour from 0^ to
OjL2. In Figure 7.6(a) and (b) it is shown that both wages and labour
supply will rise in the next point of equilibrium as the budget line
shifts upwards. However, if wages rise above OW'3, the supply curve
‘bends’ backwards. It is evident that the backward bending supply
curve of labour is determined by two forces: income effect and
substitution effect. We know that the substitution effect is always
negative. An increase in the price of any ‘good’ will always lead to a
fall in the demand for these goods. But an income effect can be
negative or positive. An increase in income will normally raise the
 SUPPLY RESPONSE

Income / Wages
188 AGRICULTURE AND ECONOMIC DEVELOPMENT

demand for all goods (including the demand for leisure) as the
individual feels better-off. In this case, this may imply ‘consumption’
of more leisure. On the other hand, if leisure is an inferior good, then a
fall in the relative price of leisure may actually lead to a fall in the
consumption of leisure (consumption of more labour) despite the fact
that a fall in the relative price of leisure raises the real income of the
individual. Notwithstanding such a rise in real income, the individual
decides to ‘consume’ less leisure. In the first part of the labour supply
curve that we have traced out (i.e. OtR) clearly the substitution effect
dominates the income effect. As we move from to OtW3, the
relative price of labour rises and hence farmers sell more labour (or
consume less leisure), hence consumption of leisure falls.
Once the level of income has reached OW'3, the supply curve of
labour bends to the right. This is due to a strong ‘normal’ income
effect which dominates a rather weak substitution effect. At E2,
farmers are already selling a lot of labour and a rise in its price will
indicate a strong rise in income. On the other hand, since the relative
rise in price has decreased, the impact of substitution effect has also
relatively weakened. The net result is an increased consumption of
leisure. The overall result of these two types of reactions are well
summarised in Figure 7.6(b) which shows that an increase in labour
is followed by its reduction as income rises.
It is now easy to summarise the basic arguments developed in this
section. A rise in income increases the price of leisure and hence a
simple substitution effect will induce farmers to offer more labour and
consume less leisure. But an increase in income and the price of leisure
also implies a fall in real income of the farmer because he also
‘consumes’ leisure. This is the income effect which will reduce the
consumption of leisure. But as the farmer’s income rises substan¬
tially, his demand for leisure also rises with a rise in his demand for
everything else. At a fairly high level of income, the supply curve of
labour bends backwards as leisure preferences rise substantially.
It is useful to remember that the above analysis rests on an
important implicit assumption in economics - that as income rises,
additional (marginal) utility due to a rise in additional income
gradually falls (see Figure 7.7). This Pigovian principle of diminish¬
ing marginal utility of money income has played an important role in
the formulation of many economic theories. As a general rule, the
application of such a principle is easily understandable. However, in
the case of LDCs, it is important to bear in mind that the marginal
utility of money income will be high as many people live on the brink
of subsistence. As such, even if the level of farm income rises (from a
 SUPPLY RESPONSE 189

Marginal
utility (dll)

Change in income (dY)

Figure 7.7: Marginal utility of income

very low level) it is very unlikely that the income effect will dominate
the substitution effect. On the contrary, it will be more plausible to
argue that the substitution effect will dominate the income effect and
as such leisure will fall. Hence, in theory, it is not easy to make a case
for a backward bending supply curve of labour in underdeveloped
agriculture. Also, the empirical evidence available so far does not
lend much support to the theory of a backward bending supply curve
in the agricultural sector of LDCs. Not so much that available
evidence fails to support the backward bending hypothesis, as that
relevant evidence of any kind is extremely scarce. More research is
needed in an LDC farming context.

6 The Concept of ‘Marketed Surplus’:

Some Methods of Estimation

The concept of‘marketed’ surplus has been regarded as important in

the literature of economic development. ‘Marketed’ surplus refers to
the amount of output which is sold in the market net of farmers’ own
consumption. So, in simplest terms, it is the difference between
190 AGRICULTURE AND ECONOMIC DEVELOPMENT

production and on-farm consumption. In this sense, the concept is

very similar to the classical theory of surplus. Assume that corn is the
only commodity produced in the economy and consumption per unit
of population is given by the fixed wage line. Hence, surplus is given
by the difference between the total product line (OP) and the wage
line (OW) (see Figure 7.8).
Given the slope of the production curve OP it is obvious that
production is subject to the operation of the law of diminishing
returns. It is also assumed that labour (= L) is the only input
available to produce corn output in the aggregate production
function i.e. Q = f(L). It is clear that the shaded area in Figure 7.8
shows the surplus. In a way, generation of this surplus is very
important because it is really the start of the process of capital
accumulation which is so vital for economic development in LDCs. It
is widely known that the lack of capital is one of the most serious
bottlenecks on economic growth of LDCs. As such, it is important to
ease the bottleneck as far as possible by increasing the size of surplus.
There are, however, different types of surplus. The surplus that
we have discussed so far is really food surplus. Other types include
labour surplus in the Lewis and Fei-Ranis-type models. Here the
concept of labour surplus implies that the marginal productivity of
some labourers in agriculture is equal to zero so that their transfer

Corn

Lobour

Figure 7.8: The origin of surplus

 SUPPLY RESPONSE 191

from the agricultural to the non-agricultural sector will not reduce the
level of output within the agricultural sector. Such surplus labour
may then be used for the industrial sector (see chapter 4).
It is also necessary to indicate that the agricultural sector can
generate financial surplus. This can be measured by the difference
between monetary receipts obtained by the agricultural sector and
the monetary payments made by such sectors. Problems of capital
formation should ease considerably if the agricultural sector can
generate substantial financial surplus or profit which could then be
reinvested for growth of both agriculture and industry as it has been
emphasized in the Jorgenson type model. Public policies like taxation
and borrowing could play a very vital role in the mobilisation of such
surplus from agriculture. It is also possible to use a positive monetary
policy whereby a high interest rate will be offered to attract more
savings, sometimes held in the form of gold, land and estates.
As regards the estimation of marketed surplus we shall first set out
the method used by Krishna (1964) in his seminal contribution. Later
we shall introduce some modifications introduced by Behrman (1968)
and others.
Let M be the marketed surplus, Q be the agricultural output
produced by the farmer, and C be on-farm consumption. Hence we
have the following definitional relationship:

M = Q — C-- (38)
If we differentiate the above equation with respect to the market price
of crop, we have:

dM dQ dC dC d\
dP ~ dP ~ dP ~ dT'dP • c=c[p,I(p)] (39)

where I = the total net income of farmers. Krishna assumes that the
farmer produces only one crop. A rise in price of this crop will raise his
income. If the farmer were only a consumer of the crop, a rise in the
price of the crop would reduce his total income in relation to the
quantity of crop actually consumed. In Krishna’s model, the farmer
is both a consumer and a producer. Hence the net effect of a rise in
price on income is the sum of these two conflicting effects (because,
due to a price rise, as a producer the farmer gains whereas as a
consumer he loses). Thus:

(40)
192 AGRICULTURE AND ECONOMIC DEVELOPMENT

If we substitute equation (40) into (39), we have:

dM_5Q_5C_McC
(41)
5P~5P <3P Mdl '
If we multiply the above equation through by P/M, and then arrange
them to obtain results in terms of elasticities, we have:

P Q P dQ
M X aP M q¥
C dC M PQ I dC
(42)
P X5P X Q "rXCX dl

/.e = rb —(r— l)(g-t-mkh)... (43)

The last equation has been obtained by using the following notations:

e= the price elasticity of marketed surplus

r= the ratio of total crop to marketed crop = Q/ M
b= the price elasticity of cash crop = P/Q x dQ/dP
g= the consumption elasticity of cash crop on the farm with respect
to its price
k = the ratio of the total value of cash crop produced to the total net
income of the cultivator = PQ/I
h = the consumption elasticity of the cash crop with respect to total
income of the former = I/C • dC/dl
By using the common statistical techniques it is possible to estimate
the parameters like r, b, etc. to evaluate the overall value of e in
equation (43).

7 Some Criticisms of Krishna’s Method and the

Alternative Approach of Behrman

Krishna’s method of estimating the price elasticity of marketed

surplus has been criticised on the following grounds:

(a) Krishna has used one crop and one price only. Such an
assumption implies that the income of the cultivator is obtained
from the sale of one crop only. Clearly, such an assumption is
questionable. As Behrman points out, the relevant income
consideration in the demand function for on-farm consumption
of the crop of concern is the total net income of the farmer.
 SUPPLY RESPONSE 193

(b) Krishna defines P as a relative price, but his use of it sometimes

suggests an absolute price. In his estimation, Krishna tacitly
assumes that the first partial derivative of income with respect to
the relative price is the same as the first partial derivative of such
income with regard to the absolute price. The ambiguity in the use
of the price variable may bias the model specification and its
estimates.
(c) Krishna tacitly assumes complete adjustment which implies the
elasticity that he estimates is relevant only for a sizeable number
of production periods after a change in price. If the number of
production periods which are necessary for complete adjustment
is high, the partial adjustment after one or two or three periods
may be of much more interest to policy makers than the complete
adjustment. Krishna does not mention what this elasticity would
be if only enough time for partial adjustment has passed.

7.1 Behrman’s estimation

Behrman follows Krishna in his definition of marketed surplus:

M =Q—C

Since it is not adequate to allow only the absolute price of cash crop to
alter (i.e. Pt) Behrman proceed to consider the total price from all
sources of income of a producer of the crop other than that from the
particular cash crop, (P2), and the aggregate price for all goods other
than this particular crop which is consumed by the producer of the
cash crop, (P3), Hence we have:

dM _ d{dQ) d(PJP2) dC d(PJP3) SC d\

Wx~d{Pjp2) dP, ~ diPjp^'^W, ~di'dP["'

If we use approximations, then the first partial derivative of gross

income with respect to Pt can be substituted for the first partial
derivative of net income with reference to Pt. Writing in terms of
elasticities, we have:

Pt dMt Q' f Pi/P2 55Q rqi tli

Mt dP1 MjL Q d(Pi/P2)J Lmi J
Pi/P3 dc 11 acHp^'j
C d(PJP3) jc Sljj I J
194 AGRICULTURE AND ECONOMIC DEVELOPMENT

P1/P2 0Q'
14
Q d{PJP2\ M!
I 5Q PiQ1
(45)
C d\ Q\ d(PJP2)

e ~ rbj — (r — l)[g 4 hk(l 4 bj)]-(r- l)hb2(l -k)--- (46)

where Q1 = planned production of all other goods, b: = the price

elasticity of the cash crop with reference to relative price (i.e. Pj/P2),
and b2 = the price elasticity of the planned goods (Q1) other than Q,
the cash crop produced with reference to Px/P2.
Behrman then re-estimates positive elasticities obtained by
Krishna and comes up with negative elasticities. Needless to say that
the implications of positive price elasticities are completely the
opposite to those which are negative. In the first case, ‘policy
implication for mobilising surplus will be to raise the price of crops
whereas in the latter case, the implication is to reduce price to raise
surplus.’ It is obvious why so much empirical research has been done
in the LDCs to obtain a fairly accurate estimate of price elasticities of
marketed surplus. However, when Behrman uses his own equation to
obtain the elasticities of marketed surplus for Thai cultivators with
respect to changes in both the absolute and the relative price level, he
has obtained positive elasticities since total production response has
been positive and the offsetting income effects on consumption are
absent. In the absence of income effects on consumption, the
estimated price elasticity of the marketed surplus of Thai rice is found
to be higher than the elasticity of price of total supply.

7.2 Some estimation methods

Behrman uses the following per capital demand equation in order to

derive income and price elasticities for on-farm rice consumption in
Thai agriculture:

C. = a04a1(P1/P3)t4a2Itn4ul- (47)

where Px/P3 = the ratio of price to the alternative crop price, and I"
= the normal income of the consumer in period t. It is given by the
following equation:

It = b0 + It-, + b1(Il_1-Itn_1)4Vt

= Z (1 ~ W) (b1It_,_i 4 Vt)... (48)

i=0
 SUPPLY RESPONSE 195

If we substitute equation (48) into equation (47), the reduced form can
be written in the following way:

Ct = a2b0 + aobj + (1 - bJC,., + a1(P1/P3)t

— aj(l — bx)(P i/P 3)t_ i + a2b1It_1 + Wt... (49)
where W, = error term.
Readers familiar with econometric methods will recognise that the
structural parameters of the model of Behrman are overidentified.
Given some assumptions about the error term, Behrman uses a non¬
linear maximum likelihood estimation method to estimate the
parameters. Interestingly enough, Behrman does not find any
substantial change in demand due to changes in income or price.
Behrman himself admits that such a surprising result may be due to
the following reasons:

(a) Data aggregation: the income or price effect on total rice

consumption may not be significant since a higher price of a
certain quality of rice may induce farmers to switch over to the
consumption of lower quality rice without changing the quantity
consumed.
(b) The quality of data that have been used might not have been good
enough to yield a more reliable set of estimates.

Notwithstanding these limitations, Behrman uses zero estimates

for price and income elasticities (i.e. g and h, respectively) in his
final equation for the elasticity of the marketed surplus with respect
to price (i.e. e in equation (46)), and with an estimated sales ratio,
r = 2.375, positive values for elasticities of marketed surplus have been
generally found for two different formulations for aggregate supply.
Behrman concludes that for both types of formulations, estimates
of price elasticity of marketed surplus are positive as the income
effect on consumption was absent while the price effect on pro¬
duction has been positive. Next, the partial adjustment may be of
more interest than total adjustment as the latter may need a long
period. Also, in the absence of an income effect, the measured
elasticity of marketed surplus with respect to price is higher than the
elasticity of total supply with respect to price. One of the major
implications of Behrman’s findings is to follow a positive price policy in
Thai agriculture.
The estimation method of Behrman has not been universally
followed. Nowshirvani, for instance, used a modified Nerlove model
to obtain the desired level of acreage under cultivation (AtD) with the
196 AGRICULTURE AND ECONOMIC DEVELOPMENT

following set of equations:

(50)
(51)
(52)

where T = time trend, R = rainfall, Y = yield of a crop, /J = price

expectation coefficient, and y = area adjustment coefficient. It is
interesting to note that equation (51) suggests that peasants form
their price expectations on the basis of both past price and yield-
changes. Farmers may also discount that part of price alteration
which is due to random supply changes. In the case of rice (though
not in the case of other crops) it has been observed that such a
hypothesis is valid when first differences of prices have been regressed
on the first differences of yields.
On the basis of his study for two states in India (UP and Bihar),
Nowshirvani concludes that generally the estimated price coefficients
for subsistence crops have been insignificant but significant for
cash crops. However, a simple Nerlovian model which uses only a
weighted average of past prices to account for price expectation may
not be adequate. The degree of commercialisation of a crop may have
a crucial effect on expectation formation.
Other forms of the Nerlovian model which have been used can be
stated as follows:

log A, = log b0 + log bj log At_, + bi log Pt_, + b3 log T... (53)

where A = acreage under cultivation, P = average farm price/whole¬

sale price, and T = the trend variable. In incorporating the effect of
rainfall (=Wt), and allowing for the effects of relative prices and
relative yields, we can write:

where Pt_, =one period lagged own crop price to alternative crop
price and Yt_, = one period lagged crop yield relative to alternative
crop yield. If lagged relative gross income per acre is GRt_, then
the equation (54) may be rewritten in the following way:

However, care must be taken to define gross income (=GRt) to

avoid the problem of multicollinearity (interrelationships between
any two variables in a multiple regression equation which are
supposed to be independent of one another) by using both gross
revenue and acreage together.
 SUPPLY RESPONSE 197

A number of other distributed lag models have been reported by

Parikh (1971) in his econometric analysis of supply response models
for Indian cereals. Parikh has obtained mixed results (i.e. both
positive and negative price coefficients for wheat and rice) for
difference Indian states - a conclusion which has been confirmed by
the findings of Cummings (1974) for rice crops in India.
It must, however, be mentioned that most evidence from India
tends to suggest a greater positive influence of price on the production
of rice and wheat than other food crops. The most consistent pattern
of positive estimates on the district level was found in Punjab (India
and Pakistan), Andhra and Tamil Nadu (previously Madras pro¬
vince). But in Assam, Maharashtra and Karnataka (in India), mixed
results have been found. In a large majority of districts, both rainfall
and trend coefficient estimates are found to be positive and statisti¬
cally significant to account for variations in both rice and wheat
acreage. In arid and semi-arid regions, the impact of rainfall has been
found to be striking.
Not surprisingly perhaps, in arid and semi-arid regions of India
and Pakistan, the impact of rainfall has been particularly significant.
Cummings claims that his supply model has performed satisfactorily
in explaining rice and wheat acreage variations in most parts of India
and Pakistan. It is useful to note that he has rejected the use of yield
variable in his model because of its unreliability, though his supply
model is basically Nerlovian, and includes both area adjustments and
price expectations. His basic model can be written in the following
simple way:

At = f(Pt-1, Wt> At_j, T)... (56)

where At = crop acreage in period t, P,_, = farm harvest price divided
by working class cost of living index, Wt = an index for rainfall which
shows deviation from normal rainfall during the period just preced¬
ing and during sowing, and T = the trend variable. Notice that
multicollinearity may exist between T and At_,. Cummings uses OLS
regression analysis to estimate equation (56). Due to the difficulties in
the identification of the parameters which influenced both the
adjustment and the expectation coefficients, Cummings, in his
subsequent formulation of the model has imposed restrictions on the
expectation coefficient. A different regression has been run for a
specific value of the restriction which varied from zero to two and on
the basis of the minimum sum of squares, the best one was chosen.
The index of weather that Cummings and Askari have discussed is
also of special interest as it really stands for an index of variation in
rainfall during the critical period of sowing and growing crops when
198 AGRICULTURE AND ECONOMIC DEVELOPMENT

soil moisture is badly needed. However, other types of weather

indices may also be used. But the use of the working-class cost of
living index as a deflator of farm harvest price may raise some
questions about the assumed similarity of taste (demand) of the
working class and the peasants. Although yield has not entered as an
argument in the postulated function of equation (56), in one
subsequent study Askari and Cummings have used yield expectation
as one of the explanatory variables of acreage.
As regards the ‘testing’ of the price responsiveness hypothesis for
other countries. Table 7.1 at the end of this chapter shows results for
different crops in difference countries. In the case of the Philippines,
although the price mechanism played a reasonably efficient role in
resource allocation (i.e. inter-crop cultivation), its positive role in
raising agricultural output has not always been confirmed due to
weak response of yield and output growth to a positive price change
(Managhas et al., 1966). The supply response model that has been
used in Philippines is basically similar to the one formulated by
Krishna. On the other hand, in the case of Indonesia, positive price
elasticities have been observed in many, but not all regions
(Mubyarto, 1965).
The basic Nerlovian model, in a slightly modified form, has also
been used to test the impact of land reform on supply responsiveness
of peasants to price changes. Using the data for some selective
Middle East countries (Egypt, Syria, Iraq, Jordan and Lebanon), the
authors have hypothesised that if land reform had a favourable
impact on the responsiveness of the farmers, supply elasticities should
tend to have higher positive (less negative) values after land reform.
Despite the lack of adequate data on time series in the case of Iraq and
Syria, Askari and Cummings claim that land reform had a favourable
effect in raising output and surplus. However, in the case of Iraq and
Syria, land reform did not have much impact on output.
It must be borne in mind that farmers react differently to different
types of crops, and a high supply elasticity for one crop does not
necessarily mean that peasants will also react in the same way for all
agricultural crops in general - a point that has been confirmed by an
empirical study by Swift (1969) for Chile. Positive price elasticity for
Peruvian rice cultivators has been observed by Merrill who used a
Nerlovian price expectation model to obtain his estimates of
elasticities. Merrill argues such ‘economic rationality’ is due to the
strong market orientation of a few large landowners in Peru (Merrill,
1967). It has already been indicated that in a large part of South Asia
farmers have shown a fair degree of price sensitivity. Evidence from
 SUPPLY RESPONSE 199

African countries also tends to confirm that fanners do respond to

price movements. In the light of such evidence, the case for a positive
price policy for agricultural development is strong indeed.

8 Perennial Crops

In the case of perennial and cash crops, similar supply response

models have been used and tested (see Bateman, 1968; Behrman,
1968; Olayide, 1972, Tomek, 1972; Wickens and Greenfield, 1973;
Saylor, 1974; and Ghosal, 1975). Price responsiveness of crops like
cocoa, coffee, cotton and rubber has been estimated by the use of
econometric models which tried to explain, say, supply of actual
stock of cocoa trees ( = Sc) by the expected real producer prices for
cocoa (= Pp and expected real producer prices for alternative cash
crops (= P'). The adjustment in the true stock is basically Nerlovian
as it is a proportion to the difference between actual and planned
stock of tree. Hence:

St —St_, = <5(Sf—St_,)... (57)

where S — the stock adjustment coefficient.
It is assumed that S? is a function of expected prices. Thus:

St = C0 + C1P^ + C2P: + Ut... (58)

Equations (57) and (58) have then been used with a price expectation
equation (discussed earlier) to obtain a relationship between stock
adjustment and price movements. The estimates of short- and long-
run elasticities for four different regions in Ghana in this sort of stock
adjustment model have been found as positive (Bateman, 1968).
Behrman points out that it is more useful to consider planned area
under cocoa (A[*) as a function of expected prices of both cocoa and
the major substitute crop, which is coffee. Thus, Behrman obtains the
following equation:

Atp= a0 + ajPj;, + Pgt + U,... (59)

Due to the lack of information about area under cocoa, Behrman
writes equation (59) in the following way:

Qt = b0 + £Y,At_, + b2P‘ + b3P‘_, + V,... (60)

where P* = the actual cocoa price at time t, Y; = the average yield per
unit area i years after planting, and Qt = production of cocoa. The
200 AGRICULTURE AND ECONOMIC DEVELOPMENT

problem of infinite series has been solved by a first difference

formulation that could be written as follows:

Qt — C0 + C1AQl_l + C2AQt_2 + C3APt

+ C4APt_, + C5APl_2 + C6APt_3 + C7APt_ni
+ C8APt_„2 + C9AP:_n) + C10AP:_n2 + Wt... (61)

where nx = the age at which trees first bear fruit and n2 = the age at
which second major rise in yield takes place.
Although Behrman obtains positive short- and long-run price
elasticities in eight African and Latin American countries, he points
out that in six of the countries there is little evidence of significant
short-run price responsiveness. Despite this finding, Behrman cau¬
tions against a policy of raising high prices by supply restrictions
unless such a policy is pursued along with stockpiling. It is important
to stress that, except for Venezuela, Ecuador and the Dominican
Republic, Behrman obtains high values of long-run supply elasti¬
cities. As such, the policy of maintaining a high and stable price may
eventually produce ‘cocoa mountains’. If this is true then it follows
that, in practice, it is necessary to follow the following three policies:
(a) a strict restriction on cocoa planting; (b) producers should operate
independent of the movements of world prices; and (c) cocoa
marketing boards should work as efficiently as possible.
In the case of coffee, attempts have been made to use the basic
Nerlovian model to test the price responsiveness of coffee producers
in a number of LDCs (see e.g. Frederick, 1965, 1969; Saylor, 1974).
Interestingly enough, prices did not turn out to be significant variable
in many cases in explaining variation in coffee acreage in Uganda.
Clearly Frederick’s study suggests that other factors (e.g. weather)
should also be taken into consideration in explaining coffee acreage
variation - a point which has been emphasised by Arak (1967) in her
study on the Brazilian coffee supply. Taxes and quotas could
influence price expectations considerably. Planting may be regarded
as a function of the proportion of existing trees (which are at least 10-
years-old) and the ‘optimal’ age distribution of coffee trees. The stock
adjustment model can be regarded as a function, albeit simplistically,
of expected coffee prices, existing stock of trees and the cumulative
planting. In a different formulation of the same type of model,
readers may wish to explain the ratio of new planting to the existing
stock, of coffee trees. The change in the age induced planting at
t( = AA“) can be considered as a function of the proportion of existing
 SUPPLY RESPONSE 201

trees older than ten years (=Tt). Hence:

AA: = f(Tt) = 2(Tt - T*) 2 > 0... (62)

where k = a function of expected coffee prices, T* = the ‘optimal’ tree

age distribution, and <5AAjy<5Tt > 0.
Expected coffee prices have been generated by Arak with the help of
the exchange rate and the New York market price of coffee. Thus we
have:

Pe = PmXm... (63)
where Pe = the expected price of coffee, Pm = the market price of
coffee, and m = the fraction of quota. The overall supply equation,
following Arak, can then be written as follows:
t-i

AA* - Uj (a0 + ajPt) T192o— X AA;

i= 1921

+ ^ilTt-(b0 + b1Pj)]. (64)

If we wish to explain the proportion of new planting to the existing
stock of trees in t, the above equation can then be rewritten as follows:
aa*
= u (a0 + ajP() T j920 X AA;
i = 1921

+ *2[Tt-(b0 + b1P?)]... (65)

where AA; = planting of trees in time i, T1920 = stock of coffee trees in
1920, Sj = the stock of coffee trees in time t, and Tt = the proportion
of existing trees other than ten years. With the aid of equations (64)
and (65), Arak obtains the reduced form of her model with which she
estimates that the short-run elasticity of Sao Paulo coffee planting
(measured annually) with respect to expected coffee price is + 2.02.
However, when she assumes that the age distribution of the tree-stock
can change the difference between planned and actual stock (i.e. /ij
and n2 shown as a proportion of trees which are at least 10-year-old,
i.e. Tt), then the estimated short-run elasticity rises slightly to + 2.28.
Limitations on data have prevented Arak from reaching conclusive
results in other areas, though the estimated elasticity (planting) with
respect to expected price has been positive in all areas. She concludes
that high export taxes have significant impact on price expectations
though public policies designed to control the planting of new coffee
trees have been virtually impotent.
Arak’s results are, however, questioned by Bacha (1968) who has
202 AGRICULTURE AND ECONOMIC DEVELOPMENT

introduced production cycles in his model of the coffee economy.

Bacha uses the following equations to show the relationship between
output of coffee (Qt) and area (At) with the following equations:

Qt = aAt... (66)
if t is even
Qt = bAt... (67)
if t is odd

where a = the output to area ratio in off years and b = output/area

ratio in peak years.
Bacha’s supply equation is then expressed in terms of lagged output
(= Qt—i) and past prices. It is assumed that price expectations of
Brazilian producers change slowly. We then have:

Qt = a0 + a^, + a2Pt_4 + a3DtPt_4 + a4Qt_2... (68)

where Dt = the dummy variable and equals zero if t is even and 1 if t is
odd. Bacha’s estimates of short and long-run supply elasticities for
coffee have turned out to be +0.23 and + 1.00, respectively. For
Latin America as a whole (1943-60), such short- and long-run
elasticities have been observed as + 0.276 and + 0.518, respectively,
while for Africa (1943-60) they are estimated as + 0.239 and + 0.374,
respectively.
In the case of other perennial crops like tea, evidence from India
suggests that although the short-run acreage elasticity with respect to
price has been positive, its value has not been high and usually varies
between +0.024 to +0.157, whereas output elasticity has been
slightly higher (varying between + 0.142 and + 0.351). However, all
price parameters have been found to be statistically significant at the 5
per cent level (see Rajagopalan and Meenakshisundaram, 1969).
As regards rubber, Behrman’s estimates of elasticity of supply in
Malaysian and other producing countries may be mentioned. Using
the simple supply response model in terms of current price and
introducting expected yield, rainfall and area as explanatory vari¬
ables, Behrman obtains significant price responsiveness and rela¬
tively important estimates for small producers. But for large estates,
the estimates have proved to be insignificant. Similarly, other studies
by Ghosal suggest a short-run elasticity for the Nerlovian model as
+ 0.22. When Ghosal (1975) uses a simple model with the help of
current prices and a trend variable, he obtains an estimate of elasticity
of price which is +0.12. Such a low value of elasticity has been
explained partly by technical reasons where output could not have
 SUPPLY RESPONSE 203
r-
<N fN
§ 5 oo
'O Og
a; C © ©
o o o
||
*3 ^ On rt fsj
^ SO r? —> o
x'O'O'or^t^o^tt^ONXoo r- r- OO Tt
pr^voppp^^HO^nmo , i—< o (N <N
d © © d d oddddddddddd |dd
+ + + I + + + + I + I I I + + + I +1 +
s: tj- os
a (N <N o
*i ■*—
'5 d d o
o o o
- O' oo m i ^oovONor^r-rt^-TfrNj'sO'or^^rsrsmmc^^
m *-h — oo i O ^O(N't00C'-;>Hqqn0fN-«00(N(N
d d o d o d oo© odddddddddddddddd
+ + + + + + C+I + I + I I I+ + +I++I++I

C/5 . __ C/5 C/5

5® cd 00 oo 00 bo GO
.£ ctf C G G c G
£ -o a e 1 £ e
o 3 £
•jc *C 5 cre E B
3 3 £
3 E
3 3
£
a m 32 Uaa-a -s u u u u u
C C/3 ■o»3§ c/5 T3 TD ■C/5 C/5 C/5 TO C/5 C/5 o3 C/5 C/5 c/5 T3 T3
c.S > > GO G G 00 00 00 G
M tic r be bo GO G C
TO / .1; ^ G c3 cC C C C 03
.£ .5 -c C C o' G 03 , oj
■r . J3 J3 j;
£ * 2 _c c« vi "n £ *G ’C £ £ E ‘C E |-s £ £ “
3 J2 M S S £
03 3 « c o o J=
C 3 03 £ £ £ 5
3 C/5 C/5 3 3 3 To I* £< G
3U 5
^ |
ed o3
oc
OC ^ <mZZi£ 0<<0(ju<o U c/i UZU<E<<

T3 ^oaN^oornr^ON
v nO vO t • sD ,0f<l|— sor^^r-vor^r-t^t^r^oovONor^oooomoco
/^\0'0'0'0'0'0N0'ON0'0f/‘iON0|/1'0OOv0Nt
•2
2! I I I I 1 1 I I 1 I I 1 I I I I I I I I I | I | | | |
^9^!Clfnfri2oorvv0(NfN<:7'OTiniTi^«o\0'ooa'Ooooo
£ a'a'aNa'ONONONONOONO'a'ONONa'ONa'ONONONO,'ONa'ONa'ONO'
Table 7.1: Supply elasticities, by crop and region

3 3 3
TD TD T3
CQ (3 A C <U
zzz 2 c
w <->
T3
3 >
oo .e'S 03 -2 w 07 C/5
C/5 C/5 £ £ -a 1) <D
"O "O t3 c3 (3 ll, 13 13
oo oo £ £ J5 JG
bt£^ G G <D <D
..2^
o3
"3
C/5
3 <d a> c c TD "O
G Cl. v_✓
£> X> US cd . . «g
u, *3 aa cq 52 3 2 3 J- 03 3
oJ 03 cd >* 03 g 3 OJ u. O
‘‘•eg -2 « w
’c**cT‘E 3 3 03 c« "O *5
§ bO "O T3 £ 3 TO J3 oo oo a
>.
^ 3 a) <DTD
-C JS •C «» 03 cfl
3 35c3aJ=£'CiSc'1
= 3 3 G
a3c.3^3'a3 TO
os c c
03 cd 03 03 00
£££sDaBoSSS<I<O^S2i^^S o- a. CQ CQ W

uu
§■ U
CJ
204 AGRICULTURE AND ECONOMIC DEVELOPMENT

*o) •O) SO 04
§ .£> 3- O O
Oft .2 d <N c4
0 4- O
<—*
O
■*—•
||
OO r- 00 VO «o> 0 04 co 3- 0 00 ’<1-
3 S5 O *o) C4 co •o) 3- 8 co vo) SE . 04 O
O d d ^ 0 d 00 d d d d 0 0 d ob 1 d d
ID D
4- + 4- 4- 4- 4* G 4- + + + + + + 4- C 4- +

CO vo) 00 VO 10 o-
co (N <N 10 Os sO
§ & i"-<
d d d d d
oft ;s
0 0 0 + O 0
-*—»
2
00 SO Os 00 co 0 Os co 3- 00 vO OO O IO Tt
0 SO 0 <N s Os 04 0 O O s 04 O
o d d d d 0 d 00 d d d d d d d d ob d d d
4/
+ + -I- + 4- 4- G 4- 4- 4- 4- 4* 4- + + c + + +

■oG ■o *o C/5 C/5

3
G
3 G
-o
G G 00 00
3 3 3 G G

00 00 •a E E
© G
«4S s s
G
OO O TJ
1ca E3 3E
a: QC G3 iUO
i3 £
3 c/f c/T
3
«) ifl G C/5 sp-a ■&
3 .2 c c
u 43 C3 3 - 00 00 -G
_ —
u _ c <u 2
£ .2 .2 •£ g?.g C/3 3 3
■c%-c% I i-S-S-5 C 3 «-*3
S
3 «-»
-22^-2 2 *1 —- ’ C ' c
■S. 3
* s?!
~ 2 2 ~ ^ -3 ®p.*
3
M M * -g -g S £ ^ 3 ^3 a: 3 K 3 3 3 g 3 i if C- C/5 C/5
< < cq m < < S oi S 2 U U & CU ^ u 2 £3 < <

C4 CO CO vo VO) 3- 3- 3- r- r- SO Os 3- r- VO) Os O r- SO
r- VO so SO SO so SO so so SO 3* CO sC SO so CO t"- SO vC
11 11 11 11 1 1 1 11 1 1 1 1
•2 1 l l 1 l l 1 1 1 l 1
v! CO
vn s
O-
CO
r-
CO
CO
lO) VO) VO)
CO
VO)
CO
VO)
3"
VO)
VO)
VO)
3- 0 O
0 VO) «0)
00
3-
Q O
VO)
CO
VO)
CO
>0)
O
2
< ! Os Os Os Os Os Os Os os OS Os Os Os OS Os Os Os Os Os Os Os

a>
(23
i-
"O *3 D
•2 3 G 3 <D <D
a. a <1> O 3 3
oj T3
3
40 u C/5 I 00 00
T3
3
VO)
£.2
T3 ca
OC/5 <D 00 43 p .5
"O . 3 G G _o C/5 ■ '
3 3 2
c
3 cac S
^ ao a. o D. 3
a op 2 -O X> £> J3 .o So
773 .2
33 _
*—* G cx <->
a. n ctf 3 3 3 3 3 -3 S._1

Dh-g N3J3h ■S
cx
3 3 3 s J2 s =* s f3
,-c'c,'c,'2''2,'5' "O
Iraq

-C 2
>>
00 (D
3 pf -3 435 is 3 3 3 3 3
3 nt u .5
w H •—> O-i CU CUO 2 ft. a. a, a. a, SOX
Table 7.1 (continued)

§
3
■5 H
<
W § w
§■ U 2d, sc
£
 SUPPLY RESPONSE 205

no
r-
o< :§ +d
g? n 0 0
§ -3 nO'^MOiO ' t* m m <N tj- CO
>3 ^

+ 0.94
+0.50
+0.27
l <N © <N m <N
o©d©© d 1 1 od d m Tf d d
1 + + + + + + + 1 + 1 1 +

<N o
<N
§ &
°? s o d
o o
i 8 m oo mm o r- —• <N O «n o
Co

+0.39
+0.22
— 0.63
to ^ . O I —' <N o on O vO Tt oo <N
|o 1 d d d o d d d d o d d d
+ + + + + + -f 1 + + 1 +

73 73 73 73 73 73
3

Askari and Cummings

c/5 c 3 3 3 3
bO bO Cd cd cd cd cd
e 3 2s
c/T c/f c/T c/T
cd E £ 5 S> oo oo oo 00 00
o
•JC JC
3 E E
3 3
u o
6 .S
3 §
u
3

§
3

E
3 3

E
3
E
6 E E E E

Raj Krishna
d J1
I 3 •g "O 3

Cummings
c/5 73 3 3 3 3 a
fc c3 3cd c
00
„ J*:
u
^ ^ „ .*
U
~_ M „ M ^
•5 c '5 | §
3
o C C C 'C ^
33333
■c -c *|**| t—1 *j3 ^ "C
3 3 cd ^ ^ J2
s JS E 8 ^ M 1 I tC -* X -* X
o - 3 3 on c^5 00 C/5 C/5 ” C/5
^Z<CUtt U- << < < < <
1914-46

1955-66
1950-67
<N <N rr r- 00 r- On ON OO (N o (N O r-
^3 SO VO o SO o NO CO U-5 no
i
r-
i
VO
1
r-
i
NO
il 1
NO
•2
£ m o (N
1 1
tr» Tf o m
1
O m
i
r-
1 i
(N
V-5 CO o *T m cr> cn Tf SO cn NO cn
Os Os ON ON ON On ON ON ON ON ON On On On ON
1-H l-H •—H —-* i-H *—• i

C
•2
s V)1) 3
t£
w
to
o>
C/5 C/5 .£
c<
d gx
-5
u j§ CQ
-V c c « C
Punjab
Punjab
Punjab

Oh On cd C3 00 2 3
«- u. u. o ^ to u w'C c£
a cd cd cd
^■2, 8 5 SS .'■« >s cr cr 73
•*-»
♦-»
*-•
<—•
r- bk
*
5
bD
‘C
>s
’C
>> cd 03
D D ‘ 2 oi > 0- 0. cl, W C/5 C/5
BARLEY

§■
 206 AGRICULTURE AND ECONOMIC DEVELOPMENT

o O
§.5 *o Tf
os .3 d © o
go 5 o o
5 a
-3 hJ
r- on o no CM —' C" oo VO tJ- vO VO *0 ON
s —« —< NO CM 0 0*0 CM VO — vo CM CM
© O^rid © © d o d d o d cm d d
+ + + I I + + I I + + + + + +

c
3 ~
^I '3 ©
~
o
© <3 —h co o o co r- vo vo co r- ^ CO
vo CO VO —< O O *0 *— o CM CM CM
d dddddod d d o o o o d <n d O
++ i +++ i + I + + + + + + I +

C/5 C/5
T3 •o "O T3
*3
*3
60 60 3 c 3 3 3
3
04 3 a 60
C 3 3 3
a:
3 3 3
3 3
CO
"O BE E 60 60 60 -a 60
60
c> c
i s £ C C C 3 3 g? % g?
3 3
3 3 3 S J3
3
3 6 1 S i £
uo U J2 x B E
B B 1 £ E
T3 T3 c/5 "O 3 3 3 c bo 60 3
E 3
E§ 3 c 60 C
u u u
3 3
in .In 3 03
.s & £ 3 * „ •g c c/d
U
. ^
u
u
„ J3
X £ ^ ^ ^ V« M U M £ * = -c-c
. C *" C-.
GO > 3 3 E<'rja333c<33a33
1,3 *2 •rt S3 5 3 3 3
c o C/5 C/5 ~i
rj Z^-xX^X^X
N-' 3 C/5 C/5 C0 ^ C/5 ^ 'S’ 3 SC u -* I K 3< E
Jz <*> c/5 ~ cr>
in Z < < U Z U < < < < as * (75 < s< < <

o co moot'~(or-~oooo<N O vo ON
r- vo '£>'0'0^0'0'0^)h- vO VO
O CM
r- r-
VO CM
vo r-
vo
VO
CM
r-
•S I I I I I I I l 1 1 1
1 11 1 1 I
v! O CO ro-3-movoo—iTj-m CN 0 0 CO VO CO
VO VO vovovorj-covovovo VO vO —^ X VO vo VO vO vo VO
£ ON ON OnOnOncOvOnOnOnOn ON ON ON ON ON ON ON ON ON ON

s:
•2 •o
|b 3
1)
T3
3
3 *-
3 Oh > 2 j2 c c 3 3
x o 2 2 -O X) 3
3
O
£ *- ^ u. S2 ‘3 3 .2 .2 3 3 3 w 3 3
S' ^ •5L:S E | ^2 ^ a- cr . . >> a. 3 3 3
3
3 Z.S 3
3 3 J- >i *3 *3
X
>> 3 3 3 3^
E £ 3 a>
S c^QI£0hCLh on Oh Oh E W c/d in hJ
Table 7.1 (continued)

>< |
u •-
J s
w
aS ? N
§■
< w <
<G m
 SUPPLY RESPONSE 207

r^ CO
§ £ o
CO
oo
CN
v-H O
ro r-*
*

«C 3 + d d d d +
o o o o O
|1 +-1 ■«-*
CN
•3 «*j •— Os so SO
o co o SO I
so
T—I
^—; 1
oo
‘'l 1
so ro OS ro CO CN
° O SO CN SO SO
<N
co
co
so
<n
so
© o—^oo o- o o o o o o © d
+ + + I I + I I + + I I
co OS
o •'3-
co o
©
§ o CN Tf CN CN
©
SO Tt
os u
© Tf 4* © © + © ©
I
■»>»* + o o o o o o
oo (N ri OS O Tf <N ~ O O so oo rt o r-
CO
CN © © O p (N oo o »0 © O co so ■ O -H CN to co
W) © ~ © © x © © © © C/5 © © © © © © © x © © © o d
o
c ++ + I I I + + + + + + I §. + + i I

-a T3 T3
T3 c G G
G
c3
ed
5,3 S c
CO
os S3
1/5 c/f
O o
•o
c •o-’g
c h
00
G
00
c
00
G
■JK o 03
a o«
03 jS
't X e E E
C 00 S £ «s £ E E
X
c G 03 03 c c M-O 13 ~
G ^ 3 3
-G
5
•3 cdC c
o3
g
ctf
o3 13
> o- >
o3 *3 *2
a.22 o'
o3
> •- .-
o3 CJ•S g
o
VM
00 ~ S 6 * -
_ 03 O
oo jo
03 ° «-<— m
w pq «3 G G
^ u
3 *C ’C
c 3 X5 —, ^ £ 03 03 03 _ ^-g 'c ’5 '51
_, G I_ijj >»' ~03- -O
w _ -'t "J "O
3 W "O
U TU U ”5
3 "O ^ 03 3 03
§ cc -c x: o3 03 oo3 *P
<u o H ^ 5 ^ 5
TO
^ cd1 CJ
o,
o3 ^ 03 .£ -c S -s ^
j*soo 2C on
CQ CQ ct U GO CO QQ < < <

C<) m SO O so SO in iv'T}-iniAiasa\Tj-oi^rv-o CN r-
“a SO so SO •or 3 s SO SO SO r- SO
| i i l
•2 1 1 1 1 I I I I I I I I I I
k! SO Os »o Q r- 1 tj- x — O r-
<0 Tf lo ‘O SO rT SO tj- so u-i
a. Os Os Os Os Os Os Os Os Os Os Os Os

s:
•2 J3 JS

c< X u. i_
<D
G
- ,. o3
’a Q 3 cd
c
a g -S, £ o3
03 03 G C T3 c b ^ "O a -a
IE J= x: 3 3 3 3 «9 C C cfl
a. hhlXO. a, K < < Jz >%
00

2
<
0£
< O
> D
< < ffl „ oi
<
z
w
1
m oi o U
O S 0 £ 0^ < tu H
V. < < D < O o OS 0£ tq
o U oa U aC •"“J on O o >
 208 AGRICULTURE AND ECONOMIC DEVELOPMENT

(N 1/-5
§ ^ m CN
o rn
o O
II O vO vo OOt^'t’-oo^^OOooMvuri^ —« m
>3 fej VO m oo cnJ—•c^Ommmcvw^OvqfNTfTtoOoq . ^ co
m © O 0000<N000c500000Ttc<i0 loo
I + + + + + + +I++I + I + I+ + + + + 14-

oo oo
3 C1
r-
^ ‘3 O o
o o
s J3 ,
rf ov oo Os osoococofot'"<N<N©os~HOOooroos'—■
(NOnOa'^TtrtTtOlNMrrincO'n
—< o
<N oo O oo *—• <N <
O o o O ooooocioooooooooo o © ©
+ 1 + + + + + + + I 4-4- I + I 4- I 4- 4-4-4- 4- I +

T3 *3 33 •o
G G G G
3 3 3 cd GO
c
OD GO GO GO £
g g G G
3 03 g
•js: J3 w 3
S 1 1 I <U (j
3
£ 6 B i T3 •-

►.A Kf
GO
3 3 3 Kf\ hf. A *sA Kf\
GO GO GO GO GO GO rrt
»*A »*A »-.GOA KA
C/5C/5C/iC/5C/5C/5CC/3C/5C/5C/5C/5C/5C*5
GO GO GO GO GO GO
« c'3 § C c« T3
3 GO C
u
„
u
„ M
u o c c c c c c > G G G C3 G C C3 .fS-S S C 03
‘C c •c 'c s ■c ‘5 E E E E E 1 js £ £ £ £ E E £ SB)_ .C
CA E '£
«s =2 3 « .3
E E E E E E -a E E £ E E E E E 3 S g
^ ac
■3® J X o E J2
3 3 3 3 3 3 ^ 3 3 3 3 3 3 3 3 3 0 o .3 -
< S u u u u u u 2 uuuuuuu UUIZ a3U<

«N <N 00 00 oo OO OO CO
r- r- vo r- r- VO vo vo VO vC VO sO'OvOvOvOvO'OvOvOvOvOvOvo
1 l ii ii
CO
ii 1
vo
11
»o
11
CO VO
1 i
Os
1
1
i
1
r-
I I I I I I I I I I I
OvTt-m^OO-OfOn^vO
SO VO ‘ZT VO 1/-5 vo vo CO CO Tf vo Tf
Os Os OS OS Ov ov os Os Os OS Os Os
T-H

3
-o •r-' o
« E
z
•2 •o.E S
So a>

<<
T3
3
1 d vOsS <D
a>
-C -C ^ -C T3
C/5 C/5
M e fc. 3
.00-0
w cc 2 2 -D •£
3 3 b33 53 O ts 3
3 ^ -0 X Mh cd u
3
cd _ D,
3 !3 5 O ^ W5 2 <3 3 2
2 GO"g3).2 lab-E) 3 Uh X) i-; 3
O' 23 .2,-2 C/5 -O 3 T3 2 .2. x
3
T3 JJ
3 g GO c 3 ^ £--* c
p •£> ed 8 'C 15 5 c5 IS *5'2
c ^ C G
3 3 o
C/3 00 £ W <OS2o,o;S<hcihCQcq0h CQ Oh co O S £5
Table 7.1 (continued)

C/2 C/2
H z H
W
< Q
UJ Q Di
aa 2
H D z ^
W S Q <C
s ^ C/2
J z < w D
<
Crop

< O C/2 CL 5
a: u < ^ a:
S a DQ C/2 Dfi O
 SUPPLY RESPONSE 209

se =\ *o
3 no
<< 3 © -H O

||
■3 ^
o
>n ^'t tj- (N rt CM
ro
CM
o
■*—>
OOMrtON'noOfN't'HTtinON
o
*—>

'—; r- O —- *0 CM NO 3 Tf ONOOOr-rNOONOrfu^fNfN
© m' © © © ^ cm :<nooooooooo
I ++++ I + ++++++++ I +++
s: ■*fr ON oo
3 rs n r- NO co
© o © o
o o o
o 3 hO^vOO'^fNfNfN'^m NO O'X't lO'OiOHO^mfNO'O
-HOOO-i^ffwN(Sh(Nm oo 'OOTj-rnm—nfNr--
oooooooooocm CM ooooooooooooo
I ++++ I +++++ ++++++++ I ++++

XJ -o
c c C/5 C/5
bo bO cd cd b0 bo bo
c G G G G
i s bo bD £ EE
■jc
£
3
s G G
3 £
3
E E
3 3
6 £ -G
3 u u bo
£ £ cd cd rs U u o
T3 c c/5 c/5 "O G 3 3 G G K ot « a C/5 T3 '"O C/5
c bo bfi c 3
u u
-C J3 T3 C JP cd bflC C bO
cd JS § g c cd OC cd cd C cd cd C ^
t* • « ._r ^
c x .5 >
cd i
u- _C B £ „ w ^ C ‘C c c «* -JS 3 2 -c g -C 'C ‘g w
cd < Cd * ,5 C 3 Cd c
* -O
0* cd B B I-
11"
3
Cd
3^
« - I « -* 5 ^ S n
3 3 —00 C/5 ^ Cd Cd ra o O02Jc/53c/5C/53s-/
3 c/5 C/5 3 W
U U < z I: o< < ocoi^^^uScuccuZ

NO ON oo r- OO r- 00 •O r- o m On ON 00 NO OO oo OO ON
sC NO NO NO NO wn O NO o NO NO NO CO NO NO NC SO NO sC NO NO
•S 1 1 1 11 1 11 ii i 1 ii 1 ii i 7 1 1| 1 1 11 1 11 11 1 ii
v! Q r- oo 00 m oo o <N CN o 3 r- CO r^> NO Tf oo
vn O Tj- to >ZY m d“ un NO <N fN i/n sO o un ■^t >o NO NO NO NO CO
£ ON ON ON ON ON ON On ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON

3 3
-O *0

3 zz -s
.O — —i <D
Jo
E E 09
cd cd.>n l_
H H G Oh
^ G
cd cd
i Cfl W -fi Im r- 3
I cd CTJ
1 i ±1 73 S 2 2 d o
c
b0
n ed cd g S'‘■a •« "2

Q z
w
w
o
H
C/5
& z
M
H
O
J u
 *

210 AGRICULTURE AND ECONOMIC DEVELOPMENT

o
sO tJ- r- oo o
§ £ <N r* On oo oo
<*C 3 r}- o © o d
o o o o
||
oo r- oo ON O x n ^nn-os'noo'nt^^to1^
CN Tf o o 5 n 'O |»-<o^Hr^ooooo'Or^i—inO'O
d d o odd ! HHoddddddd-o
+ + I + + 4-+ + 4- + + + + + + + + + +

r- — O Os r—
no r- oo r- Tf
d d o d
o o o
o .« - -H O OO M O' ON v-t ONoooooooo^Ooor-oor^^Tj-ooo
O Tf Tt rt fO O Tf oo rnriinnoO'^or^'cior^inr-’-'Ooro
55^ o o © © d o © d oddddddddddddddd
++++++ I + I 4-4- + + + + + + + + + 4- + + + 4-

T3 -o *3
G G G c/> C/5
3 cd cd 00 00
G G
c/f c/f c/f
oo 00 oo £ 6
§ G G G £ §
•JC 3 s
£ £ £
3 cd U U
Tr>
>
£ £ £
G C/5 C/5 3 3 3 TO c/5 T3 (A C/5
-G 00 00 u u c 00 G 00 00
C/D G C u
cd G cd .S c 51 a G G
£ *c *c -c '5 G -G cd _ G cd •£ cd
G G cfl .« td « t C « G 42 cd d Ge e.g
6 „
G 4^
■^i-^DC^CElj-srS in x: O ■ r -G G X> X} C G -O
c/5 ^ oo ^ c/5 JP C G
i—.O 3 c/5 3 cd c/5 cd 3 3 cd
U 55 55 < < < ~
So<z«wi£a:< u a: < cc u u c*

—< ON r- r- ON r- <N (N — Wd r- NO CN m On ON (N ON ON <N ON <N OO ON ON

NO wn NO
| mii ii
NO r- r- r- NO NO NO m NO m NO NO NO NO NO NO X NO cn
1 1 ii ii 1 1 1 i
1 1 1 i 1 1 1 1 1 1 1 1 1
iS OO m o o Os df m OO Ud ON 4 d 4 ON ON cn o ON
Tf «TN •vd ON T-H NO s l/d Tf © d •TN */■') Tt Tj- »XN IZ5 Tj- _
ON ON ON ON OO ON ON ON ON ON ON ON On ON ON ON ON On ON ON O' ON ON ON
2

G 8
s:
cd
o cd .£ 43
> o
•2 C
o
"O
G O TD
CdC/5
£ 00o
cx!>
£ Q < cd cd cd O
G ts OO "2
00 oo 00 *6
Td 3
G G G 3
G
cd
G
cd
G
cd
^ 2^0 o o Im
cu 3
5 — CO CQ CD -
C/5 cd cd g
Cd .22 .22
a a d. £ u. l-
is 3
43 cr 3 Q,
2 cd cd cd 00 00 00 cd 3 if* M e j2 g £ £ £ S.-S£-c *4 ‘C
hScLhOhO- WWW C/D P H
Table 7.1 (continued)

Z £ u.
o -S <
Lh
cr O zH tu
§■ o ^«-» w
u id P
 SUPPLY RESPONSE 211

§ On vO <N VO ^ sO
Os VO Tf Tt- (N —
's © © o <N —3 ©
o o o O O
il
SO
Os
^ n oo Tt M VO VO Q (N lO
vO OO tJ (N Tf <N O O ^ O (N
—• m
<n oo os vo m
— Vo OC fN m r- VO —<
O ©©©O 0 0 0—3 © © © © © O © o —3 o © ©
+ ++++ + I ++ I + + I + + + + + + + +

VO o o
§ £>
© © © o o
S o o o o
o a ti- o as o £iiCi2!Cvo9£^Srs|T-'00"OoooN<OTi-
vo m ro rt ■'tC'lrnp—;pp(NC-'OlovDrf-HfNr<-i vo o <N OO
vo «n o
o' © o' © o’ oooooooooooooooo dodo
+++++ + 4- I + + I + + + | + + + + + + + + + +

C/3 C/3
OX) OX) oo OX)
c c c 3

£ S £ E
© E £ E
■JK £
3 3
3 c u u
-a 3 <n
OX)
K C/3 C/3 ’TO
OX) OX) c
o3 c ^O ca3
c OX) > -3
3 s
c TO 3 3 rt 3 (V c3 s •- cn 03 >
c3 03
e fc a u. £ £ u- £ W •g ’c b
-O X) Cd C r; Cd
1 .§
EC/5 2^ 5 -3 ^
£ bx 1 E >* o ^ 3 3 £ ^ <D
.53 cd >©r3r‘0^‘2c2-:3:3c'5:::3^c/33'OaOcSa3'O
3 > 03 3 T3 -XO c5 -*
oC U c*
iuo^<<uu<uz<u<QZa:^z oo 2 <

fNO'fOvcxr^i^^r^t^Mt— [ [— [^^orfotNoon^tTt'tin^t
^3
■2
'OlOvOvD'O'O'ClO'OiO'OiOVOinvO'oS'O'tvO'tSSSvoS
I I I I I ! I I I I I I I I I | | || !ll|||[
v! ^!^222122,0'^'t,i,''10'00rt®',1voO'0'o-HOMr'4

3
_o
*35
3
’>
T3 w
ctf •3 cd
c z
•2 vO 3=Z
Sc 0> 45 -3
C/3 —h

"O 03 £ 1i
33 33
cS is o3 a
C/3
a
C/3
i)
C/3
i>
C/3
a> cr3
-3 H,
T3
o3 2(2 »
3 -a TD TD ■-
c/5 — it?
o3 nj c/5 Oh Oh
b u cd c« «g '5.
d3 -53 J3 J3 E 03 i-
03 .2 >
*— i_ . ^Sa
OX) "ob OX) OX) N .52 OX) d
-3 *o
?5S
3 3 3 3 3
3 3 5 cS£;'« ^
o3 03
^ «
OX) cd
§ -2P J2 - - c3 rb 111
£fflcofflcQh<<fflOSSS£a,coZ2DD Oh 03 < 2 a.

O
U
U 0£
-J < <
< ca
§■ o
£2 O D
C/3 £— C/3
212 AGRICULTURE AND ECONOMIC DEVELOPMENT

o
3 vo s
© ©
o o
H <—•
OO 'O ^ o ^ OO in oo ooO^h- — ^ <N
n oo n
cn o r- Tt OO OO a\ M ^ ri rt n m

^ d d d-^dddd © © O © © © © ©

4- + -f + 4- + 4- + + + + + + + + + + + +

, r-
O 00 VO <N
s:
-* m> 00 (N <N <N 00
0 O O d r4 d d
a; •5
■«>a1 0 0 O 0 O 0 0
■*-*
03 -*—»
o a <N , m3 00 m CO <N m fN m3 VO 0 ■*t 0 <N m 0 00 0
7f SO 1 vO m3 0 VO i-H • <N O <N <N r^*
co hj CN d d d d
O 0 1 1 d d d 0 d d d d d d d d
+ + -1- + + 4- + -1- 4- + + + + + + + + + 4-

c
ju
o 3
-o
c
cd

c c G G G
cd cd cd <u
G -O
cj o
* r *c cd cd co
73
„ cd s
73
£ £ S S 2
*0 'O .ti .ti X X flJ X
D <L> - -
!-•■* ^-h cd -G *G *G »-G -G *-G w v w w ^ a^
U(U2U(uuuu(uu.ih-m-i-c3u|-c3 c? G cd
CD CQ CQCQOCQQ3CQ0QCQaafeU,2SSCQCQ<CQ CO CD CD

fN CN ro ro m PO m m m m OO OO •^t 7t 73- O m3 m O m3 O OO
VO VO VO VO VO vo VO vo VO vo vo CO m VO VO vO X X m3 f^3 vo r-* vO VO vO
1 1 1 1 1 1 1 11 11 1 I1 1 1 1
•S 1 1 1 1 1 1 1 1 1 1 l 1 1
Ov Ov r- OO r-- r- r~~ r- vo vO vO X m OO 0 m> G OO <N cn m
Tf r}- rf rt 7f 7t 7t 7f tj- tj- 7t <N <N $ 7t ^t T^ ro <N cd m3 G- m3 7t m3
£ Ov Ov Ov Ov Ov Ov Ov Ov Ov Ov Ov Ov Ov Ov Ov OV Ov OV Ov Ov Ov Ov Ov Ov Ov

-a
G
73 C/3 cd
a/ * to
<U
X N
3 3 cd 2
2 .o cd
T5 ^ o
•2 £ a
<D G 3 3
3 cd
C/3 X
cd cd to CO
CeC « s
OD CU CU
*3 3 •c “>'3
<D G
cd cd cd CD tS 6 O O o «2 .5 3
S, O o 0> C/3 cd cd „
cO .2
■ g T5 c
cd U 2 o I cd cd CO CO
73
C
cd
G
cd .2
G Jr- - >, « cd — 3 -a 12
cd, cd cd -&<% §
73
- <U C G «
N 'S
G o n OJ G G ^
S. G G C
XI .3.3 a
X X .SP-2P C cd
0§Gg0530)«30<D(D<L)^^
DO 00 (D
c * 9
t-
9
U.
a OOZZ^UcqwQ>DD^^<cp CO CO fflU J
Table 7.1 (continued)

U4
< w
o U-
u u.
§■
O O
U U
 SUPPLY RESPONSE 213

Ti¬
§ £ c's
=< c o o
o
il
>3 ^ ^ On Q ^ (N
^ ^ ^ (N (N
O O O O O

+ + 4- + +

c O ir Tt VO o OO Os
fS O m rf rN c<\ eN m
O © ' © «n © © © © ©
o o O o § o o o o o
•*-* •*-» © •*-*
o
^^0^2 + Tt (N (N »
o -- (N o O - (N o (N M
0 0 0 0*0 O o c> o o’ © ©

a
c
Cd
a. G C C _
o cd
£
cd
£
cd w cd d
£ -a J= TD T3
GO c
cd 5 E E p E >. o >>'>>
cd ,c u-g-S-gJ2^i2iS'5'S
DC U trtfflcDcoOOOOOO

’7!’rc5r^LOTf'r^rN,r^t^'TtT^
ts sOvOsO'OsOvOsOr-vOvOvOvO
.c I I I I I I I I I I I I
v! -oona'hO'XOwooa'O'
(Nd,|^dd,d,d,|^d',td,,t
£ Ov^OnOn^sOnOvOsOnOnOnOn

cd
•c
<L>
o o
GO

s:
"6 O ^ z
•2 G
u
cd cd cd U

c<
£ c/5
E tS
c/5 4>
C/5
cd
cd W
cd _
cd cd
£ v g IS i* 1 « cd
.. cd cd
S' S' S'— c 'c 'c -c ‘c _ _
_
-O ..
cd cd Cd 2*0 ^ X) GO GO GO G{
£^2fS=ZjZZZZ

on
J j
w
06
W O z
CQ ss Q£

<« JJ PJ
§■
wD << Hi
H Btf a- a-
214 AGRICULTURE AND ECONOMIC DEVELOPMENT

been varied much in the short-run and partly by the argument that
producers are not profit maximisers.

9 Conclusion

It is interesting to observe on the basis of the evidence available so far

that producers of primary products in developing countries tend to
behave ‘rationally’ in general, though there are exceptions, parti¬
cularly in highly subsistence economies. Although it has been
observed that in many cases, farmers are ‘price responsive’, it should
not necessarily imply that other variables do not matter. Indeed, in
many cases other variables, like weather condition (rainfall and
temperature), soil moisture, family size, education, irrigation, farm
size, wealth, income and resource endowments, could have significant
effects. A number of these variables have not always been tested
vigorously. Future research may well be directed in testing the
significance of both price and non-price variables in affecting area and
output of agricultural crops in LDCs.

References

Ady, P. (1968), ‘Supply functions in tropical agriculture’. Bull.

Oxford Inst. Econs and Stats, vol. 30.
Arak, M. (1968). ‘The Price Responsiveness of Sao Paulo coffee
growers’, Food Res. Inst. Studies, vol. 8, no. 3.
Askari H. and Cummings, J.T. (1976), Agricultural Supply Response:
A Survey of the Econometric Evidence, Praeger.
Bacha, E.L. (1968), ‘An Economic Model for the World Coffee
Market: The Impact of Brazilian Price Policy’, PhD thesis, Yale
University.
Bardhan, P.K. and Bardhan, K. (1970), ‘Price Response of Marketed
Surplus of foodgrains’, Oxford Economic Papers, July.
Bateman, M. (1968), Cocoa in the Ghanaian Economy: An
Econometric Model, North-Holland, Amsterdam.
Behrman, J. (1968), Supply Response in Underdeveloped Agriculture:
A Case Study of Four Major Annual Crops in Thailand: 1937-1963,
North-Holland, Amsterdam.
Cagan, P. (1956), ‘The monetary dynamics of hyper inflation’ in
Friedman, M. (ed.), Studies in the Quantity Theory of Money,
University of Chicago Press, Chicago.
 SUPPLY RESPONSE 215

Cummings, T.J. (1974), Supply Response in Peasant Agriculture

Price and non-price factors, PhD thesis, Tuft University
Dantwala M.L. (ed.) (1970) ‘Symposium on farmers’ response to
pnces , J. Indian Soc. of Agr. Stat., vol. 22, June.
Falcon, W P. (1964), ‘Farmers’ response to price in a subsistence
economy: the case of West Pakistan’, Amer. Econ. Revi., Papers and
Proceedings, May.
Fisher, F. (1963), ‘A theoretical analysis of the impact of food surplus
disposal on agricultural production in recipient countries’ J. Farm
Econo., vol. 45, November.
Fredenck, K.D. (1965), ‘Coffee Production in Uganda’, PhD thesis
MIT.
Frederick, K.D. (1969), ‘The role of market forces and planning in
Uganda’s economic development 1900-1938’, Eastern African
Economic Review, vol. 1.
Freebairn, D.K. (1969), The Dichotomy of prosperity and poverty
in Mexican Agriculture’, Land Economics, 31-42, February.
Ghoshal, A. (1975), ‘The price responsiveness of primary producers-
a relative supply approach’, Amer. J. Agr. Econ., vol. 52, February.
Khatkhate, D.R. (1962), Some Notes on the Real Effect of Foreign
Surplus Disposal in Underdeveloped Economies, Quarterly Journal
of Economics, vol. 76, May.
Krishna, R. (1963), ‘Farm supply response in India and Pakistan: a
case study of the Punjab regions’, Economic Journal, vol. 73.
Lipton, M. (1967), ‘Should reasonable farmers respond to price
changes?’, Modern Asian Studies, vol. 1.
Managhas, M., Recto, A.E. and Ruttan, V.W. (1966), ‘Price and
market relationships for rice and corn in the Philippines’, J. Farm
Econ., vol. 48, August.
Mathur, P.N. and Ezekiel, H. (1961), ‘Marketed surplus of food and
price fluctuations in a developing economy’, Kyklos, vol. 14.
Merrill, W.C. (1967), ‘Setting the price of Peruvian rice’, J. of Farm
Econ. vol. 49, February.
Mubyarto, S. (1965), ‘The elasticity of marketable surplus of rice in
Indonesia: a study of Java-Madura’, PhD thesis, Iowa.
Nerlove, M. (1958), The Dynamics of Supply Estimation of Farmers’
Response to Price, Johns Flopkins University Press, Baltimore.
Nowshirvani, V. (1968), ‘Agricultural supply in India. Some theoreti¬
cal and empirical studies’, PhD thesis, MIT.
Olayide, S.O. (1968), ‘Some estimates of supply and demand
elasticity for selected commodities in Nigeria’s foreign trade’, J. Bus.
and Soc. Studies, vol. 1, September.
216 AGRICULTURE AND ECONOMIC DEVELOPMENT

Owen, W.F. (1966), ‘The Double developmental squeeze on agricul¬

ture’, The American Review, vol. 56, No. 1, 43-70, March. ?
Parikh A. (1971), ‘Farm supply response: a distributed log analysis ,
Oxford Inst. Stats. Bull., vol. 33.
Rajagopalan, V. and Meenakshisundaram, V. (1969), ‘Travails of the
tea industry - An economic appraisal’, Indian J. Agr. Econ. vol. 24.
Saylor, R.G. (1974), ‘Alternative measures of supply elasticities: the
case of Sao Paulo coffee’, Amer. J. Agr. Econ., vol. 56, February.
Schultz, T. (1960), ‘Value of US farm surpluses to underdeveloped
countries’, J. Farm Econ., vol. 42. ,
Swift, J. (1969), ‘An economic stwdy of the Chilean agrarian reform ,
PhD thesis, MIT.
Tomek, W. (1972), ‘Distributed log models of cotton average
response: a further result’, Amer. J. Agr. Econ., vo. 64, February.
Wickens M. and Greenfield, J. (1973), ‘The econometrics of Agricul¬
tural supply: an application to the world coffee market, Rev. Econ.
and Stats., vol. 55, November.
Yotopoulos, P.A. and Nugent, J.B. (1976), Economics of Develop¬
ment, Harper & Row, New York.
 8 Institutional Constraints
on Agricultural
Development and
Remedial Policies

In this chapter we consider the constraints imposed on agricultural

development by defects in the institutions controlling the distribution
of land, access to agricultural capital and credit, and the competitive
structure of agricultural markets. We also consider the motives for
and possible effects of reforming policies in these areas.

1 Inequitable Landownership and Land Reform

In this section we examine systems of agricultural landownership

and land tenure; the meaning of and motives for land reform; the
possible benefits and costs of land reform, including its effects on
farm output and the marketed surplus; the limitations of land
reform; and cooperative farming and tenancy reform as policy
alternatives to individual farm ownership.

1.1 Landownership and tenure systems

Taking a world view, systems of agricultural landownership and

tenure are both diverse and complex. To simplify, we make the broad
distinction between ‘traditional’ and ‘modern’ systems. Under tra¬
ditional systems the ownership of land may be either communal or
private. Under communal or tribal ownership, farmers have indi-

217
218 AGRICULTURE AND ECONOMIC DEVELOPMENT

vidual rights of cultivation, but not necessarily exclusive use of the

land. For example, grazing rights are often held in common. Under
private ownership, the rights of ownership and cultivation may either
be exercised by the same person (the owner-farmer) or separately by
landlords and tenants. In the traditional mode the landlord-tenant
system has generally been feudal in character with tenants paying and
landlords receiving rent in the form of either a share of the crop or
labour services. Traditional systems of commercial ownership still
persist, in much of Africa, for example. Feudal landlord-tenant
systems still prevail in much of Latin America, the Middle East and
Southern Asia.
Under modern systems of land ownership the dichotomy is
capitalism versus socialism. In the capitalist model landownership is
private, with landowners exercising the option of either farming the
land themselves or letting it to a tenant usually paying a fixed cash
rent. In the socialist model the land is owned by the state, although
the responsibility for cultivation is often given to co-operative groups
or ‘collectives’. The collective farmers are required to meet production
norms and delivery quotas set by the state.
This system of classification which is summarised in Figure 8.1,
oversimplifies reality. In the LDCs, many gradations exist between
traditional and modem modes of ownership and tenure, as well as
between pure capitalism and pure socialism.
The use of the term ‘land reform’ most commonly refers to the
redistribution of landownership from traditional and feudal-type
landlords to their previous tenants or wage labourers. But, in
principle, the meaning of land reform can be extended to cover any
socially beneficial change in a country’s system of agricultural
landownership or tenure arrangements. So, for example, under some
circumstances the change from a primitive system of communal or
tribal landownership to advanced communal or even private owner¬
ship might yield a substantial social gain in terms of more intensive
use of the land and a larger marketed food surplus. However, the
remainder of our remarks about land reform relate mainly to the
redistribution of privately owned land from the few to the many.

1.2 Meaning of land reform: redistribution of land and agrarian reform

Even when land reform is restricted to the transfer of private

ownership rights, the term is ambiguous in that it fails to distinguish
between the redistribution of land ownership alone, comprehensive
‘agrarian reform’, and tenancy reform.
 INSTITUTIONAL CONSTRAINTS 219

Figure 8.1: Systems of land ownership and tenure

Because even feudal-type landlords commonly supplement the

cultivation rights granted to their tenants with services such as the
provision of credit or a marketing outlet for farm products, the mere
redistribution of landownership from landlords to tenants is unlikely
to advance agricultural development. Even though a landlord may
have exploited his tenants, the freedom conferred on them by
owning the land is a doubtful gain if they are unaccustomed to
independence and therefore unable to fend for themselves as farmers.
With the landlord gone, his former tenants are likely to need
alternative sources of credit and marketing services, and even advice
on how best to care for crops and livestock. Thus, in addition to
220 AGRICULTURE AND ECONOMIC DEVELOPMENT

transferring the ownership of land from landlords to farmers,

substantial government intervention is often needed to provide the
beneficiaries of redistribution with the back-up of credit, marketing,
extension and other services needed to induce increased farm output,
improved farm incomes and a larger marketed surplus. The term
‘agrarian reform’ is sometimes used to distinguish the mere redistri¬
bution of land from genuine land reform combining redistribution
with a complete ‘package’ of complementary institutional reforms.
Where so-called land reform has failed, with its intended beneficiaries
abandoning the land in large numbers, the reason has frequently been
that the vital difference between land redistribution and agrarian
reform was overlooked.
Land reform is costly. Equity demands that landlords who
surrender their land should receive compensation - though, histori¬
cally, some land reforms have been confiscatory. However, the
amount of compensation may be less than the full market value of the
land, particularly under a socialist regime. Where land is sub-divided
the new, smaller farms may need fences, water supplies and other
equipment. But these are merely the costs of redistributing the land.
Agrarian reform is much more costly still, due to the costs of
installing and operating ancillary services such as credit supply,
marketing facilities and agricultural extension. Agricultural service
costs which were concealed within the landlord-tenant system now
fall directly upon the public Exchequer. It is clear why the successful
implementation of a land reform may depend upon a measure of tax
reform (Lewis Jr, in Southworth and Johnston, 1967).

1.3 Motives for land reform

Although the demand for land reform frequently comes from the
‘grass roots’, its underlying goals are primarily social and political,
rather than economic. Thus, although development economists tend
to see land reform as a possible means of accelerating LDC
development through the economic betterment of peasant farmers,
the farmers themselves may be unaware of this possibility. However,
in societies where the mere ownership of land gives social status to the
holder, the desire to own land can be intense. The possible use of land
as a factor of production is a secondary consideration. A landowner’s
wealth derives from the value of the land, which in turn reflects the
capitalised value of the rent. But, in the context of land reform, the
source of rent payment is the psychic income the owner derives from
merely possessing the land, rather than income from farming. Thus
 INSTITUTIONAL CONSTRAINTS 221

grass roots pressure for land reform comes from the popular demand
for a fairer distribution of wealth rather than for a better farm
income. But the motives for land reform also include the redistri¬
bution of political power.
In feudal societies wealth is concentrated in the hands of large
landlords who also dominate politically. In true traditional societies
the peasants may accept their politically subordinate role as the price
of their landlord s protection, particularly in times of economic
hardship. But, in the course of the transition from traditional to
modern society, landlords may seek to exert their economic power. It
is then that the landlords’ role may come to be seen as exploitative
rather than paternalistic, leading to popular pressure to remove their
political and economic power simultaneously (Dorner, 1972, ch. 3).
Evidence that the distribution of agricultural land in some LDCs is
in fact heavily skewed in favour of the rich and powerful has been
presented in chepter 2, section 2. The evidence is accompanied by a
discussion of how unequal access to land exacerbates rural poverty.
Despite our argument that the conscious motive for land reform
amongst the people is primarily social and political, a large part of the
objective economic case for it must be income redistribution in favour
of the rural poor. But the actual success of land reform in this respect
is uncertain depending on factors discussed in the next section.

1.4 Benefits and costs of land reform

An idealised view of the benefits of land reform may be taken. The

argument runs that as a consequence of transferring landownership
rights from landlords to their former tenants, newly-independent
farmers have both the resources and the incentive to accumulate
capital. The resources derive from the rent they no longer have to pay;
complete security of tenure provides the incentive. Apart from
taxation, the whole agricultural surplus (output minus personal
consumption) now accrues to the cultivator who cannot be dis¬
possessed. Aggregating from the individual farm to the level of the
agricultural sector as a whole, the benefits of land reform are
supposed to include both an enlarged domestic agricultural output
and a bigger marketed surplus of food and other agricultural
products. Thus food shortage is averted, food prices remain relatively
low to the benefit of nascent industry and urban consumers, and
pressure on the balance of the overseas payments is relieved by a
declining food import bill. It is also claimed that land reform, which
222 AGRICULTURE AND ECONOMIC DEVELOPMENT

breaks down large farms or estates into smaller-sized agricultural

holdings, promotes greater efficiency in the utilisation of agricultural
resources, particularly land. We referred in chapter 6, section 1.2 to
the evidence of Berry and Cline (1979) in support of the hypothesis
that an inverse relationship exists between farm size and crop yields
(or the productivity of land), at least in traditional agriculture. The
same authors infer from this evidence that, assuming that the
productivity of new small farms (created by land reform) is the same
as existing farms of the same size, the twin goals of greater equity and
higher productivity can be reached simultaneously by sub-dividing
large farms into smaller ownership units virtually without limit
(Berry and Cline, 1979, ch. 2). The only exceptions to this rather
startling conclusion are differences in land quality and ‘economies of
scale in a few products’. Our own view is that in drawing this
inference Berry and Cline overlook some rather obvious dis¬
economies of ‘minifundia’ which are too small to yield an adequate
livelihood to an occupier.
These are some of the claimed benefits of land reform, but how do
they compare with reality? Whether the immediate beneficiaries of
land redistribution will in fact save and invest in agriculture is a moot
point depending upon numerous economic and social factors. We
have stressed how in peasant agriculture technical innovation is
obstructed by economic uncertainty and lack of knowledge, as well as
by ingrained social attitudes that are inimical to change (ch. 2,
section 1.3). In this chapter we have emphasised that those to whom
land is transferred do not necessarily see the reform as opening the
way to augmenting their income from the land: rather they may view
mere possession of the land as the ultimate goal. Moreover, newly-
independent farmers may lack the experience to farm successfully on
their own, particularly if ancillary services previously supplied by the
landlords are not available from another source.
Without higher output at the level of the individual farm there can
be no spread of land reform benefits from agriculture itself to the
larger economy. But even if farmers do increase their output, they
may increase their own consumption by an equal amount. Although
such increased consumption may count as a gain in welfare, there is
no increase in the marketable surplus. Indeed, as explained more fully
in chapter 3, section 4, where we summarised Griffin’s analysis of
marketable surplus transfer mechanisms, the abolition of rent
payments under land reform could even reduce the size of the
marketable surplus. Were this to occur, the additional social costs of
 INSTITUTIONAL CONSTRAINTS 223

land reform, in terms of higher urban food prices, higher manufactur¬

ing costs and the balance of payments cost of larger food imports
could be considerable.
In breaking with the past so abruptly land reform is economically
and socially disruptive. Time is needed to establish a new structure of
agricultural production with supporting ancillary services such as
credit, marketing and agricultural extension. Thus, in the short
term, there is almost bound to be some decline in agricultural
production as an aftermath to land reform. This decline may be
especially marked where reform involves switching from feudal-type,
large-scale agriculture to small-scale family farming. So, for example,
the break-up of Latin American hacienda, with the land distributed
amongst the former peons, contrasts with merely transferring the
ownership of existing small-scale farms in South-East Asia from
landlords to their former share-cropper tenants. Ceteris paribus, the
expected disruption of production would tend to be much greater in
the former situation than in the latter (Dorner, 1972, ch. 5).
The empirical evidence of the net benefits of land reform is mixed.
The record indicates a variety of results ranging from successful to
unsuccessful reforms with a number in between for which the verdict
is indecisive. The group of countries with successful reforms includes
Egypt, Taiwan, South Korea and, probably, Iran. In all these
countries an effective redistribution of landownership was combined
with improved agricultural productivity and higher output. The
unsuccessful group may be considered to include Mexico, where the
ejidos have lagged behind private farms in productivity and output
growth, and Peru, where the economic viability of small-scale farms
has been difficult to establish, despite the formation of co-operatives.
This group also includes Boliva and Iraq, where the output decline
following reform was both dramatic and persistent. The indecisive
group includes Japan, where a successful first reform was followed by
a second reform which worsened the problem of under-sized farms,
and India, where the objectives of land reform legislation have tended
to be defeated by incomplete implementation and evasion (World
Bank, 1975, pp. 252-62: Southworth and Johnston, 1967
pp. 285-93).
The case of Mexico, where land reform was only partial, so that
reformed and unreformed sub-sectors of agriculture continued to
operate side by side, is especially interesting. The results of a recent
comparative study of land reform (ejidat) and private farms across 32
Mexican states revealed a substantial discrepancy in average pro-
224 AGRICULTURE AND ECONOMIC DEVELOPMENT

ductivity of labour (and farm income) in favour of the non-reformed

farms. However, this result was attributed not to less efficient resource
utilisation on the land reform farms but to their unequal access to
land and capital. A larger share of these scarce resources for the land
reform sector was the policy implication (Nguyen and Martinez-
Saldivar, 1979).
Even with relatively successful land reforms the success has been
qualified by unresolved problems. One such problem is that the larger
and more successful of pre-reform farmers have tended to gain a
disproportionate share of the post-reform benefits. Another is that
landless labourers, with no access to land before reform, have
frequently gained little or nothing.

1.5 Limitations of land reform

If it is true that agricultural resource efficiency is inversely correlated

with farm size, then the redistribution of land appears to achieve the
twin goals of greater equity and improved efficiency. However, in
some countries the density of the rural population is so high that the
ideal of providing every rural household with enough land to yield an
adequate livelihood from agriculture alone is unattainable. There is
simply not enough land to go round amongst all those with a claim
based on equity.
Some countries have sought to impose farm size ‘ceilings’ in an
endeavour to release more land for redistribution to families with
only very small areas of land or none. But in India it was found that
even with the ceiling as low as 20 acres, the amount of land released
would be sufficient only to bring the minimum size of holding amongst
all existing farm occupiers up- to 5 acres. Landless households,
representing more than 20 per cent of all rural households (see
chapter 2, section 2) would remain landless. The same dilemma has
been observed in other LDCs, especially in South East Asia (World
Bank, 1975, p. 219). Under such conditions land reform alone cannot
cure rural poverty or eliminate unemployment. Indeed, it is clear that
in large part a solution to these problems must come from the
acceleration of employment opportunities created outside the agricul¬
tural sector.

1.6 Alternatives to ‘land to the tiller’

Although the use of the term land reform is sometimes restricted to

the granting of individual ownership rights to former agricultural
 INSTITUTIONAL CONSTRAINTS 225

tenants, there are alternative ‘reformist’ approaches to raising

agricultural output and incomes. We refer briefly to two of these.

1.6.1 Co-operative Farming

As an alternative to breaking up large farms or estates into family¬
sized farms, large farms may be transferred intact to co-operative
groups of producers. Compared with small-scale individual owner¬
ship, co-operative farming spreads the scarce factors of skilled
management and capital over a larger output and facilitates group
decisions on indivisible factors such as irrigation. It also reduces the
risk that, following an equalising land reform under private owner¬
ship, marked disparities of income and wealth within the agricultural
sector will eventually re-emerge as, due to greater enterprise or good
fortune, some farmers succeed in bidding resources away from other
farmers. Readers are referred to Putterman (1983) for a more
extended discussion of the issue of ‘restratification’. A possible
drawback of co-operation is a loss of individual incentive to greater
enterprise and effort, but this depends on how individual rewards
within the group are organised.
The choice between individual and co-operative farm ownership
may be influenced by ideology, as well as by economics, but even if
eastern-style collective farms are excluded, there have been co¬
operative farming experiments in many parts of the world. The
record of achievement has been varied with some successes and many
failures. Some of the most successful examples have been concerned
with ‘land settlement’, rather than with the reform of an existing farm
structure. Settlement schemes in Israel {kibbutzim) are notable in this
respect. These have been widely used as a model in other countries,
but only rarely with anything like the same degree of success as in
Israel (Southworth and Johnston, 1967, pp. 302-3; Dorner, 1972,
pp. 54-62).
The case for the co-operative farming alternative may be strength¬
ening with time due to changing agricultural technology. In discuss¬
ing the relationship between farm size and efficiency in an LDC
context, we earlier expressed reservations concerning the claim that
HYV technology is neutral to scale (chapter 6, section 1.2). Much of
the economic rationale of co-operative farming derives from the
opportunities afforded to members to exploit economies of large-
scale production. Thus, ceteris paribus, co-operative farming is
favoured by the emergence of new agricultural technology with a
potential for yielding economies of scale, such as field mechanisation
for example.
226 AGRICULTURE AND ECONOMIC DEVELOPMENT

1.6.2 Tenancy Reform

Some advantages of land reform may be attainable without actually
changing the ownership of the land. It may be possible to improve
the economic status and productivity of farm tenants through land
tenure reform. An efficient landlord-tenant system precludes gross
inequality of bargaining power between the parties. Yet, in practice,
farm tenants in LDCs frequently lack legal rights. Thus, a first step
in tenancy reform may be the enactment of legislation to codify and
limit the powers of landlords. Such a code might limit the unilateral
raising of rent, improve security of tenure and provide for independ¬
ent arbitration to settle disputes. A second and complementary step
would be to oblige landlords and tenants to enter into legally binding
tenancy agreements consistent with the code of rights. If properly
enforced, such measures would revolutionise feudal tenure systems
by compelling landlords to let land on fair and equitable terms. But
there must be enough political will to ensure both the enactment of
legislation and the oversight of its enforcement.
A good landlord-tenant system possesses the virtue of functional
specialisation. Landowners concentrate on estate management,
maintenance and administration, leaving tenants free to concentrate
on actually farming the land. As previously mentioned, landowners
may also provide ancillary services, such as production credit, a
marketing outlet, and even farming advice. Share-tenancy is a device
for spreading the risks of production between tenant farmers and
their landlords: without such a system farmers would have to bear the
whole of the risk themselves. Contrary to much popular supposition,
the principle of share-tenancy is not socially inefficient; but it may be
socially oppressive in practice due to the abuse of monopoly power by
landlords (Griffin, 1976, pp. 122-3).
Although tenancy reform avoids the destruction of landlords’
ancillary services, complementary institutional reforms may be
needed to supplement or improve those services. By providing
farmers with another source of credit or an alternative market
outlet, government can help to ensure that landlords offer their
services at competitive prices.
A complete land reform can be very expensive due to the high
cost of compensating landowners for the loss of their property
rights. The comparative cheapness of tenancy reform is one of its
major advantages. The major drawback of tenancy reform is
political rather than economic, namely, that it does not alter the
structure of political power (Mellor, 1966, pp. 260-1: World Bank
1975, pp. 222-3).
 INSTITUTIONAL CONSTRAINTS 227

1.7 Summary and policy conclusion

(1) The motives for land reform, as seen by its beneficiaries, derive
primarly from the social and political aspirations of the rural
population. Farmers may have to be educated to see land reform
as a means of improving their productivity and incomes.
(2) The agricultural benefits of land reform depend on much more
than the mere redistribution of land ownership. A complete
agrarian reform is needed entailing redistribution plus a complete
package of supporting ancillary services.
(3) Although the possible economic benefits of land reform include a
larger agricultural output and higher farm income, together with
a great marketed surplus of agricultural products, the real¬
isation of these benefits in practice depends upon numerous
factors, including the form and content of government policies.
Due to its disruptive character, land reform is virtually bound to
result in some short-term loss of agricultural output.
(4) The empirical evidence on the agricultural benefits of land reform
reveals a mixture of successes and failures. A general failing of
virtually all reforms is that very few of the benefits have reached
very small-scale farmers and landless labourers.
(5) In very densely populated countries the agricultural area released
by land reform may be insufficient to provide adequate-sized
farms for all rural families; that is, rural unemployment and
poverty cannot be eliminated by land reform alone.
(6) Co-operative farming and tenancy reform offer alternative and
possibly cheaper reformist approaches to raising agricultural
output and incomes than conventional land reform. But these
alternatives fail to satisfy other goals such as the redistribution of
private wealth and the re-alignment of political power.
(7) Strong government direction and leadership are needed to enact
land reform legislation, ensure its implementation, and provide
its beneficiaries with adequate incentives and ancillary services
to induce higher investment, output and farm income as well
as a larger marketed surplus of agricultural products.

2 Capital and Finance in Underdeveloped Agriculture

The problem of capital scarcity in underdeveloped agriculture is

generally well known. The role of financial and credit institutions in
promoting capital accumulation in this context has received con-
228 AGRICULTURE AND ECONOMIC DEVELOPMENT

siderable attention in many countries in recent times. It is important

to understand the nature and composition of rural monetary
institutions in LDCs to formulate appropriate monetary and credit
policies to act as catalysts to promote the level of economic
development. In this section we shall first discuss the nature of the
credit problem in underdeveloped agriculture. Next, an attempt will
be made to analyse the demand and the supply side of rural money
markets. Then, we shall describe the structure of rural interest rates in
some LDCs. It is generally believed that such interest rates are very
high and sometimes approach the level of usury. Merchants-cum-
moneylenders are sometimes considered as the main agents of
exploitation through usury. In extreme cases, it has been argued
that an ‘exploitative’ mode of production created through usury,
prevents the growth of the agricultural sector as a whole in LDCs.
(See, Bhaduri, 1973, 1977.) Later, we shall discuss such views and
their criticisms. At this stage, it may be useful to note some general
features of rural money markets.

2.1 Functions of rural money markets in LDCs

A money market (rather than a capital market) usually caters for the
demand for an supply of short-run loanable funds. We may note the
following major functions of rural money markets:

(a) A major function of a money market is to allocate savings into

investment and promote a more rational allocation of resources.
(b) It is generally well acknowledged that the growth of an agrarian
economy could be significantly retarded due to a lack of savings
and investment. An efficient money market raises savings and
investment by promoting liquidity and ensuring the safety of
financial assets. This function is of crucial importance in LDCs
where savings and investment behaviour leave considerable
room for improvement. It is well known that in the rural
economies of most LDCs savings too often take place in the form
of gold-hoarding and land-holding rather than in the holding of
financial assets. Hence, notwithstanding the ability to save, the
lack of an efficient and developed money market deprives the
society of financial assets which could lead savings into fruitful
investment.
(c) A money market generally promotes financial mobility as funds
could be transferred from one sector to the other in a manner
which economises transaction costs. Such economies of scale are
 INSTITUTIONAL CONSTRAINTS 229

considerable when many people are involved in such transactions

with an overall growth in the flow of total funds. It is noteworthy
that an efficient money market is crucial for providing elasticity in
the flow of funds.
(d) An efficient rural money market is essential for implementing the
monetary policies of the central bank. Clearly, monetary policies
are unlikely to be very successful if a large part of a predomi¬
nantly agrarian economy works independently of the monetary
transactions. This may sometimes happen in a barter economy
where goods are usually exchanged against goods, and money is
no longer a medium of exchange. Under such conditions, changes
in money supply are unlikely to alter substantially the aggregate
demand in the society and monetary policies will be largely
ineffective.
(e) Rural money markets in LDCs are seldom homogeneous.
Broadly, they can be divided into two major parts: organised, and
unorganised. The organised section usually comprises the central
bank, the commercial banks, co-operative banks, credit societies.
They usually operate within the provisions of the Banking
Companies Act of different LDCs. The organised sector gener¬
ally maintains accounts which are open to audit and periodic
inspection.

On the other hand, the unorganised sector operates outside the

legal framework. They maintain sorts of accounts which are not
always very sophisticated. These accounts are not open to inspection.
Indeed, a great deal of secrecy covers the financial operations of the
unorganised financial sector. It generally consists of moneylenders,
landlords, merchants, traders, indigenous bankers, pawnbrokers,
and even friends and relatives. Blending of moneylending with other
types of economic activities is a special feature of such an unorganised
sector. The other features include informality in dealings with
customers, personal contact with borrowers, simple systems of
maintaing accounts, flexibility of loan operations and secrecy about
financial transactions.
The moneylenders are important sources of funds in the agricul¬
tural sectors of many LDCs and it is important to discuss their
operations at some length in a village economy. Usually they have a
very good knowledge of the character and repaying capacity of the
borrower. They are quite flexible in their operations, but grant loans
largely against personal security. Sometimes they also offer consump¬
tion loans. In case where moneylenders are also the landlords, there is
230 AGRICULTURE AND ECONOMIC DEVELOPMENT

a tendency to ‘over price’ the loan at the time of lending and

‘underprice’ them at the time of repayment. This complex mechanism
could be understood more clearly in the light of what has been called
the ‘characteristics of a semi-feudal agriculture’ (Bhaduri, 1973,
1977). Such characteristics can be summarised as follows:
Share cropping. This is a system under which the landowner leases
out his land for at least one full production cycle and the net harvest
(i.e. gross harvest - seed required for the next harvest) is shared
between the landlord and the tenant on some mutually agreed, legal
basis. However, in many cases, tenancy rights are not legally very well
protected and terms of contract are quite complicated.
Perpetual indebtedness. Quite often it has been witnessed that tenants
are heavily indebted, particularly to their landlords. Repayment of
debt along with interest payments by tenants generally signifies a
large fall in the available harvest and the tenant is compelled to
borrow for consumption for his survival. Because of the presence of
such consumption loans, tenants remain perpetually in debt.
Landlords as the moneylander. The problem is tenants are aggravated
by another characteristic of a semi-feudal agriculture - the absence
of a complete specialisation. Landlords frequently act as creditors to
their tenants and consumption loans are usually advanced by the
landlords at high rates of interest. Hence the tenant leases his land
from the same person to whom he is perpetually indebted and this
reduces him to the abject status of a traditional serf. Hence, the semi-
feudal landlord exploits the tenant both through usury and through
his property rights on land.
Inaccessibility to the market. It has sometimes been claimed that the
tenant has no access to the organised money market. Under such
circumstances, it follows that the only creditor of the tenant is his
landlord who assumes the power of a pure monopolist. He lends
against a future harvest at a high rate of interest and the tenant has
very little option open to him but to accept such rates. It is also argued
that the tenant has little access to the commodity market, and as such
he cannot sell at the highest price. Generally, he has to sell when
harvest prices are at their lowest just after harvesting to meet the
scheduled date of repayment. On the other hand, when prices are at
their highest, tenants could be left with very little food for themselves
and they frequently need consumption loans just to survive.
Hence, lack of access to money and commodity markets is
supposed to be the reason for usurious interest rates in LDCs. In the
extreme, the ‘mode of production’ can be totally exploitative as the
 INSTITUTIONAL CONSTRAINTS 231

landlord-moneylender as the single buyer of the product and as a single

lender in the money market squeezes out the whole of monopolistic
profit.
It must be admitted that although such a view of the determination
of the rural interest rate and the mode of exploitation is interesting,
the empirical evidence available so far has not substantiated the
‘interlocking of factors’ theory (see Bardhan and Rudra, 1978;
Bardhan, 1980). More specifically, the hypothesis that tenants and
agriculturists have no other option but to sell to a single landlord
who is also the only moneylender has not always been upheld in
practice. It is possible to note cases of usurious rural interest rates in
an exploitative situation where the monopolist-landlord-moneylender
does successfully extract monopoly gains. On the other hand, as a
general rule, some find it difficult to accept, particularly in view of the
evidende available so far. Hence, it may be useful to explain the
economic, as opposed to the institutionalist, argument to account for
high rural interest rates.

2.2 An economic theory of rural interest rate determination

From the supply side, the rural interest rate (= r) may be regarded as
the sum of (a) administrative cost (= a), (b) risk premium (= p) (due
to the probability of default in repayment), (c) opportunity cost of
lending (= y), and (d) the monopoly profit (= n) (see Bottomley,
1971). Hence, we can write:

r = a + /? + y + 7t

The administrative cost (= a) is not supposed to be very high as the

moneylenders, as explained above, do not incur high administrative
costs to run their moneylending activities. The opportunity cost of
lending is also very small as the organised financial institutions offer a
very low rate of interest on savings - a system which has been
described by some as one of financial repression (see McKinnon,
1973; and Shaw, 1973). Furthermore, interest rates advanced by the
rural financial intermediaries on deposits (e.g. Post Office savings
banks, primary credit co-operatives, rual banks) hardly change. All the
interest, then, centres on the two other variables, P and n . For the
purpose of simplification, we can then write:

r= i(P, n)
232 AGRICULTURE AND ECONOMIC DEVELOPMENT

It is generally argued that the higher the risk premium (which can
be approximated by the probability of default), the higher will be the
rural interest rate. The relationship between the monopoly profit
and r is also positive. However, if the real income and output in the
agricultural sector rises, then this will reduce the probability of default
and hence the risk premium will also be reduced; which will also
account for a fall in r. A fall in rural interest rate will weaken the
‘monopoly power’ of moneylenders in the rural economy too.
In sum, an increase in the level of output and real income in the
agricultural sector will increase the repayment capacity of the
farmers. This would clearly increase their credit worthiness by
reducing risk-premium which in its turn will lower the high level of
rural interest rates that prevail in LDCs.
Institutional factors, like the growth of the credit co-operatives and
the rural banks, also pay an important role in influencing the rural
rates. Usually, these credit co-operatives and the rural banks charge
interest rates which are much lower than those charged by the
moneylenders. Despite such a differential in the lending rates, it is
interesting to point out that the credit co-operatives and the rural
banks failed to make a substantial progress in lending and saving
activities of the village economies of most LDCs. Such a failure may
suggest that the demand for credit is not very sensitive to a low rate of
interest. It may also suggest that the pattern of loan administration
that has usually been followed by the organised societies in the rural
areas is fairly complicated for rural people. Banks, in many instances,
follow the orthodox principles of lending and, given the general
impoverishment of agriculture in LDCs and the low level of average
real income, it is obvious that farmers are not generally considered as
very credit worthy customers. Formal business practices by the
organised lending institutions, lack of an adequate number of rural
banks and the amount of transport costs are some of the other factors
which could explain the lack of progress of organised finance.
Sometimes farmers in LDCs need consumption rather than produc¬
tion loans, so it may be useful for the organised lending agencies to
open up the possibility of financing such consumption loans. Indeed,
in many instances, it is difficult to distinguish between consumption
and production loans in under developed countries. The traditional
idea that all forms of consumption loans are unproductive may not be
always valid. Mirrlees (1975) has shown how an increase in consump¬
tion in LDCs could add to the efficiency and productivity of labour
which, in its turn, has a salutory effect on economic growth and
development.
 INSTITUTIONAL CONSTRAINTS 233

Some have argued (e.g. Bhaduri, 1973) that farmers in LDCs are
prevented by their landlords from adopting modern farming methods
which might have increased their income since such a rise will increase
repayments of borrowing by farmers and release them from the
clutches of the landlords-moneylenders. To maximise the income
from usury, it is in the landlords’ interests to keep farmers-tenants
heavily indebted for as long as possible. As a result, in the agricultural
sector of most LDCs, technical progress has not been introduced; and
even when it has been introduced its scale has been very limited. A
semi-feudal agriculture is once again perpetuated with usury and
indebtedness as the chief instruments of exploitation by the land¬
owning class.
On the other hand, given the complete dominance of the landlords
over the poor farmers and given the principle of profit and income
maximisation, it may be argued that the landlord can maximise his
income by ‘expropriating’ the gains from increased productivity
flowing from the adoption (rather than non-adoption) of new
technology without resorting to usury. It may be argued that
investment in the land will be prevented by the landlords since
gains through usury are larger than those from increased gains in
productivity. But such an argument is quite unsatisfactory
because if landlords have enough control to prevent innovation,
they should also have sufficient power to obtain from the peasants
additional gains from increasing productivity via technical progress.
In any case, usury as a mode of expropriating the ‘surplus’
could be unneccessary in labour surplus economies where the
bargaining strength of the poor peasants and tenants is very weak. It
should also be mentioned that the empirical evidence available so
far does not support the strong inter-linkage between the moneylen¬
ders, merchants and landlords (i.e. between the product and the
credit market) as suggested by the institutional school (see Rahman,
1979, Bliss and Stern, 1981, Ghatak, 1983). On balance the evidence,
albeit sketchy, seems to suggest that rural interest rates are likely to be
lower the higher the growth rates of agricultural output and
repayment, the lower the risk premium and the probability of default
in loan repayment, and the greater the spread of the co-operatives.
High rates of rural interest rates are also significantly correlated with
low per capita rural income. It is certainly true to say that a near-
subsistance agrarian economy only helps the moneylenders or
loan-sharks to strengthen their strange-hold over the rural economy.
It could also enable them to obtain usury. However, the real
‘explanation’ of the rural interest rates is quite a complex issue
234 AGRICULTURE AND ECONOMIC DEVELOPMENT

since, income apart, institutional factors like the social and legal
relationships between debtors and lenders and the nature of the
contract between them could play a major role. But, on the whole, it
seems fairly likely that a rise in agricultural output and rural real
income should reduce the risk premium and uncertainty in lending,
reduce the rural interest rates and contribute substantially towards
weakening the monopoly power of the moneylenders.

Rural interest rate

Repayment

Figure 8.2: A composite view of rural interest rate determination

 INSTITUTIONAL CONSTRAINTS 235

2.3 A simple geometric representation of the composite view

of rural interest rate determination

It may be useful to describe a synthetic view of the rural interest

rate determination by using a simple diagram (see Figure 8.3). On the
vertical axis, we measure the interest rate and the level of income
on the horizontal axis in the first quadrant. The relationship
between r and Y is assumed to be inverse for the reasons
already stated. In the second quadrant, the positive relationship be¬
tween the monopoly power of the moneylenders (= n) and the rural
interest rate is described. In the third quadrant, the association
between repayments and the power of the moneylender is regarded
as inverse because the higher the repayment, the lower will be the
power of the moneylender to compel the farmers to pay high interest
rates on grounds of high risks of default. In the fourth quadrant a
positive relationship between real income and repayment is shown
because the higher the level of income, the greater will be the
repayment of the farmers. Thus we have;

r = f(Y, R, n)

and fj <0, f2 < 0, f3 > 0.

It is then easy to work out that a low income level like OYx will
generate low repayment (=OR1), a strong stranglehold of the
moneylenders over the farmers (= Onx) and a high rural interest rate
(= OrJ. A high rate of growth in the agricultural sector will raise the
real income to OY2. This, in its turn, will increase repayments to OR2,
reduce the degree of risk and the probability of default, and reduce the
monopoly power of the moneylenders to On2, which will then reduce
rural interest rate to Or2.

2.4 Financial policies for agricultural development

At the policy level, it is thus necessary to raise the rate of agricultural

growth and real income within the agricultural sector to reduce the
usurious rural interest rate. It is also necessary to attack the problem
from the institutional standpoint. In many cases, a complex set of
socioeconomic and legal frameworks help to perpetuate an age-old
system of land tenure which generates an unequal access to resources
including the availability of credit. It has now been confirmed that the
landlords in LDCs who own large plots of land and earn higher income,
pay (relatively) lower interest rates in comparison with the small
236 AGRICULTURE AND ECONOMIC DEVELOPMENT

landowners. It is also known that a very high proportion of credits

from the rural banks and co-operative usually gravitates towards the
large landowners as they have more sound collateral (generally land
and equipment) to be very credit-worthy clients for the organised
financial agencies. Clearly, small farmers, tenants and landless
labourers are at disadvantage to be considered credit-worthy cus¬
tomers. Here, a more egalitarian distribution of land through a series
of land reform programmes should improve the economic standing
of the poorer section of the peasantry in the eyes of the organised
financial institutions who will then be more willing to lend to this
particular group of the rural sector. A more equal distribution of
credit by the organised credit agencies will also generate more
competition for the loan-sharks and moneylenders, who will then be
forced to charge lower interest rates since the organised financial
institutions which generally operate in the rural sector of the LDCs
charge rates of interest which are much lower than those charged by
the moneylenders. Where political will exists, it is extremely import¬
ant to carry out such land reform programmes whenever the situation
permits.
The other policy that may be pursued by the monetary authorities
to mobilise financial savings from the rural sector is to raise the
deposit interest rate offered by the organised financial institutions. It
is well known that the rural banks and the credit co-operatives in most
LDCs usually offer very low interest rates to the savers. As such,
incentives to save in organised financial agencies are lacking. Given
the shortage of credit supply in many areas, one would expect the
equilibrium interest rate for a rational allocation of resources will be
much higher. The system of keeping interest rate very low by the
organised financial agencies has been called a regime of ‘finanacial
repression’. The following diagram illustrates (see Figure 8.3). Inter¬
est rate in rural areas is measured in the vertical axis and savings and
investment are measured on the horizontal axis. When the interest
rate is too low, say Orl5 savings are OS! and investment demand will
be equal to O^ and a gap of excess demand of SJj will emerge. If
the interest rate is raised to Or2 savings will rise to OS2 and the
excess demand will be narrowed to S2I2. A more rational allocation
of resources will require the interest rate to be raised to Ore. A rise in
interest rate from r2 to r2 and a fall in investment simply irons out
inefficient allocation of resources as shown by the shaded area.
On the basis of the available evidence, it has been observed that
a policy of very low interest rate in many LDCs has led to a steady
attrition of organised finance (Chandavarkar, 1971). A policy of
financial reform in such LDCs should then include the programme of
 INSTITUTIONAL CONSTRAINTS 237

Figure 8.3: Efficacy of a positive rural interest rate policy

raising rural interest rates to an equilibrium level (or around it) to

attract more savings for investment in agriculture. It is, however,
assumed that .savings will rise with a rise in interest rates. There is
some evidence to confirm that such an assumption is plausible in
some LDCs (Gupta, 1970). In any case, the policy of maintaining a
very low rate of interest in underdeveloped rural areas has courted
such a dismal failure that perhaps it is now necessary to follow a
policy of financial liberalisation.

3 Market Imperfections and Marketing Policy

In this section we examine the role of marketing in agricultural
development and the objectives of government marketing and price
238 AGRICULTURE AND ECONOMIC DEVELOPMENT

policy, the imperfections of agricultural marketing in LDCs, market¬

ing reform policies, price policy and price stabilisation mechanisms,
and the role of the state in agricultural marketing.

3.1 Role of marketing in agricultural development and the

objectives of government marketing and price policy

Agrarian societies are characterised by a high degree of self-

sufficiency at the village level and even at the level of individual
households. Hence, at this stage of development, the demand for
commercial marketing services, such as the transport, processing and
strorage of food, is extremely limited. But as a consequence of
economic growth and development, as signified by urbanisation,
rising living standards and an increasing demand for services of all
kinds, the volume of economic resources devoted to agricultural
marketing inevitably grows.
Marketing and production are interdependent. Producers must be
convinced that a remunerative market exists for their products,
particularly ‘new’ products, before they can be induced to produce
commercially. An attractive market prospect combines a ‘good’ price
with an assured sales outlet. This simple, self-evident truth has
sometimes been overlooked in agricultural development planning.
Marketing projects have failed both because an adequate price
incentive was not backed up by a neccessary addition to the market
infrastructure (e.g. processing facilities) and because infrastructure
was provided without a sufficiently attractive price.
Government policy for agricultural marketing may be seen to have
four principal objectives: first, to encourage (and occasionally limit)
production to meet growing consumer demand; second, to distribute
non-farm inputs to agricultural producers; third, to protect con¬
sumers from unscrupulous trading practices and unduly high prices;
and fourth, to promote ‘efficient’ marketing.
The types of policy adopted to realise these objectives vary widely in
different countries according to their stage of economic development
and their government’s ideological stance between pure capitalism
and pure socialism. A broad choice exists between regulation and
government participation. Where government confines itself to the
regulation of markets, trading remains in the hands of private
enterprise or producer organisations - government merely prescribes
fair trading practices’, such as the use of standard weights and
measures, the prescription of grade standards and the prevention of
fraudulent practices (and possibly also the regulation of prices).
 INSTITUTIONAL CONSTRAINTS 239

Where government also participates in marketing it performs trading

activities either in competition with private enterprise, or as the sole
(i.e. monopoly) buyer/seller of a particular commodity or commod¬
ities. A primary but insufficient reason for government participation
in trading is the desire to control prices. If the government wishes
to regulate prices it can do so without exercising trading powers.
Price policy is an important sub-set of government marketing
policy. One objective of price policy is to regulate the agricultural
sector’s internal terms of trade so as to maximise the overall rate of
economic growth. But the terms of trade objective may be com¬
plemented by a price stabilisation objective. The economic rationale
of price stabilisation is that, where producers are risk-averse, it may
promote more efficient resource allocation. Allocation or pricing
efficiency requires that the marketing system be responsive
both to changes in consumer demand and changes in the conditions
of supply, such as factor cost and technological choice. But the full,
economic efficiency of a marketing system demands technical as well as
pricing efficiency.
The realisation of technical efficiency in marketing is contingent
upon goods moving from producers to consumers at minimum cost,
subject to the proviso that all marketing services for which an
effective demand exists are actually catered for. Thus an improve¬
ment in technical efficiency does not necessarily mean a reduction in
the marketing margin; rather, it may involved an improvement in
marketing services such that the marginal utility of the extra service
(to market users) exceeds the marginal cost of providing it. But if, due
to market imperfections, the marginal private and marginal social
costs or benefits of a market innovation diverge, the social benefit:
cost ratio is the appropriate efficiency criterion. So, for example, in
a country with serious unemployment or underemployment in the
agricultural sector, although a policy of reducing the number of
marketing intermediaries might be effective in reducing the direct
costs of marketing, if the policy also increases unemployment the
consequent added social cost may outweigh the incremental benefits
to producers and/or consumers of a narrower marketing margin.

3.2 Imperfections of agricultural marketing in LDCs

We have selected three major weaknesses of agricultural marketing in

LDCs for brief discussion: (i) infrastructural deficiencies; (ii) the
weak bargaining position of producers; and (iii) producers’ lack of
information.
240 AGRICULTURE AND ECONOMIC DEVELOPMENT

3.2.1 Deficient Infrastructure

Although a country’s general communications infrastructure - its
roads, railways, ports, postal services, telecommunications, etc. -
serves the economy as a whole, and not just agriculture, the adequacy
of these services vitally affects the structure and costs of agricultural
marketing. Adequate transport is especially crucial. Without rail¬
ways or roads suitable for the use of motor vehicles, the cost of
long-distance transport becomes prohibitive, thus greatly limiting
the size of the market. The development of agricultural marketing
may also entail the provision of a more specialised infrastructure.
Building are needed for the assembly, buying and selling, and storage
of produce and even more specialised facilities, such as slaughter¬
houses and cold stores, may also be required.
Progress in agricultural marketing is typically constrained by low
standards of infrastructural provision in both these areas, the general
and the specialised. The problem persists due to underinvestment.
The provision of social overhead capital is generally unattractive to
private investors who are also deterred from investing in the
provision of more specialised infrastructural services for agriculture
by what are considered to be unacceptably high risks.

3.2.2 Weak Bargaining Position of Producers

The competitive structure of agricultural markets is invariably such
that the number of market intermediaries is greatly exceeded by the
number of individual producers. Thus the bargaining position of
individual producers tends to be weak. But degrees of ‘effective
competition’ amongst traders buying from farmers varies greatly
between countries and even within them. Development economists
hold widely divergent views on the position of the competitive norm
in LDC agricultural markets between the extremes of pure competi¬
tive and pure monopsony, possibly according to where they have
gained their field experience. Some take the view that, if not pure
competition, then ‘effective’ or ‘workable’ competition is the norm,
and oligopsony (and even more so, monopsony) the exception in
characterising the behaviour of first-hand buyers of agricultural
products in LDCs. In other words, according to this view, farmers
generally have a real choice of market outlets so that buyers are
obliged to compete, by offering a better price or an improved service
in order to procure supplies (see e.g. Bauer and Yamey, 1957,
pp. 186-7). The opposing view is that, at least in some countries,
buying power is so concentrated and collusion amongst buyers so
great that sellers do not have an effective choice of market outlets,
 INSTITUTIONAL CONSTRAINTS 241

and are therefore obliged to accept whatever price the buyer offers. It
is also held that potential competitors are deterred by the high costs
of entry into trading with farmers. The obstacles to entry include the
cost of advancing credit to suppliers, the threat of price warfare by
established traders, and even intimidation with threats of physical
violence (see e.g. Simpson, 1970, on market behaviour in Sudan). In
many areas it is customary for the farmer to rely on the final buyer of
his crop for an advance of credit to fund crop expenses or even the
purchase of food and other consumption necessities to bridge the gap
between seed-time and harvest. Where this practice is followed,
‘buyer attachment’ effectively precludes competition for the purchase
of the crop at harvest time.
The available fragmentary evidence on the structure, conduct and
performance of the markets serving peasant agriculture in LDCs is
conflicting. It suggests the co-existence of differing degrees of
competitiveness depending on numerous economic, social, cultural
and geographical factors. However, it seems fair to conclude that, as
in developed countries, where small farmers are completely un¬
organised their bargaining position in the market place is weak.

3.2.3 Lack of Market Information

The farmer’s bargaining position is further undermined by his lack of
information on current market prices, crop prospects and prospec¬
tive changes in demand, such as export orders. Unlike developed
countries, LDCs typically lack a formal market news service and, in
any case, many farmers are unable to receive up-to-date news reports
by mass media. Although it is claimed that news travels fast by word
of mouth, even to isolated areas, traders naturally tend to be better
informed than farmers.

3.3 Marketing reform: objectives and implementing policies

The principal objectives of marketing reform follow from the three

major market imperfections we have identified:

(1) to improve the quantity and quality of infrastructural provision,

especially in transport and other forms of communication;
(2) to strengthen the bargaining power of farmers vis a vis the market
intermediaries through whom they buy and sell; and
(3) to improve the quantity and quality of agricultural market
intelligence.
The means of attaining (1) and (3) can be dealt with very briefly.
242 AGRICULTURE AND ECONOMIC DEVELOPMENT

Infrastructural development, and the collection, processing and

dissemination of market intelligence, are both pre-eminently the
responsibility of government. Indeed, as we have already remarked,
because general infrastructural developments are, in effect, additions
to social overhead capital, government funding is virtually inevitable.
As for the provision of market intelligence, it is generally easier for
government to inspire confidence in its impartiality than it would be
for a private intelligence agency. In the rest of this section we
concentrate on the choice of means for achieving objective (2) - that
of strengthening the bargaining power of small-scale farmers. First,
we consider the potential role of voluntary marketing co-operatives,
and then that of statutory marketing boards, or ‘compulsory co¬
operatives’. We shall conclude that public policy can play a very
important role in promoting marketing reform by both strategies.
The potential role of marketing co-operatives with voluntary
membership is to provide their members either with a ‘new’ (i.e.
hitherto unavailable) service, or to replace an existing but inadequate
one. The reason for dissatisfaction with the existing service may be
that its cost is considered to be unnecessarily high, or its quality
unacceptably low, or both. However, the record of voluntary
agricultural marketing co-operatives in LDCs is not encouraging.
Commonly-encountered problems include an inability to retain the
loyalty of members to obtain competent yet incorruptible managers,
and to prevent subversion of the co-operative by trading competitors.
Fulfilment of at least three conditions appears to underlie the
successful establishment of voluntary marketing co-operatives. They
must be capable of:

(1) providing something which is actively wanted by the potential

membership and to achieve which members are prepared to
sacrifice part of their individual freedom for the common good;
(2) eliciting an active response from their members and retaining
their loyalty and confidence; and
(3) making visible progress in reducing marketing costs or improv¬
ing the quality of marketing services, often in competition with
rival traders.

Experience suggests that two further factors are favourable to

successful establishment. First, it is easier to achieve success by
providing a specialised yet relatively uncomplicated service than by
attempting to give a comprehensive set of marketing services. For
example, co-operative cotton ginning and transport services (for
cotton and other crops) have quite often achieved success. Second,
 INSTITUTIONAL CONSTRAINTS 243

the probability of success is greater for products with a limited choice

of marketing channels than for products for which there are many
different outlets. For example, an export crop which can be sold only
to overseas markets has the advantage that no opportunity exists for
more profitable domestic disposal and thus co-operative members
are more likely to remain loyal.
Government sometimes attempts to assist agricultural co¬
operatives not only by according them privileges, such as tax
concessions, but also by granting monopoly trading rights. Generally
though, this seems inadvisable since the spur of competition en¬
courages efficient marketing by both co-operatives and private
enterprise.
Agricultural marketing boards differ from voluntary co-operatives
in that they are more directly controlled by government (they exercise
delegated powers) and membership is complusory. Technically,
marketing boards are producer organisations in the sense that
producers or producer representatives share in the management. But,
in LDCs, government nominees (officials or politicians) tend to
dominate the management, so that boards are under closer and more
direct government control than their counterparts in developed
countries. The possible functions of marketing boards, and the
powers they may exercise, are wideranging. A useful classification
scheme (Abbot and Creupelandt, 1966) separates them into six main
types. A major distinction exists between non-trading types, which
leave the existing marketing structure unchanged, and trading
boards, which transform the structure, either by trading in compe¬
tition with private enterprise or by assuming monopoly trading
powers. In ascending order of degrees of market intervention, the
principal functions forming the basis of classification are:

(a) advice on presentation and product promotion,

(b) market regulation (trading standards),
(c) price stabilisation (without regular trading),
(d) competitive trading, and
(e) monopoly trading (export market only, or export and domestic
markets).

A primary function of marketing boards is to enhance producers’

market returns. This may be achieved either by reducing the costs of
marketing, or by selling additional services, or even by curtailing
supplies to raise the selling price. Another major function is to reduce
market uncertainty through price stabilisation. In addition, market¬
ing boards in LDCs are sometimes used by government as a
244 AGRICULTURE AND ECONOMIC DEVELOPMENT

convenient means of taxing the agricultural sector. It has been

objected that, by exercising this fiscal function, marketing boards
effectively ‘rob’ producers of part of their market returns (Bauer,
1954). But if it is accepted that extraction of part of the agricultural
surplus may be a legitimate objective of government policy in LDCs
(see chapter 3, section 4.1) then this may be an effective means of
collection (Blandford, 1979).
In principle, because of their compulsory membership, marketing
boards possess much greater bargaining power than voluntary co¬
operatives. But, in practice, board operations are very difficult to
appraise objectively for several reasons (Abbott and Creupelandt,
1967). Relevant information is difficult to obtain due either to the use
of inadequate or inappropriate accounting and recording proce¬
dures, or to the deliberate secretiveness of governments wishing to
avoid political embarrassment. Even if more adequate data were
available, valid comparative analysis would still be very difficult
because of the widely differing conditions in which boards operate.
Due to the lack of objective evidence we can only compare the
potential achievements of agricultural marketing boards with the
possible costs. Potential achievements include: enhanced producer
returns through improved market presentation; economies of scale in
selling and other marketing activities; easier access to capital for
market development; price stabilisation; and raising returns to
producers by controlling supplies and practising market
discrimination.
But are these goals achievable by some alternative means at a lower
cost? In principle, some of them probably are, either by direct
government regulation (but not trading), or by government colla¬
borating with voluntary marketing co-operatives, or even with the
private sector. Marketing board achievements depending on the use
of monopoly trading powers are, of course, exceptional. There is no
substitute for monopoly. However, the hidden social costs of trading
monopolies are well known, not only to consumers but also to
producers compelled to trade with them. Although the balance
between increased efficiency from exposure to competition and from
economies of scale need not invariably favour competition, it will
often do so even for producers.

3.4 Price policy and price stabilisation mechanisms

The general level of agricultural prices fixes the internal terms of trade
between agriculture and other economic sectors. Governments often
 INSTITUTIONAL CONSTRAINTS 245

seek to influence the purchasing power of agricultural products for

two reasons: first, to give farmers a bigger incentive to expand output
for the market and, second, to accelerate the transfer of capital from
agriculture to the remainder of the economy. Since these are
potentially conflicting objectives, requiring opposing agricultural
price adjustments, finding the ‘correct’ balance is no easy task. From
the A sector’s point of view, a positive price policy weights the terms
of trade in agriculture’s favour, whereas a negative price policy does
the opposite. LDC governments are frequently charged with pursu¬
ing a negative price policy; slow agricultural growth is often
attributed to this. Whilst such criticisms are doubtless sometimes
valid, there is no obvious economic argument to justify giving
agriculture a permanent terms of trade advantage.
The economic case for government intervention to stabilise
agricultural prices is inherently stronger than the case for raising
them above the mean level or trend dictated by market forces. Buffer-
stocks and buffer-funds are widely used as stabilising mechanisms,
with agricultural marketing boards frequently acting as the stabilis¬
ing agency.
Buffer-stocks are built up to raise market prices in periods of over¬
abundant supplies, and run down to lower prices when current
supplies are less than normal. In Figure 8.4, SL is the market period
supply curve when the crop yield is ‘low’, and SN and SH, respectively,
represent years of ‘normal’ and ‘high’ yields. Thus, without price
stabilisation, supplies and prices fluctuate from Oqj and Ph in
low-yield years, through Oq2 and Pn in normal-yield years, to Oq3
and Pj when yields are high. Assume Oq3- Oq2 = Oq2 - Oqj. In
times of glut the buffer-stock agency removes Oq3 — Oq2 from the
market in order to force the price up from P, to Pn: in years of
relative shortage stocks amounting to Oq2 — Oqt are released to
force the price down from Ph to Pn. Thus Pn is the stabilised producer
price. Ignoring the costs of storage and handling, and the interest
due on deferred income, producers gain in glut years what they lose
in years of shortage. These offsetting gains and losses of marginal
revenue are respectively represented by the triangular areas A and
R. In this hypothetical and idealised example the stocks are
self-liquidating over a period of time in which there are equal numbers
of high- and low-yield years. In practice, however, the ex ante
determination of, and strict adherence to, the stabilised price, Pn,
presents problems. In order to estimate reliably the equilibrium
market price in a ‘normal’ year, the buffer-stock agency needs a
complete model of the market which both specifies all the variables
246 AGRICULTURE AND ECONOMIC DEVELOPMENT

affecting supply and demand, and also permits estimation of their

parameters together with the probability distribution of stochastic
errors.
Even if the agency is able to estimate the market-clearing price in a
normal year, producers may well exert pressure for the price to be
stabilised above that level. The consequences of yielding to such
pressure can be followed through in Figure 8.4. Designate the level
around which producers desire the price to be stabilised the ‘target
price’. Suppose the target price is not set at Pn but at Ph. Then the buffer-
stock agency can no longer choose between accumulating and
maintaining stocks at the current level. Given the supplies of Oq1( the
target price Ph can be obtained from the market without changing
stocks. But if supplies are Oq2 or Oq3 the agency is
 INSTITUTIONAL CONSTRAINTS 247

compelled to add to its stocks in order to obtain the target price.

I hus over time, stocks can only accumulate, they can never decline-
in other words they are no longer self-liquidating. Moreover, unlesJ
stocks are self-liquidating a buffer-stock scheme cannot finance itself,
hver-accumulating stocks require periodic capital in injections to
defray the ever-rising costs of holding them.
Buffer-/w/jJ schemes possess the advantage that the costs of holding
physical stocks are avoided. The fund accumulates revenue by
taxing producers in high price years, and disburses revenue by
subsidising them when market prices are relatively low. Again, the
financial viability of the scheme depends on pitching the target price
near to the level which equates supply with demand in a ‘normal’
season.

3.5 The state versus private enterprise in marketing

The case for some government regulation of agricultural marketing is

self-evident. Compliance with fair trading practices and
adherence to grade standards, for example, must be policed; overseas
trade must be monitored and, if necessary, controlled. But how
desirable is it that government should actually participate in market¬
ing? In seeking an answer to this question we first consider why
government may choose to participate (including some non¬
economic reasons). We then consider a number of a priori reasons
against participation (FAO, 1969, pp. 94-106).
From observing how governments behave, the possible reasons
why they decide to participate in marketing appear to include:
(1) Attempting to obtain more control over national resources and
the national economy in the furtherance of national develop¬
ment, by facilitating the control of prices and foreign exchange,
and by simplifying the collection of taxes, for example.
(2) Attempting to obtain increased bargaining power for domestic
producers in export markets, by setting up monopsonistic export
marketing boards.
(3) Attempting to develop new markets and products which the
private sector is unable or unwilling to enter or handle.
(4) Attempting to dislodge foreign trading enterprises from domestic
markets for ideological reasons, to gain ‘local control’ in markets
where indigenous private enterprise is absent or unable to
withstand foreign competition.
(5) Pandering to popular prejudice against private middlemen as a
class, regardless of whether they are nationals or foreigners.
248 AGRICULTURE AND ECONOMIC DEVELOPMENT

Whilst (l)-(3) possess an underlying economic rationale for govern¬

ment participation, (4) and (5), which are non economic, clearly do

An important part of the case against participation derives from a

number of potential handicaps of public enterprise in marketing,
these are:

(1) Exposure of such enterprises to political pressure, possibly

affecting the appointment of staff, choices of buying and selling
contracts and many other policy or managerial decisions. In
extreme cases staffing may occur entirely within a system of
political patronage.
(2) Overhead costs are excessive and operating costs unnecessarily
high because of over-staffing and injudicious investment for
reasons of‘prestige’ or due to inadequate planning, for example.
(3) The civil servants appointed to manage public enterprises lack
the judgement, experience and self-confidence needed to make
sound commercial decisions; they are trained to contain public
expenditure, within the constraints of a fixed budget, rather than
to make profitable investments.
(4) Governments expect public marketing enterprises to show a
commercial profit and, at the same time, undertake financial loss¬
making ‘social service’ functions, without the benefit of financial
subsidy. So, for example, a marketing board or public marketing
corporation may be required to accept produce from all indiv¬
idual producers on the same terms, regardless of the size,
frequency and timing of their deliveries.

Certain potential strengths of private marketing enterprise are

implied by these potential weaknesses of government or public
enterprise. Thus, private entrepreneurs tend to be superior in business
acumen, commercial judgement and risk evaluation. They are also
less open to political pressures and government interference in
discriminating between commercial and non-commercial marketing
services. But, despite the shortcomings of public marketing enter¬
prise, a valid case for a measure of government participation remains,
based on any or all of the above-listed economic reasons for
participation, (1)—(3).
Rather than relying exclusively on either public or private market¬
ing enterprise to achieve the goals of agricultural marketing policy,
better results may be achieved by adopting a policy of partnership
between government and the private sector (Fenn, 1972). The
 INSTITUTIONAL CONSTRAINTS 249

economic rationale underlying this approach is a form of compara¬

tive advantage. For their part, private concerns are frequently
capable of lower cost operations and tend to be better able to provide
the drive and initiative needed to sustain managerial efficiency and
innovation. But public intervention is needed to control and direct the
activities of private enterprise in accordance with development plans
and policies. Also, public enterprise may be needed to function and
assume risks for social gains where private enterprise is absent.
Moreover, definite policies are needed to reduce market imperfections
and stimulate more effective competition in the private sector. The
state may also be able to finance capital projects, such as processing
plants and storage space, where private investment in such facilities is
inadequate.
A precondition of fruitful partnership between the public and
private sectors is a willingness to co-operate on both sides. But
government officials and politicians in LDCs sometimes display an
attitude of indifference (or even hostility) towards the private sector.
A remedy may lie in improved education and training to give
administrators, and even politicians, a better understanding of the
benefits of the economic services provided by private sector market
intermediaries.

3.6 Summary and conclusions

The basic objectives of agricultural marketing policy in LDCs include

infrastructural development, improving the quantity, quality and
dissemination of market information, strengthening the bargaining
position of individual producers, and reducing uncertainty by
stabilising prices. In pursuit of these objectives general infrastructural
development and the improvement of market information services
are primarily the responsibility of government. Government is also
responsible for enforcing necessary marketing regulations as well as
devising and implementing policies to enhance marketing efficiency
by reducing market imperfections and making competition more
effective. Price stabilisation may be undertaken either by government
or by a producers’ organisation. The extent of government trading
activities is likely to depend in part upon the extent and quality of the
services offered by private enterprise and partly on government’s
ideological stance. A good case exists for active collaboration
between government and private marketing enterprise in order to
provide agricultural marketing services at minimum cost, subject to
the proviso that the quality and range of services offered is consistent
250 AGRICULTURE AND ECONOMIC DEVELOPMENT

with the effective service demands of both producers and consumers

of agricultural products.

References

Abbot, J.C. and Creupelandt, H.C. (1966), Agricultural Marketing

Boards: their Establishment and Operation, FAO, Rome.
Abbott, J.C. and Creupelandt, H.C. (1967), Agricultural marketing
boards in developing countries’. Monthly Bull. Agric. Econ. & Stats.,
16(9).
Adams, Dale W. and Nehman, Gerald I. (1976), ‘Measuring farm
level credit use: a Brazilian example’. Agricultural Finance Review, 39:
77-9, April.
Adams, Dale W. and Tommy, Joseph L. (1974), ‘Financing small
'farms: the Brazilian experience, 1965-69’, Agricultural Finance
Review, 35: 36-41, October.
Agency for International Development (1973), Spring Review of
Small Farmer Credit, 20 vols., February-June.
Bardhan, P. and Rudra, A. (1978), ‘Interlinkage of hand labour and
capital relations: an analysis of village survey data in East Asia’,
Economic and Political Weekly, pp. 367-84.
Bauer, P.T. (1954), West African Trade, Cambridge University Press,
Cambridge.
Bauer, P.T. and Yamey, B.S. (1957), Economics of Underdeveloped
Countries, Cambridge University Press, Cambridge.
Berry, R.A. and Cline, W.R. (1979), Agrarian Structure & Productiv¬
ity in Developing Countries,- Johns Hopkins University Press,
Baltimore.
Bhaduri, A. (1973), ‘A study in agricultural backwardness under
semi-feudalism’, Economic Journal, 83, pp. 120-37.
Bhaduri, A. (1977), ‘On the formation of usurious interest rates in
backward agriculture’, Cambridge J. Econ., 1, pp. 341-52.
Blandford, D. (1979), ‘West African export marketing boards’, in
Hoos, S. (ed.), Agricultural Marketing Boards: An International
Perspective, Ballinger, Cambridge, Mass.
Bliss, C. and Stern, N. (1982), Palanpur: A Study of an Indian Village,
Macmillan, London.
Bottomley, A. (1971), Factor Pricing and Economic Growth in
Underdeveloped Rural Areas, Crosby Lockwood, London.
Chandavarkar, A.G. (1971), ‘Some aspects of interest rate policies in
 INSTITUTIONAL CONSTRAINTS 251

less developed countries: the experience in selected Asian countries’,

IMF Staff Papers, 18, pp. 48-112.
Dorner, P. (1972), Land Reform and Economic Development,
Penguin, Harmondsworth.
Fenn, M.G. (1972), ‘The entrepreneur in agricultural marketing
development', Monthly Bull. Agric. Econ. & Stats., 21 (7/8).
Food and Agriculture Organisation of the United Nations (1969),
State of Food and Agriculture, 1969, FAO, Rome.
Ghatak, S. (1975), ‘Rural interest rates in the Indian economy’, J. Dev.
Studies, 11, 3, pp. 190-201.
Ghatak, S. (1976), Rural Money Markets in India, Macmillan, India.
Ghatak, S. (1983), ‘On inter-regional variations in rural interest rates
in India’, J. Developing Areas (forthcoming).
Griffin, K.R. (1976), Land Concentration and Rural Poverty, Mac¬
millan, London, (2nd edn.).
Gupta, S.G. (1975), ‘Interest sensitiveness of deposits in India’,
Economic and Political Weekly, 20, pp. 2357-63.
Lewis, S.R. (1967), ‘Agricultural taxation in a developing economy’,
in Southworth, H.M. and Johnston, B.F. (eds), Agricultural Develop¬
ment and Economic Growth, Cornell University Press, Ithaca.
McKinnon, R.I. (1973), Money and Capital in Economic Develop¬
ment, Brookings Institution, Washington, DC.
Mellor, J.W. (1966), Economics of Agricultural Development, Cornell
University Press, Ithaca.
Mirrlees, J. (1975), ‘A pure theory of underdeveloped economies’, in
Reynolds Lloyd (ed.), Agriculture in Development Theory, Yale
University Press, New Haven.
Nguyen, D.T. and Martinez-Saldivar, M.L. (1979), ‘The effects of
land reform on agricultural production, employment and income
distribution: a statistical study of Mexican states, 1959-69’,
Economic J., no. 355.
Putterman, L. (1983), ‘A modified collective agriculture in rural
growth-with-equity: reconsidering the private unimodal solution,
World Development, 11(2), pp. 77-100.
Rahman, A. (1979), ‘Agrarian Structure and Capital Formation: A
Study of Bangladesh Agriculture with Farm Level Data’, un¬
published PhD thesis, Cambridge.
Shaw, E.S. (1973), Financial Deepening in Economic Development,
Oxford University Press, New York.
World Bank (1975), The Assault on World Poverty, Johns Hopkins
University Press, Baltimore.
9 Population and Food
Supplies

The relationship between population and food supplies has been the
subject of major economic debate for at least 200 years. T .R. Malthus
published his famous ‘Essay on the Principle of Population’ in 1789.
But Malthus was responding to Adam Smith’s more optimistic view
of the economic implications of population growth, as expressed in
‘The Wealth of Nations’ published in 1776. Other classical econo¬
mists in the British tradition, like David Ricardo and John Stuart
Mill, were much influenced by Malthus and his stress on the
restrictive influence of diminishing returns.
In what are now termed the developed countries the discipline of
economics has now lost some of its reputation as the ‘dismal science’.
Disbelief in the ‘iron law of wages’ has been encouraged partly by the
declining relative importance of agriculture and other primary
industries in national products and partly by the modernisation of
agriculture itself, which has diminished the relative importance of
land as a factor of production. Moreover, developed countries
generally now enjoy the advantage of a relatively stable and low rate
of population growth.
A very different view is apt to prevail in many of today’s LDCs,
where agriculture and other primary industries are still dominant.
Moreover, most of agriculture is still ‘traditional’ (see chapter 2) and
a high rate of population growth prevails (higher than historic rates
at a comparable stage of development in what are now the developed
countries). In such a situation it may well seem that there is much less
reason to discard the pessimistic Malthusian view of the prospects for
improving the economic lot of the broad masses of a country’s
population.
This chapter falls into two main sections. First we outline the

252
 POPULATION AND FOOD SUPPLIES 253

classical, Malthusian-type model of economic development. We then

contrast this with two alternative models of population and food
supplies-a contra-Malthusian model in direct antithesis to the
classical model, and an ‘ecological model’, postulating the existence
of ‘natural’ checks on uncontrolled population growth in culturally
stable human societies. Section 1 concludes with an attempt to
synthesise these various approaches to explaining the relationship
between population growth and food supply. In Section 2 mal¬
nutrition is defined and its incidence and extent in LDCs is examined.
In so doing we distinguish between deficient aggregate food supplies
and the maldistribution of food. The causes of famine are also briefly
examined. The section concludes with a brief examination of some
alternative policy approaches to ameliorating the problem of malnut¬
rition in LDCs.

1 The Classical Model

The basic argument of the classical theory of development, as

exemplified by Malthus, can be summarised in five steps:

(1) Increased population causes a parallel increase in the demand for

food.
(2) This increased demand for food can be met either by bringing
‘new’ land into cultivation, or, by cultivating existing cultivated
land more intensively than before, i.e. by applying more labour.
(3) Land is not only scarce (i.e. limited in supply) but also variable in
fertility. The most fertile land is cultivated first. Then, in response
to continuing population pressure, less and less fertile land is
successively brought into cultivation. Thus, the least fertile land
being cultivated at any particular time is to be found at the
‘extensive margin of cultivation’. (‘Fertility’ is synonymous with
‘productivity’ in this context.) Thus, as a consequence of bringing
additional land into cultivation at the extensive margin, the
marginal productivity of agricultural labour is bound to decline.
(4) If additional labour is applied to existing land (at the intensive
margin) its marginal productivity will again decline due to
diminishing returns.
(5) Since diminishing marginal returns to agricultural labour are
inevitable and unavoidable, food production will always tend to
grow less rapidly than population. In the long run population will
always expand to the limit of available food supplies at the
254 AGRICULTURE AND ECONOMIC DEVELOPMENT

subsistence level. That is, although the rate of growth in food

supplies may temporarily get ahead of population growth - due,
for example, to an unexpected and temporary decline in popula¬
tion or the fortuitous discovery of a more productive method of
farming - continual population growth must inevitably force per
capita food supplies back to a bare subsistence level in the long
term.

More formally, the classical model implicitly assumes that the

production function expressing the relationship between population
(as a proxy measure of agricultural labour input) and aggregate
agricultural output is continuous through time. Thus, in Figure 9.1
where the vertical and horizontal axes respectively measure aggregate
agricultural product and population, TP and AP are the total and
average product functions. The shape of the TP function is consistent
with the classical model assumption that aggregate agricultural
product is doubly bound by the constraint of diminishing returns
at both the intensive and extensive margins of cultivation. The model
assumes that agricultural wages are determined by the average
productivity of labour. Thus the limit to population increase is at
OPm where AP is at the minimum subsistence level OM. If
population exceeds this limit people will starve. The slope of the
total consumption function TCS, is consistent with OM. An agri¬
cultural surplus exists provided TP > TC. Such a surplus may be
mobilised either for investment to provide employment outside
the farming sector or to raise productivity and living standards
within agriculture itself (Nicholls, 1963). But, with population at
OPir following unconstrained population growth, the surplus is
zero.
The obvious policy implication of the classical model is that in
predominantly agrarian societies curtailment of population growth is
the sole feasible means of materially improving living standards for
the majority of the population. It is easy to criticise this model with
the benefit of hindsight. The classical economists did not foresee the
extent to which European countries would be able to expand their
agricultural production at both the intensive and extensive margins
of cultivation, by means of land improvement, improved communi¬
cations and the adoption of scientific methods of farming. At the
time of Malthus, most of the major scientific advances which have
revolutionised developed country agriculture during the last 150 to
200 years, including the invention of chemical fertilisers and the
selection and breeding of higher yielding crop varieties and strains of
 POPULATION AND FOOD SUPPLIES 255

livestock, still lay in the future. Equally important, the extent to

which population pressure in the Old World would be relieved by
emigration to the New World was not foreseen: nor was it anticipated
that an ocean transport revolution would effectively enable much of
Western Europe’s extensive margin of cultivation to be transferred to
the New World.
Although the classical economists can be absolved from failing to
foresee the inforeseeable, they are vulnerable to the criticism that it is
unrealistic to assume that land at the extensive margin of cultivation
must necessarily be of lower fertility than any supra-marginal land.
Economic location theory postulates that the rent of land is jointly
determined by its fertility and its distance from the market. If either
factor is constant the other - and the rent - are free to vary. Thus,
there is no more reason why all plots which are equidistant from the
market should be equally fertile (or attract the same rent) than that
256 AGRICULTURE AND ECONOMIC DEVELOPMENT

all equally fertile land should be the same distance from the market:
both propositions are absurd. Land at the extensive margin of
cultivation is likely to consist of a mixture of relatively infertile land
close to the market and relatively fertile land further away from the
market. Moreover, in a market economy land is valueless unless it is
accessible, regardless of its inherent fertility or distance from the
market (in the sense of a line drawn on a map). Thus, even in
developed countries, the extent of the area that can be used for
commercial agriculture is limited in the short term by the existing
communications infrastructure. In LDCs the short-term limiting
effects of an underdeveloped transport infrastructure on agricultural
expansion are likely to be even more significant.
So, contrary to one of the basic premises of the classical model,
there is no reason to expect declining marginal physical productivity
of labour at the extensive margin due to declining soil fertility. The
principal constraint on expanding production at the extensive margin
is likely to be rising marginal costs (especially transport costs) rather
than declining marginal returns.
Despite this technical criticism, the Malthusian model may still
have policy relevance for some present-day LDCs. These countries
lack the population pressure safety-valve of massive overseas
emigration. Moreover, they are unable to import large quantities of
foodstuffs unless available on concessional terms. Most importantly,
population growth rates in what are now the developed countries
never reached the very high levels of up to 3 per cent annum which
characterise today’s LDCs. Whereas the developed countries had
many years in which to adapt their birth rate to a gradually declining
death rate and rising living standards, many LDCs have experienced
a later and much more rapid decline in the death rate, whilst the
majority of the population (much of it in agriculture) remains very
poor. Thus, in the LDCs, birth rates have remained at, or very close
to, their traditional high levels.

2 Contra-Malthusian Model

Whereas the classical model postulates that, in the long run,

population growth is limited by food supplies, the contra-Malthusian
model argues that food supplies increase in response to population
growth. These sharply opposed views of the relationship between
population and food supplies result from equally contrasting views
of the process of technical change in agriculture. The classical (and
 POPULATION AND FOOD SUPPLIES 257

neo-classical) tradition treats the discovery of improved technology

as an exogenous shift variable, i.e. discovery is fortuitous. Since, by
definition, improved methods of production offer the means of
reducing costs of production, discovery is rapidly followed by
adoption on the part of profit-maximising producers. In contrast,
the contra-Malthusian hypothesis that food supplies increase in
response to population growth rests on the assumption that a stock
of improved technology is accumulated (again fortuitously) but
adoption depends on population pressure. (Boserup, 1965).
Boserup’s concept of the dynamics of agricultural technology
concerns change in cultivation methods and the choice of tools in
relation to changing systems of land use. Population density is the
primary determinant of the system of land use, the method of
cultivation, and the choice of tools at a particular time. As population
pressure increases, progressively more intensive systems of land use
are adopted, combined with consequential changes in methods of
cultivation and the choice of tools, in order to offset any tendency for
food output per capita to decline, due to diminishing returns. In more
formal terms, the model postulates that although in the short run
there may be diminishing returns to agricultural labour, in the long
run the aggregate agricultural production function will always shift
upwards in response to population pressure, at whatever rate is
required to maintain output per capita. The shift variable is
agricultural technology, as expressed by the system of land use, the
method of cultivation and choice of tools. Thus Boserup explicitly
rejects the classical model notion of a continuous agricultural labour
productivity function: in her view the true function is discontinuous
and the reason for discontinuity is technological change.
The structure of the Boserup model, assuming a Cobb-Douglas-
type production function, is formally presented in the appendix.

2.1 Differential intensity of land use systems

Boserup identifies a succession of increasingly intensive systems of

land use in primitive agriculture and correlates these with increasing
population density. The lowest level of population density is
characterised by the system of‘forest fallow’, with a fallow period of
20-25 years between successive crops. At a somewhat higher level of
population density, forest fallow gives way to ‘bush fallow’ in which
the fallow period is reduced to 6-10 years. Shortening the fallow
period raises the labour: land ratio as well as output. In response to
further population growth and increase in the agricultural labour
258 AGRICULTURE AND ECONOMIC DEVELOPMENT

supply, bush fallow is successively followed by ‘short fallow’, with a

fallow period of 1-2 years only, and ‘annual cropping’, with the
fallow period reduced to less than a year. The most intensive system
of land use - at the opposite extreme to forest fallow - is ‘multi¬
cropping’, under which the land is cropped more than once annually
with virtually no break between successive crops. Other things being
equal, the multi-cropping system is most labour intensive as well as
yielding the highest output per unit of land area per unit time period.

2.2 Method of cultivation and choice of tools

Forest fallow farming, the most extensive system of land use, is

virtually a ‘no-cultivation’ system. The land is not broken, the farmer
merely fells and burns the forest and plants his crop in the ashes. The
essential tool is the axe and little else is needed.
The system of bush fallow calls for the use of the hoe (or digging
stick) to help clear the land for planting and also to keep down weeds
during the growing season. As the fallow period is successive¬
ly shortened, so the ‘weed problem’ intensifies and, by the
annual cropping stage, the use of the plough - to bury weeds
before planting - becomes essential. The introduction of the plough
also entails setting land aside for the grazing of draught animals
(motor tractors being strictly outside the ‘primitive agriculture’
terms of reference). Thus the incremental production resulting from
the use of the plough must exceed the production lost by diverting
land to grazing. Multi-cropping frequently involves the adoption of
irrigation, requiring specialised equipment, in order to break the
constraint on production imposed by the natural ‘dry season’.

2.3 Land use intensity and labour productivity

A further important feature of Boserup’s theory is that, because the

necessary number of cultivations and frequency of attention to crops
is inversely related to the length of the fallow period, the length of the
working day in agriculture increases with intensification of the system
of land use, i.e. forest fallow farmers work the shortest hours and
those practising multi-cropping the longest. This is the exact opposite
of the widely-held view that the short working day is most prevalent
in densely populated countries, and vice versa. But long fallow
farmers are not disguisedly unemployed in the sense of Sen (see
chapter 3, pp. 52-4). According to Boserup, their leisure preference
is high and they are prepared to sacrifice leisure at the margin of their
 POPULATION AND FOOD SUPPLIES 259

customary working hours only when forced to do so by population

pressure which increases the relative scarcity of food. Only then are
they willing to switch to a more intensive system of farming entailing
not only changed methods of cultivation but also longer working
hours. Underlying this argument is the assumption that in primitive
agriculture the motivation for work is the attainment of a fixed
income target (with the income consisting primarily of food) rather
than income (or utility) maximisation. Thus, in Figure 9.2, an
individual producer’s utility surface, as signified by indifference curves
Ix, I2 and I3, is superimposed on his production possibility surface,
as indicated by the curves Pl5 P2 and P3. Work time is assumed to

Real
Income

Figure 9.2: Fixed income target model

260 AGRICULTURE AND ECONOMIC DEVELOPMENT

be the sole variable input. The labour time response curve to

opportunities for gaining real income through improved methods of
cultivation is indicated by SS'. Whereas the neo-classical producer
with unlimited wants could move all the way from Wj to w3 work time
in response to the opportunity of increasing his real income from yt
to y3, the producer with a fixed income target of yfyf would not be
prepared to exceed wm work hours unless forced to do so in order to
maintain the income target.
A major policy implication of the contra-Malthusian model is that,
at least in the long run food output will always tend to keep pace with
population growth through changes in the intensity of cultivation
and choice of tools. However, food output response to growing
population pressure may be inhibited by policy constraints such as a
deficient marketing and distribution infrastructure and unfavourable
movements in agriculture’s terms of trade, (Boserup, 1975). Despite
some empirical evidence lending support to the hypothesis that in
LDCs technical change in agriculture responds to population
pressure, such as the results of a study of rice cultivation in Sierra
Leone (Levi, 1976), Boserup’s model is clearly open to several
criticisms.
First, the model provides a much more plausible explanation of
how food supplies might keep pace with population growth in a
sparsely populated country than in a densely populated one. Suppose
that a country has no idle land so that agricultural expansion at the
extensive margin of cultivation is precluded. Suppose further that the
transition to the most intensive known system of cultivation has
already occurred, but the population is still rising due, for example, to
the favourable effect of continuing improvement in standards of
public health on the death rate. What options then exist for
maintaining per capita food supplies? Within the constraints of the
present model the two main options would appear to be either
(a) attempting to increase aggregate food output by applying more
labour within the existing system, or (b) progressing to an even more
intensive system of land use and choice of tools. Whereas the
successful application of (a) might well be frustrated by very low or
even zero marginal returns to labour, the feasibility of (b) could
depend upon the scientific discovery of some hitherto unavailable
technology. But the concept of technical progress in agriculture being
dependent on scientific discovery and the effective transmission to
farmers of the fruits of agricultural research and development, is
absent from the Boserup model. Thus this model scarcely offers
an effective reply to the argument that the only possible escape
 POPULATION AND FOOD SUPPLIES 261

route from the Malthusian trap for the most densely populated
poor countries is larger food imports provided on concessional
terms.
A second serious criticism also concerns the nature and costs of
technical change in agriculture. Boserup emphasises the economic
incentives needed to induce producers to change their production
techniques. However, the model fails to account fully for the costs of
adopting new methods. Although the investment costs of intensifying
the system of cultivation are not entirely neglected, there is too much
emphasis on the special case of farmers utilising their own low
opportunity cost labour to create additional farm capital to the
neglect of types of investment necessitating the purchase of modern
capital inputs from outside the agricultural sector. For example,
although hand digging may be the ‘best’ means of excavating a well if
labour is sufficiently plentiful and cheap, it is less likely that manual
labour can feasibly be substituted for a mechanical pump with equal
effect as the means of extracting the water. But, for a poor peasant
farmer, finding the money capital to buy a pump, and possibly other
irrigation equipment too, may present a virtually insurmountable
obstacle. Such problems are ignored by the Boserup model.
Moreover whereas most LDCs now plan to accelerate their rate of
agricultural development and raise per capita food output, the
Boserup model is too long term and insufficiently specified for use as a
planning model. Furthermore, although rapid population growth in
LDCs may have contributed to the urgency of the international
search for improved agricultural inputs such as high-yielding
varieties, this model does not account for the process whereby the
fruits of research conducted mainly in developed countries have been
successfully transmitted to and adopted by farmers in LDCs.
A third criticism is that Boserup presents little convincing evidence
to support her contention that primitive agriculturists typically
exhibit a strong leisure preference. Although it is not irrational for
farmers to value their leisure above the disutility of work at the
margin of their working hours, the hypothesis lacks plausibility in a
context of relative poverty, short working days and ample leisure.
Much stronger empirical support is needed to carry with conviction
the argument that the hypothesis has general validity.
However, despite these criticisms, the contra-Malthusian model,
deserves serious attention for its novel analytical insights into the
process of agricultural development as well as for the counter¬
arguments it provides against the over-gloomy prognostications of
neo-Malthusians.
262 AGRICULTURE AND ECONOMIC DEVELOPMENT

3 Ecological Disequilibrium
The central hypothesis of Wilkinson’s ecological model of economic
development is that ‘changing ecological circumstances, centring on
the relationship between population and resources, force societies to
exploit their environment in new and often more difficult ways’
(Wilkinson, 1973, p. 4). The approach is multi-disciplinary, drawing
extensively on biological, anthropological and sociological evidence,
as well as upon the literature of economics, in order to support the
central thesis and related arguments.
In the natural world ecological equilibrium is the norm. ‘Rather
than over-exploit their resources [natural populations] build up a
pattern and a rate of resource use which the environment can sustain
indefinitely’ (ibid., p. 21). Wilkinson presents anthropological
evidence that primitive societies apply cultural restraints upon the
rate of resource use. Specifically with respect to the rate of popu¬
lation growth, these include abortion, infanticide and taboos upon
sexual intercourse during lactation. But, in mankind, ecological
equilibrium is not necessarily a permanent state. Ecological dis¬
equilibrium can occur as and when the constraints which maintain
human societies in ecological equilibrium break down. For example,
traditional beliefs and customs may be undermined due to contact
with alien cultures. Thus, traditional methods of population control,
like infanticide, have been abandoned due to the teaching of
missionaries that such practices are wicked.
When ecological equilibrium breaks down society will try to find
ways of developing its technology in order to increase the yield from
the environment - in agriculture more intensive systems of land use,
frequently involving the sacrifice of leisure, are adopted.1 Like
Boserup, Wilkinson contends that pre-industrial societies are charac¬
terised by a pronounced leisure preference; only in industrial societies
does the leisure/consumption (or work) trade-off shift markedly in
favour of consumption. But again, the empirical evidence supporting
this contention is weak.
The apparent policy implication of the ecological disequilibrium
model is that in LDCs, where the traditional cultural constraints on
excessive population growth have broken down and new agricultural
technology is incapable of closing the gap between food supplies and
food requirements, a new social consensus is needed to curb
population growth. But is such a policy feasible? And how might it be
implemented? The logic of the model suggests that re-establishment
(and maintenance) of ecological equilibrium might depend on large
 POPULATION AND FOOD SUPPLIES 263

numbers of individuals either conforming to new forms of social

taboo, or behaving altruistically. But in many LDCs, where the
pursuit of individual self-interest is encouraged to promote the
common goal of more rapid development, and where pluralism may
be gaining at the expense of consensus, such an option may be
unrealistic. However, although most countries with a population
policy rely primarily on the voluntary acceptance and use of birth
control techniques to limit family size, some have resorted to direct
social and economic pressures to discourage parents from having too
many children.
In China, for example, restricting family size to one or two children
has long been one of the main planks of social planning. Raising a
large family is anti-social since it diverts resources from the more
important tasks of self-education and productive employment.
Consequently, the Chinese government, as well as advocating family
planning, have exerted very strong social pressures to discourage early
marriage (which was traditional in pre-communist China). More
recently, social pressures to limit family size have been supplemented
by economic pressures, at least in some urban areas, where one-child
families receive preferential treatment with respect to housing and
taxation, whereas families with more than two children are dis¬
criminated against. Similar discriminatory policies against large
families have been adopted and applied in Singapore, a non¬
communist country (Birdsall, 1980, p. 60). Barriers to the wider
adoption of such policies in LDCs include the implied infringement
of parental liberty to determine their own optimum family size and
the view that, in any case, children should not be made to suffer
for the ‘social deviation’ of their parents.

4 Synthesis of Population and Food Supply Theories

A principal weakness of the classical model is its underestimation of

the potential role of technical progress in offsetting the effects of
diminishing returns to agricultural labour accompanying population
growth. The classical economists also mis-specified the extensive
margin of cultivation by not allowing adequately for the development
of communications giving better access to agricultural land both at
home and overseas. This mis-specification exaggerates diminishing
returns in agriculture. Nevertheless, the Malthusian model may still
have policy relevance for some present-day developing countries with
a high rate of population growth and severe technical or capacity
264 AGRICULTURE AND ECONOMIC DEVELOPMENT

constraints on either expanding domestic agricultural production or

increasing the volume of food imports.
The contra-Malthusian model, as interpreted by Boserup, draws
attention to the possibility of a relationship between population
growth and the adoption of new agricultural technology (which
Malthus missed). But Boserup- has nothing to say either about the
process whereby new agricultural technology is generated - either
within agriculture itself or in a separate agricultural service
sector - or about obstacles to its adoption - such as capital rationing
and risk-aversion (as discussed in chapters 2 and 5). Boserup’s model
is also difficult to square with the ‘internationalisation of agricultural
R & D’ whereby the fruits of research conducted in developed
countries are transferred to the LDCs. This process is exemplified by
the yield augmenting affects of the ‘seeds-fertiliser revolution’ which
has occurred in certain LDCs, particularly in South-East Asia during
the past 25 years (see chapter 6). The thesis that higher population
pressure will always induce increased productivity of land via the
adoption of more intensive cultivation methods is almost certainly
invalid for some of today’s very poor and most densely populated
countries. But, more generally, the long-term secular relationship
between population and agricultural output in most countries lends
more support to Boserup than to Malthus. However, a historical
relationship is not necessarily a fool-proof guide to the future.
The theory that population growth is limited by food supplies is
also rejected by Wilkinson, except under conditions of ecological
disequilibrium where cultural restraints on population growth have
broken down. Ecological disequilibrium forces the adoption of new
technology as the means of increasing agricultural output. Thus, the
theory of ecological disequilibrium is similar to the contra-
Malthusian model, except that Wilkinson treats technological inno¬
vation as ‘second-best’ to adherence to cultural restraints on
population growth in the race between population and food supplies.
Wilkinson’s theory poses the problem, for today’s over-populated
LDCs, of finding new population restraints or institutions to replace
outmoded traditional methods which would be unlikely to regain
social acceptance in the modern world. Recent developments in
China and one or two other LDCs suggest that, at least in some
societies, it may be possible to solve this problem.
Both the Malthusian and contra-Malthusian explanations of the
relationship between the rates of growth in population and food
supplies assume that causation is uni-directional. The difference
between them is that the direction of causation between the two vari-
 POPULATION AND FOOD SUPPLIES 265

ables is reversed. Whereas Malthusians assume that an autonomous

food supply growth rate (cause) limits population (effect), contra-
Malthusians assume that an autonomous increase in population
(cause) induces a corresponding increase in food supplies (effect).
Uni-directional causation is open to the criticism that it rules out
feedback. An alternative hypothesis is that as growing food supplies
induce population expansion, so population growth induces further
growth in food supplies. It is immaterial whether the causal chain is
initiated by an autonomous increase in food supplies or population.
The crucial point is that feedback is built into the system so that
neither variable is uniquely causal. A resolution of the opposing
viewpoints of Malthusians and contra-Malthusians might be found
in a model of this kind.
Each of the three models of the relationship between the growth of
population and food supplies which we have reviewed is open to
criticism. But, despite inherent weaknesses, each model also deserves
attention for any positive aspects which add to our knowledge and
understanding of a complex problem and which may also inform the
deliberations of policy-makers in LDCs.
One further aspect of this relationship deserves mention. It is naive
to suppose that if the growth of aggregate food supplies keeps pace
with population all individuals comprising the population will
necessarily be adequately fed. This is because, except in a society
where all households have direct access to land, there is a two-way
race between: (a) population and economic participation (or employ¬
ment) and (b) economic participation and food supply. The unem¬
ployed are bound to have difficulty in obtaining an adequate food
supply regardless of the aggregate supply situation (Poleman, 1981).
(We discuss food redistribution policies, to meet the needs of the
unemployed and other socially deprived groups, in section 6 below.)

5 Malnutrition in Developing Countries

Malnutrition refers to a state in which diet is inadequate, either in

quantity, quality, or both, for normal health and physical well-being;
thus adequate nutrition is a normative concept. Whereas, in a market
economy, an individual’s food consumption primarily depends upon
his income and tastes (assuming constant prices), his food require¬
ments depend upon the application of nutritional norms, including
allowances for interpersonal differences in age, sex and occupation,
as well as upon factors like climate variations.
266 AGRICULTURE AND ECONOMIC DEVELOPMENT

The conventional method of assessing the nutritional adequacy of

per capita food supplies for a country or region is to construct a ‘food
balance sheet’. This compares per capita food supplies - including
imports, and after making some allowance for wastage - with
objectively determined per capita nutrient requirements. The nut¬
ritional status of the country (or region) is judged to be inadequate
only if per capita supplies are exceeded by per capita requirements.
This approach to the measurement of malnutrition is open to
several criticisms (Poleman, 1981). First, the assumption that per
capita food requirements are homogeneous, even amongst people of
similar age and the same sex, is clearly very crude. There must
inevitably be a tendency to overestimate ‘reference man’s’ food
requirements in order to err on the safe side. Second, estimates of
available food supplies are highly likely to underestimate the actual
amounts available for reasons such as deliberate underreporting by
farmers (to reduce taxation, for example) and statistical errors such
as incomplete enumeration and underallowance for the volume of
subsistence agricultural production (intrinsically difficult to mea¬
sure). Third, and most importantly, by concentrating attention on
what per capita food consumption would be if available supplies were
equally distributed within the population, the food balance sheet
approach ignores the whole problem of the maldistribution of food
supplies. This requires some elaboration.

5.1 ‘Overt’ and ‘silent’ food problems

The conventional explanation of hunger focuses on the supply of food

and physical food scarcity as signified by an average or per capita
shortfall of supplies below consumption requirements. This has been
termed the ‘overt’ food problem (Reutlinger, 1977). But, by assuming
that members of the relevant population group are either all hungry or
all adequately fed, this explanation overlooks the possibility of a
proportion being adequately fed whilst the remainder are hungry due
to unequal distribution of the available food supply. Reutlinger calls
this the ‘silent’ food problem.
The key to ‘who goes hungry’, regardless of whether aggregate
food supplies are ‘scarce’ or ‘plentiful’, is the distribution of effective
demand, or what Sen (1981) calls ‘exchange entitlements’. Any
economic change causing a reduction in per capita real income, such
as reduced working hours, unemployment, or a sharp rise in the price
of food or other necessities, will tend to ‘squeeze’ the consumption of
low income households. From simple comparative statics analysis it
 POPULATION AND FOOD SUPPLIES 267

can be deduced that, ceteris paribus, some reduction in per capita

food consumption is the expected consumer response to any real food
price increase. But, provided that the price-elasticity of demand is
lower for ‘rich’ than for ‘poor’ families - a seemingly reasonable a
priori assumption - the scale of the cut in per capita consumption will
be inversely related to family income. Whereas rich families may
scarcely need to curtail their food consumption, some poor families
will be virtually priced out of the market.
The reason for a rise in the food price may be a rightwards shift in
the aggregate demand schedule rather than a leftwards shift in
aggregate supply. For example, speculative hoarding by high income
families cduld produce this effect. Whether the expected shortage
actually materialises or not, aggregate demand must increase in the
short term as the hoarders build up their stocks, as well as buying for
current consumption, and low income families again suffer from
reduced food consumption as the price rises. But poor families can
also be forced to curtail food consumption despite food price
stability. Thus any economic misfortune which reduces the amount of
family income, such as falling wages or rising unemployment, will
induce a leftwards shift of the demand schedule in the poor families
sub-sector of the market. Ceteris paribus, this could be expected to
induce a parallel shift in the corresponding segment of the aggregate
demand schedule, and some consequent price reduction. However,
the effect of this price reduction on family real income may only be
marginal in relation to the initial income loss. Of these three
situations in which low income families suffer a reduction in
food consumption, only the first, in which aggregate food sup¬
plies are curtailed, conforms to the conventional explanation for
hunger.

5.2 Famine

Carried to its extreme, hunger becomes famine and people die from
starvation. The causes of famine have been analysed by Sen (1980a)
who uses the term ‘food availability decline (FAD)’ to signify the
most usual and obvious explanation. That is, people starve because
per capita food supplies are suddenly reduced to the starvation level
by crop failure or some other event which greatly reduces aggregate
food supplies below their normal level. But Sen finds that when the
facts surrounding certain major famines are examined, FAD emerges
either as an incomplete explanation, or even as a wholly un¬
convincing one.
268 AGRICULTURE AND ECONOMIC DEVELOPMENT

An alternative explanation of famine is that those who starve lack

the means to obtain sufficient food to remain alive. Sen terms this the
‘entitlement failure’ explanation. Entitlement failures are of two kinds:
direct entitlement failures, relating only to food producers, in which a
peasant farmers’ crop fails and he is unable to feed himself either by
self-supply or by trade; and trade entitlement failures, which relate to
those who can normally afford to purchase a sufficient quantity of
food, but are priced out of the market by a sudden increase in
demand. In a ‘slump famine’ direct entitlement failure predominates,
whereas trade entitlement failure is the overriding factor in explain¬
ing a ‘boom famine’. But, despite its theoretical elegance, the
exchange entitlement failure explanation of famine suffers from being
much less amenable to measurement than FAD.
It is important to appreciate that the FAD and entitlement failure
explanations of famine are not mutually exclusive. Sen cites the
notorious Irish potato famine of the 1840s as an example of slump
famine. Supplies of other foodstuffs, such as grain, were available in
Ireland during the famine despite the devastating failure of the potato
crop. But, because of the dominance of the potato in the accustomed
diet of Irish peasants, the obliteration of the crop by potato blight did
represent a substantial FAD, as Sen himself admits. Indeed, FAD
may be the dominant cause of some famines. At the same time, the
dire consequences of crop failure in Ireland were greatly exacerbated
by government’s failure to organise more effective measures of
famine relief. In Sen’s phraseology, such measures could have
done much to offset the Irish potato producers’ direct entitlement
failures.
Thus, explaining the Irish potato famine appears to combine FAD
with slump famine conditions brought about by the devastating loss
of income suffered by potato producers. Sen cites the Ethiopian
famine of 1974, which was concentrated in one province of the
country, as an example of a ‘pure’ slump famine, i.e. famine in a
specific geographical area, due to massive entitlement failure for the
agricultural population, without substantial FAD in the country as a
whole. In contrast to the Ethiopian famine, the Bengal famine of 1943
is cited as an example of boom famine. The statistical record indicates
that per capita food availability in Bengal during the famine period
was not exceptionally low. But the real incomes of agricultural
labourers, the prime victims of the famine, were severely curtailed by
food price inflation and increased unemployment induced by a
combination of adverse weather conditions and disruption caused by
India’s involvement in the war with Japan.
 POPULATION AND FOOD SUPPLIES 269

5.3 Amelioration of malnutrition: policy options

Analysis of the causes of malnutrition in LDCs indicates at least

three possible policy approaches to ameliorating the problem:

(1) lowering the population growth rate;

(2) augmenting food supplies (domestic and/or imported); and
(3) redistributing available supplies from the adequately or ‘over¬
fed’ to the ‘under-fed’

5.3.1 Population Control

A comprehensive analysis of population control is beyond the scope
of this book and we confine ourselves to a brief examination of
possible reasons for the persistence of high fertility rates in LDCs,
particularly in rural areas or where agriculture predominates.
Whereas crude birth rates in low income countries still average about
30 per thousand population, compared with only about 15 per
thousand in industrial market economies, the crude death rates
appear to be virtually the same, on average, at 10 or 11 per thousand
population (World Bank, 1981, Table 18).
Fertility appears to decline with increasing urbanisation. But since
this, and other socioeconomic trends favourable to falling fertility,
such as rising literacy and declining mortality and poverty, tend to be
interrelated, it is difficult to isolate the effect of any one of them
(Birdsall, 1980, p. 42). The urban-rural fertility differential hypo¬
thesis states that, other things being equal, fertility in LDCs is lower
in urban than in rural areas. The reasons postulated for the
differential include not only higher urban incomes and a wider choice
of consumption goods in urban areas, but also more advanced
standards of education and superior family planning services. The
empirical evidence is mixed. The most convincing evidence of a
significant differential is provided by Latin America where, during
the 1950s and 1960s, fertility was lower in urban than in rural areas in
practically all countries. The evidence is less convincing for other
continents, and particularly in Africa where ‘comparable and reliable
measures of fertility differentials are generally lacking’ (United
Nations, 1977, p. 284). However, this conclusion is as much based on
an absence of reliable data as on positive evidence to disprove the
hypothesis.
The causation and duration of urban-rural fertility differentials
remains obscure. The general relationship between fertility and
economic growth seems clear enough, i.e. fertility declines as per
270 AGRICULTURE AND ECONOMIC DEVELOPMENT

capita income rises. Thus urban-rural fertility differentials may

reflect no more than the expected transition from high to low fertility,
which is part of the development process, with the urban population
merely leading. Long period demographic evidence from some now
developed countries, like Japan, suggests that although rural fertility
remains higher initially, the gap eventually closes (United Nations,
1977, p. 290).
A high rural fertility rate probably reflects the relative poverty of
farmers - and, a fortiori, landless labourers - as well as lower
literacy, but it may also reflect a stronger economic motivation to
invest in children as potential contributors to the family income. The
economic motivation for family formation combines consumption
with investment aspects, as discussed by Cassen (1976). On the
consumption side, the costs of educating and rearing children, plus
the opportunity costs of parents’ time and earnings possibly forgone,
have to be set against the psychic income derived from parenthood.
On the investment side, the potential ‘yield’ includes not only a direct
contribution to family income when the child is old enough to work,
but also relief labour to cover parents’ temporary inability to work
(due to illness or accident) and the support of parents in their old age.
On balance, Cassen appears to be dubious whether, on a strictly
economic calculus, children do typically represent a good investment
prospect for poor households in LDCs. Relevant empirical evidence
is extremely meagre, but a study in Bangladesh (Eberstadt, 1980)
pointed to the relatively optimistic conclusion that the average village
boy can be credited with the triple achievements of earning enough to
pay for his food by the age of 12; paying for all his past consumption
by the age of 15; and paying for the consumption of his sister as well
by the age of 21.
We conclude that lowering the rate of population growth, by
means such as improved family planning services, is probably a viable
approach to improved nutrition in LDCs, especially in the long term.
In the shorter term, the decline in fertility may be retarded by the
perceived economic advantages of a larger-sized family, particularly
in agriculture.

5.4 Augmenting food supplies

This policy option is discussed at length in chapters 5,6 and 7. Suffice

it to repeat here that the constraints on expanding domestic food
production in LDCs are complex and include institutional rigidities
which are very resistant to rapid change. Thus, like population
 POPULATION AND FOOD SUPPLIES 27)

control, this approach is too long-term for dealing effectively with the
most pressing aspects of the problem of malnutrition.

5.5 Redistribution

For the reasons already analysed in this chapter hunger is primarily a

function of poverty. If aggregate food supplies are inadequate for all
to have sufficient for survival, the relatively well-off are much less
likely to starve than the poor. And if, as is much more likely,
aggregate food supplies are adequate to meet consumption needs
given equality of distribution, the poor remain exposed to the threat of
starvation.
In an economy consisting exclusively of poor people all might be
hungry and the problem of relieving hunger could be resolved only by
raising incomes all round. But, coming closer to reality, where the
distribution of income is such that only a proportion of households
are ‘poor’ whilst the remainder are ‘rich’ - i.e. relatively better-off it
may be feasible to relieve hunger by redistributional measures alone,
i.e. without raising the aggregate or average level of income. Such an
approach is open to the critical argument that the best means of
curing poverty (and hunger) is by maximising the rate of economic
growth in order to make everybody better off. There are two decisive
arguments against this criticism and in favour of deliberate redistri¬
butional measures to ameliorate hunger. First, due to the uncertain
but most likely unequal distribution of its benefits, economic growth
does not necessarily cure poverty. Secondly, in any case, the ‘trickle
down’ approach to the relief of hunger is too slow.
Flaving opted, in principle, for redistribution as the most effective
short-term method of relieving hunger, government remains con¬
fronted by two major problems: a) identifying the hungry, and
b) devising the most effective means of raising their per capita food
consumption, particularly in the most needy or ‘target’ groups. An
obvious and commonly used means of identifying the hungry is the
‘income approach’, assuming that there is a direct and reasonably
close correlation between per capita food consumption and house¬
hold income. Provided appropriate data are available, this relation¬
ship can be estimated statistically, most usually in the form of data
showing average levels of per capita consumption by household
income group. The results of empirical observation in a number of
countries confirm the expected systematic increase in average calorie
availability per capita with rising household income (or household
expenditure) (FAO, 1977, Tables ii 1.3-1.7).
272 AGRICULTURE AND ECONOMIC DEVELOPMENT

The same approach might be used in order to test for a systematic

difference in per capita food consumption between urban and rural
households. A food consumption differential in favour of urban
households might be postulated a priori because of the income
differential which commonly exists in favour of those in urban
occupations. We are unaware of empirical evidence with a bearing on
this hypothesis. But available evidence does suggest that at the same
or similar levels of household income per capita food consumption
tends to be higher in rural households. This is possibly because
whereas farm households usually possess self-produced food sup¬
plies, poor urban households typically have much less access to the
resources required for subsistence food production. Urban house¬
holds are consequently more dependent on cash income to purchase
food. Lacking access to land, rural landless labour households are
also deprived compared with farm households.
A possible alternative to trying to relate actual calorie intake to
perceived nutritional needs is to attempt to observe the point where
households endeavour to improve the quality of their diet rather than
increasing its quantity (Poleman, 1981). This approach is based on
the principle that, as a nation becomes richer, the composition of its
diet shifts from almost complete reliance on bulky foods rich in
carbohydrates, such as root crops and cereals, to the adoption of a
more varied diet emphasising less bulky foods with more protein.
Bennett’s potato/wheat substitution model, based on historical
changes in the diet of American immigrants and their descendants, is
a well-known application of this principle (Bennett, 1954). Poleman
cites more recent developing country evidence, from North-East
Brazil for example, suggesting that very poor people do in fact
substitute more preferred for less preferred starchy staples as
opportunity arises rather than merely consuming more calories.
A major weakness of the household income and consumption
approach to identifying the hungry is that it conceals the facts
concerning food distribution within the family. Yet it is known that
young children and pregnant and lactating women are especially
vulnerable to protein-energy malnutrition (PEM). Accurate in¬
formation on the extent of worldwide PEM is lacking though, on
the basis of various approximations and assumptions, Poleman
estimates that, in 1975, between 62 and 309 million women and
children might have been ‘at risk’. PEM is related to bad sanitation
and health factors, as well as to inadequate food intake as such, so
that the ‘synergism of malnutrition and infection is difficult’ to
 POPULATION AND FOOD SUPPLIES 273

separate into its component elements’ (FAO, 1977). But bad health
and low food consumption are both linked to poverty.
Reasonably accurate measurement of the extent of malnutrition in
a country, and proper identification of the most needy malnourished
‘target groups’ are important objectives of nutrition policy. If the
extent of malnutrition is overestimated or exaggerated not only does
the problem appear more intractable than it really is, but scarce
resources are wasted in giving relief to people who do not need it
(Srinivasan, 1981).
Because of the close link which exists between poverty and hunger,
a general redistribution of income in favour of the poor represents
one approach to the amelioration of hunger. In principle, a
choice of redistribution methods exists. First, there is the ‘socialist
solution’ of guaranteed employment for all who are able to work,
even though this entails the loss of the individual’s freedom to choose
the type and place of employment, as in China.2 Second, there is the
possibility of a guaranteed income for all citizens through social
security benefits, as in rich ‘capitalist’ countries, despite the cost to the
national budget and the political resistance of vested interests. Third,
there is the pursuit of full employment with adequate minimum
wages for all, within the context of a free enterprise or ‘mixed’
economy, despite the well-recognised problems of achieving and
maintaining steady growth in the face of inherently unstable market
forces.
In spite of the formidable difficulties of achieving significant
redistribution of income by deliberate government policy, particular¬
ly in market-type economies, evidence is accumulating to show what
can be achieved, even in LDCs. So, whereas in Taiwan and Sri Lanka,
for example, economic growth has been successfully combined with
greater equity in the distribution of income, in other LDCs, such as
the Philippines and Brazil, growth has been accompanied by
increasing income concentration (Fields, 1980; Sen, 1980b). Thus,
although the growth and distribution record of some LDCs appears
to confirm the well-known Kuznets curve hypothesis (that, during the
process of economic growth, declining income concentration is
preceded by a phase of increasing income inequality) the evidence is
by no means uniform. Some developing countries have actually
achieved ‘redistribution with growth’ (Chenery et ai, 1974). However,
as a means of alleviating hunger, the general income redistribution
approach suffers from the twin problems of its relative slowness and
its failure to deal directly with the nutritional problems of especially
274 AGRICULTURE AND ECONOMIC DEVELOPMENT

vulnerable groups such as young children and pregnant and nursing

mothers.
More direct approaches to stimulating food consumption of low
income families and especially vulnerable groups include at least
three possibilities. First, there is nutritional education to encourage
the purchase of high-calorie foods at minimum cost. This is difficult to
put across effectively, particularly to the poorly educated and, in any
case, provides no more than a partial solution to the problem. In
many families income may still be too low to buy nutritionally
adequate amounts of even the cheapest types of foodstuffs. A second
possibility is for the government to grant food price subsidies, as
exemplified by India’s ‘fair price’ shops. This approach is limited by
the problem of effectively restricting the supply of low price food to
low income families: leakages to non-preferential consumers tend to
make such programmes very expensive. The third option is special
target group food distribution measures, such as the US Food Stamp
programme. Although the US is not a developing country, there
seems to be no reason, in principle, why LDCs should not adopt this
approach. The essential feature is that low income households must
be identified and issued with coupons which can be exchanged only
for a fixed amount of food. The coupons may either be given free to
qualified consumers or sold to them at less than face value. Such a
system is clearly open to abuse in that some stamps may get into
unauthorised hands: but no scheme is completely foolproof. A more
serious objection is that the distribution of coupons benefits only
those consumers who purchase food in the market. In particular,
low income subsistence farming households are not catered for. A
different approach, emphasising increased farming productivity, is
needed for these. From a strictly nutritional point of view, a main
advantage of food stamps compared with open-ended food subsidies
or straight income transfers to target groups, is their cost-effective¬
ness, i.e. the programme cost per unit of extra food transferred to
target group consumers is likely to be lower. A simple theoretical
model supports this conclusion (Reutlinger and Selowsky, 1976). But
a food stamp programme imposes a substantial additional admini¬
strative burden on government, and this may be one reason why this
approach to the relief of malnutrition has not been more widely
adopted in LDCs.
In practice, some combination of policy measures to combat
malnutrition in Third World countries, including both short-term
and long-term instruments, may best serve national welfare. So, for
 POPULATION AND FOOD SUPPLIES 275

example, an employment-oriented programme of economic growth

(long-term) might be combined with a food stamp programme to
give some immediate relief to the most needy families (short-term).

Appendix: Formalisation of
Boserup’s Model

Consider the Cobb-Douglas production function:

q - aLa, 0^a^ 1

where q = agricultural output, L = agricultural population (with

dependency ratio assumed constant), a = a constant reflecting a given
state of agricultural technology, and a = elasticity of output w.r.t.
labour (L). The amounts of all non-labour inputs are assumed to be
fixed.
Discrete time can be introduced by attaching numerical sub-
scripts to variables. Thus:

qi = aiM... (1)
and
q2 = a2La2... (2)
It is a condition of Boserup’s model that:

q2 _ l2
(3)
qi V"
Proceeding now to re-expressing condition (3) in terms of (1) and
(2), we deduce (from (3)) that:

q2=Y:i-qi--- (4)
Li

and substituting (1) in (4) yields:

q2=T 'aiL®... (5)

Substituting (2) in (5), we have:

a2L2 = r—‘aiLi ••• (6)

276 AGRICULTURE AND ECONOMIC DEVELOPMENT

from which:

(7)

and finally, by substituting (7) in (2):

~T (1 -«)
(8)

Equation (8) merely expresses the mathematical relationship

between the ‘required’ level of output, q2, and given values of
other model variables and parameters. It is not an explanatory
relationship in the sense of explaining how target output q2 is
achieved. The technological shift factor is readily isolated from
(7), i.e.

(9)

But this states no more than that technological state a2 is sufficiently

‘better’ than ax to enable the proportionate increase in labour of
L2/Lx to raise output by an equivalent ratio, despite the constraint of
a. The means whereby better technology raises output in the required
proportion is not specified.

References

Bennett, M.K. (1954), The World’s Food, New York.

Birdsall, N. (1980), Population and Poverty in the Developing World,
World Bank Staff Working Paper no. 404, World Bank, Washington
DC.
Boserup, E. (1965), The Conditions of Agricultural Growth, George
Allen & Unwin, London.
Boserup, E. (1975), ‘The impact of population growth on agricultural
output’, Quarterly J. Econ., LXXXIX (2).
Cassen, R.H. (1976), ‘Population and development: a survey’, World
Development, 4 (11/12).
Crosby, Jr., A.W. (1973), ‘New world foods and old world de-
 POPULATION AND FOOD SUPPLIES 277

mography’, in Crosby, Jr., A.W., The Columbian Exchange,

Greenwood Press, Westport, Conn.
Eberstadt, N. (1980), ‘Recent declines in fertility in LDCs and what
population planners may learn from them’, World Development 8(1).
FAO (1977), The Fourth World Food Survey, FAO, Rome.
Fields, G.S. (1980), Poverty, Inequality and Development, Cambridge
University Press, Cambridge.
Griffin, Keith (1977), Inequality, Effective Demand and the Causes of
World Hunger, Queen Elizabeth House, Oxford.
Hayami, Y. and Ruttan, V.W. (1971), Agricultural Development: An
International Perspective, Johns Hopkins University Press,
Baltimore.
Langer, W.L. (1963), ‘Europe’s initial population explosion’, Amer¬
ican Historical Review, 69 (1).
Levi, John (1976), ‘Population pressure and agricultural change in
the land intensive economy’, J. Dev. Studies, 13 (1).
Nicholls, W.H. (1963), ‘An agricultural surplus as a factor in
economic development’, J. Political Economy, LXX1 (1).
Poleman, Thomas (1981), ‘A reappraisal of the extent of world
hunger’, Food Policy, 6 (4).
Reutlinger, S. and Selowsky, M. (1976) Malnutrition and Poverty,
World Bank Occasional Paper no. 23, World Bank, Washington DC.
Reutlinger, S. (1977), ‘Malnutrition: a poverty or a food problem?’,
World Development, 5 (8).
Sen A.K. (1980a), ‘Famines’, World Development, 8 (9).
Sen, A.K. (1980b), Levels of Poverty: Policy and Change, World Bank
Staff Working paper no. 401, World Bank, Washington DC.
Sen, A.K. (1981), ‘Ingredients of famine analysis: availability and
entitlements’, Quarterly J. Economics no. 384.
Schultz, T.P. (1969), ‘An economic model of family planning and
fertility’, J. Political Economy, 77: 153-80.
Srinivasan, T.N. (1981), ‘Malnutrition: some measurements and
policy issues’, J. Dev. Econ., 8 (1).
United Nations (1977), Levels and Trends of Fertility throughout the
World, 1950-1970, United Nations, New York.
Wilkinson, R.G. (1973), Poverty and Progress, Methuen, London.
Woodham-Smith, C. (1962), The Great Hunger, Hamish Hamilton,
London.
World Bank (1981), World Development Report 1981, World Bank,
Washington DC.
278 AGRICULTURE AND ECONOMIC DEVELOPMENT

Notes

1. A further aspect of ecological disequilibrium is damage to the natural

environment itself, such as soil erosion caused by the cultivation of hillsides,
de-afforestation and over-grazing with livestock.
2. Another example is West Bengal where the (Marxist) government has made
‘food for work’ available to the unemployed as a practical means of
redistributing income in favour of the poor by mobilising surplus labour.
 10 Agriculture and
International Trade

1 Introduction

Trade in agricultural goods can play a significant role in promoting

economic development of LDCs. First, exports of agricultural goods
can pay for imports of capital goods, technology, manufactured
products and other essential commodities for a sustained economic
growth of developing countries. It is important to stress that exports
of agricultural goods from LDCs like Thailand and Malaysia have
helped them considerably to earn valuable foreign exchange for their
industrialisation and economic growth.
Second, many LDCs have a comparative advantage in the
production of agricultural goods. Given a trade regime which is
relatively free from control and regulations, LDCs can use their
comparative advantage in producing agricultural goods to raise their
standard of living. Indeed, in an export-led growth model of trade, it
would be to the advantage of many LDCs to especialise in the
production of those goods where they have a comparative advantage
and to export the surplus production. Such a policy will lead to the
use of trade as an engine of growth apart from ensuring a ‘rational’
allocation of resources.
Third, even when a developing country successfully raises it
standard of living, trade in agricultural goods could still remain an
important policy for a number of key industries. Witness the case of
Canada, New Zealand, Ireland, Denmark and even the United States
of America - in these rich countries, agriculture still performs a very
important role as a major export industry for stimulating economic
growth.
Fourth, a growth of agricultural trade where LDCs can play an

279
280 AGRICULTURE AND ECONOMIC DEVELOPMENT

increasingly major role could also be of considerable advantage to the

DCs. As the income in the LDCs grows with a rise in their exports of
agricultural goods, there will be a corresponding rise of the demand
for industrial goods and different types of services in these countries
which would be imported from the DCs. In effect, the market for DCs
in the LDCs will expand and trade in agricultural goods will be of
mutual benefit to both DCs and LDCs. However, we are assuming a
trade regime which is free from all types of controls and regulations.
In practice, DCs try to protect their agricultural sector by the
imposition of different types of controls for keeping out imports.
Such controls usually take the form of quotas whereby only a
fixed amount of goods from LDCs are allowed to be imported. Also,
on a number of occasions, DCs (e.g. France and the EEC) have im¬
posed levies on the imports of agricultural goods. It may be pointed
out that LDCs also use tariffs, quotas and other exchange control
regulations to protect their ‘infant’ industries from foreign compe¬
tition. Some argue that the imposition of these controls like tariff and
quotas work against the principles of free trade which is supposed to
enhance economic welfare by providing goods at the cheapest price.
Developing countries point out that while the DCs preach the virtues
of the theory of free trade on the basis of comparative cost advantage
in the production of industrial goods, they themselves do not obey
these rules while they trade in agricultural goods with the LDCs. A
similar point has been made with respect to trading in textile and
manufactured products. We shall examine these issues more carefully
later.

2 Main Features of Trade in Agricultural Goods

Most LDCs trade in a large number of agricultural goods which have

the following important characteristics.

2.1 Commodity concentration in trade

This feature implies that a large number of LDCs tend to export a limi¬
ted number of agricultural goods. For instance, Brazil tends to depend
considerably on exporting coffee; in the case of Ghana, the major
export is cocoa; while Zambia’s chief export is copper. Bangladesh
and India depend substantially on the export of jute; India, Sri Lanka
and Kenya depend considerably on the export of tea; rubber is one of
 AGRICULTURE AND INTERNATIONAL TRADE 281

the few major exportables for Malaysia; Argentine and Thailand

depend very much upon exports of food grains. In other words, trade
is heavily dependent upon the exports of a few major agricultural
products.

2.2 Geographic or market concentration in trade

Geographic or market concentration implies that most of the major

exportables are usually sold in a few markets of the industrialised
countries. Examples are the sale of tea, coffee, jute, cocoa, rubber in
the markets of Europe, North America, Australia and Japan. Such a
geographic concentration has the implication that the economic
fortunes of the LDCs are strongly related to the rise of fall of the
domestic and economic activities of some industrialised countries.
Let us assume that India and Sri Lanka depend significantly upon
exports of tea to the British market. Now, if there is severe recession
in Britain for any reason, then the market for Indian and Sri Lankan tea
will also be adversely affected.

2.3 Fluctuations in primary commodity prices

It has been argued that the degree of fluctuations in trading of

agricultural and primary goods is significantly high or at least higher
than fluctuations in prices of manufactured goods which are traded in
the world market. The figures for Brazilian coffee illustrate the point:
monthly fluctuations in prices in the New York trading market is very
considerable. Many other agricultural goods which are exported by
the LDCs show similar fluctuations. Given such fluctuations, it is
easy to understand why export earnings (i.e. price x quantity of
exports) of LDCs will also fluctuate under certain assumptions.
The point has been made that such fluctuations in export earnings
are detrimental to the economic growth of LDCs. The theory is quite
easy to illustrate. Assume that the producers of primary goods are
generally risk-averse. If risk can be approximated by an index of
instability of export earnings, then a high degree of instability in such
earnings will induce producers to withdraw resources from the
production of such goods and, under such a circumstance, output
and employment will fall. If the producers were planning to commit
investible resources in the production of those goods whose earnings
oscillate substantially, then again production plans may be cut back
and an increase of employment will not take place, and once again,
economic growth will be stifled.
282 AGRICULTURE AND ECONOMIC DEVELOPMENT

Primary commodity prices may also fluctuate in the export market

due to demand factors, two of which deserve special emphasis:

(i) The income elasticity of the demand for agricultural products is

usually less than unity. This implies that 10 per cent rise in the level of
income will raise the demand for the product by less than 10 per cent.
Hence with economic growth and a rise in per capita income, the rise in
the demand for agricultural products will tend to decline in pro¬
portional terms. This will have an adverse impact on the export
earnings of the LDCs.
(ii) The price-elasticity of the demand for agricultural product is
usually less than unity. Thus, an increase in supply of agricultural
products in the international market will reduce the price. But the
additional demand generated by a decline in price of agricultural
exports will bring less revenue than before due to price inelasticity of
products. Once again, the LDCs will be hard hit if their combined
effort to expand production (following the principle of comparative
advantage) brings down the price by such an extent that total
earnings will actually fall.

The income and price inelasticities of agricultural exportables seem

to suggest that the LDCs are in a ‘Catch 22’ situation where they
cannot win! One argument that can be advanced to escape the
impasse is to raise rather than to reduce prices of primary exports
because if the demand is price-inelastic, then an increase in price will
raise export revenue. Unfortunately, the LDCs are generally price-
takers rather than price-makers in the international market and as
such they cannot raise the price unless they can exercise monopolistic
control in the supply of a particular primary product, like oil. One of
the major reasons why the Organisation of Petroleum Exporting
Countries (OPEC) has been so successful in raising the export
earnings from oil is that the cartel of oil producers has been able to
collude in production and price planning to take full advantage of
their monopoly gains as a group. But as far as other primary products
are concerned, it is difficult to see how other LDCs can form such
cartels to create such monopolies and ‘exploit’ the market.

3 Trade Policies in Developed Countries and

their Impact on Agricultural Trade

It is generally acknowledged that some trade policies followed in the

DCs are not really conducive to the growth of exports of agricultural
 AGRICULTURE AND INTERNATIONAL TRADE 283

goods from LDCs. We shall demonstrate here the impact of two main
instruments of commercial policy which have been successfully used
by the DCs to keep out imports of agricultural goods from the LDCs.
The first instrument that has been generally used in the DCs is an
import levy. This has the effect of a tax on imports. The other
commercial policy has been export subsidy}
Consider first the case of an import levy. Figure 10.1 illustrates the
point of taxing imports very clearly. On the vertical axis, price level
(=P) is measured and quantity (=Q) demanded and supplied is
measured along the horizontal axis. The domestic demand is shown
by Dd and the domestic supply is given by Sd. The world price of the
product is given by OPw. At this price, domestic supply is given by
OSw and the domestic demand is ODw. Under the import levy
scheme (as it has been used by the Common Agricultural Policy
(CAP) within the EEC) a lower limit price is given by OPL. Let us say,
no country within the EEC will be allowed to import agricultural

Figure 10.1: Welfare cost of an import levy

284 AGRICULTURE AND ECONOMIC DEVELOPMENT

goods at a price below OPL from the LDCs. Within the EEC as a
result consumers are forced to pay OPL as the domestic market price.
Notice that when the domestic price rises from OPw (the world price)
to OPL (the market price), PlPw is the amount of levy that has been
imposed on the amount of import (= SWDW) within the domestic
market. The size of this levy will actually depend upon the farm price
policies of the EEC countries.
The impact of such a levy is easy to illustrate. Domestic supply rises
from OSw to OSL and domestic demand falls from Odw to ODL. The
decline in the balance of trade deficit is shown by SLDL instead of
SWDW. This may be regarded as the balance of payment effect of
import levy which is now ‘successful’ in keeping out the imports of
primary goods from LDCs. Production effect within the EEC is
clearly observable as output rises by SLSW.
However, the welfare cost of such an import levy is mainly borne
by the consumers. This is shown by the reduction of consumer
surplus by the area PWGH PL. It is not a cost for the community as a
whole because what the consumers have lost as a group, is partly
gained by the producers as a transfer payment (area A). The
governments also gain at the expense of consumers in tax revenue
equal to the area C (i.e. tax rate PLPWX amount of import i.e.
HlJlTN). However, the area B measures the inefficiency in domestic
production from producing SWSL more of domestic agricultural
goods. Such a production is inefficient as it costs more to produce at
home than abroad. Similarly, area D measures the consumption cost
and loss of welfare as domestic consumers are now forced to reduce
consumption by DWDL because of the import levy. Such ‘consump¬
tion inefficiency’ arises because consumers are now paying PLPW
more than the world price per unit of output consumed. In other
words, net welfare loss is given by B + D.
The other policy that has been frequently used within the
framework of CAP is an export subsidy to agriculture. This policy
has the same effects as levies. However, taxpayers should now be
prepared to bear the cost of subsidy. This is shown in Figure 10.2 by
the area which is equal to N + R 4- T. Note that the world price is OPw
but the subsidised farm price within the EEC is OPsu, while the
equilibrium price is OPe. The impact of an export subsidy is to
encourage domestic exporters to carry on exporting agricultural
output from the EEC until the home price is equal to foreign price
plus subsidy per unit of output.
Notice that with an export subsidy, net income of EEC farmers rises
by M+N + R. Although there is a rise in exports because of the
 AGRICULTURE AND INTERNATIONAL TRADE 285

Figure 10.2: Common agricultural policy and the impact of an export subsidy

subsidy, an increase in home price reduces home demand by DSsu

Dw because of the rise in prices to OPsu. The loss of welfare due to
‘consumption inefficiency’ is clearly shown by the area M + N. Also
notice that farm producers from LDCs are now compelled to face
extra competition due to subsidies in the world market.
It is clear from the above analysis that the use of protec¬
tionist policies for sheltering agriculture in DCs has almost
certainly reduced the trading welfare of the LDCs. As most LDCs are
generally exporters of agricultural products, the point has been made
that the imposition of trading embargos, even if they are indirect,
should be removed not only for the sake of LDCs, but also to remove
the types of‘consumption inefficiencies’ in DCs that we have just now
observed.
There is an additional problem about the trade in primary
products. It is often claimed that prices of primary products fluctuate
286 AGRICULTURE AND ECONOMIC DEVELOPMENT

quite sharply in the international market and such fluctuations are

regarded as detrimental to the growth and development of the
economies of LDCs. It is useful to mention that the problem of export
instability and economic development of LDCs has gained consider¬
able attention in the past and present, particularly in the context of
the North-South dialogue (.Brandt Report, 1981). In the next section,
we shall use simple economic theories to illustrate the welfare gains
from primary commodity price stabilisation.

4 Welfare Gains from Price Stabilisation

Within a comparative static model, it is easy to show welfare gains
from primary commodity price stabilisation with the following
simple
(i) The economy is competitive.
(ii) The demand and supply curves are linear and supply elasticity is
always positive while demand elasticity is always negative.
(iii) Either the demand curve or the supply curve shifts alternately
between two known parallel positions (the stochastic disturbances
can also be accommodated but are left out for the purpose of
simplification; for details, see e.g. Hailwood, 1979).
(iv) The disturbances are additive and affect only the intercept and
not the slope parameter. Thus shifts in the demand curve can be
regarded as the product of a ‘business cycle’ in the exporting country
and the shifts in the supply curve by a ‘production cycle’ in the
exporting country.
(v) Supply and demand respond instantaneously to price changes. In
other words, adjustment lags are assumed away.
(vi) The buffer-stock scheme can operate without any cost and it can
stabilise market price on a unique intervention price (once again, the
positive costs of operating buffer-stocks can be included, but are left
out to facilitate exposition).
In Figure 10.3, we show the movements of prices due to shifts in
demand (from to D2). The stabilising price P*stabilises the total
revenue which is equal to total earnings from exports regardless of
price elasticities of demand and supply. In Figure 10.3 OHGP* is the
stabilised total revenue. This is more stable than the oscillating free
market revenues at prices Pt and P2, i.e. OJIP! and OLKP2.
In Figure 10.4, we illustrate the case of price instability due to shifts
in supply from Sx to S2. Here, stabilisation may either stabilise or
 AGRICULTURE AND INTERNATIONAL TRADE 287

Price

Figure 10.3: Welfare gains from price stabilisation: demand shifts

Figure 10.4: Welfare gains from price stabilisation: supply shifts

288 AGRICULTURE AND ECONOMIC DEVELOPMENT

destabilise total revenue. Thus, if the price is stabilised at P*, total

earnings fluctuate between P*', MNO and P*' VQO. It is difficult to
say, a priori, whether this amplitude of movement in earnings will be
greater or less than the free market price fluctuations in revenue,
which is given by P'2RSO and Pf'TUO. All that can be said is that
given low price elasticity of demand, price stabilisation will also
stabilise total earnings when instability has been caused by shifts in
supply.
The proof is as follows:
Let R* = P*Q... (1)
where R* = total revenue at stabilising price, P* = the stabilising
price, and Q = quantity supplied. Differentiating with respect to Q,
we obtain:

^ = P*^=P* (2)
dQ dQ

Note that:
R = PQ... (3)
where R = total revenue without price stabilisation.
Differentiating (3) with respect to Q we get:

dR dP ^ dQ
dQ -dQ Q + PdQ

dP
Q+ P (4)
dQ

Multiplying and dividing equation (4) by P, we get:

dR
(5)
dQ

where tjd = dQ/dP • P/Q < 0 price elasticity of demand. If we divide

(5) by (2) we obtain:

dR dR* P
(6)
dQ dQ ~P*

Equation (6) may be regarded as the ‘stabilisation factor’

(Hailwood, 1979). The impact on total revenue earnings of stabilisa¬
tion price at P* will depend upon rjd. It is clear from the left-hand
 AGRICULTURE AND INTERNATIONAL TRADE 289

side of equation (6) that if the standard conversion factor > 1, price
stabilisation will also stabilise total revenue because R is more stable
with than without price stabilisation. When SCF < 1, price stabilisa¬
tion will lead to a revenue destabilisation. When the SCF = 1, the
effect on revenue is neutral. (Readers may wish to work out the
different kinds of relationships between price stabilisation and
revenue stabilisation by assuming different values of price elasticity.)
The analysis has so far been carried out in a micro context. It is
possible to highlight some macro issues regarding export instability
and economic growth.

5 Export Instability and Economic Growth:

some Macro Issues

It has been sometimes argued that a high degree of instability in

export revenue would usually mean a larger fluctuation in the income
of the export earners and as long as such income comprise a
significant proportion of the gross national product (GNP) of the
economy, the GNP would be subject to large oscillations. Many
argue that such an instability in expot income would have harmful
effects on the economic development of a country. If there is a fall in
export income, then there could be a ‘foreign exchange crisis’, and the
capacity to import (particularly food and capital goods) will be
adversely affected. Thus, there could be a serious balance of payment
crisis. A decline in the imports of capital goods will have an adverse
effect on the rate of growth of investment, which, in its turn, will
reduce the overall rate of economic growth.
So far we have assumed away the presence of uncertainty. It is
possible to take into account the role of uncertainty explicitly. Let us
assume that uncertainty in export income exists; let us also assume
risk-aversion by the producers, then, if the producers expect a large
fluctuation in their export income, they may divert funds from the
export to some other sectors of the economy where the return from
capital investment is expected to be fairly stable. This type of
speculation could clearly jeopardise the long-run growth of export
industries in the economy.
There is another problem with respect to wage rigidity. Assume
that wages are rigid downwards, then, it can be shown that a rise in
instability in export earnings could imply an increase in the rate of
inflation. The argument may be understood in a very simple way if it
is acknowledged that prices and wages tend to increase during a
290 AGRICULTURE AND ECONOMIC DEVELOPMENT

boom, but would not fall much during the ‘trough’. Note that many
countries, particularly LDCs, suffer from low supply elasticities. If
the Philips curve-type trade-off is assumed to have some validity in
the short run, then, a rise in the rate of inflation will affect
employment. Further, a high rate of inflation can adversely affect the
level of savings as people suffer the erosion in the real value of
financial savings. There could be a movement away from financial
assets with adverse consequences on the process of monetisation of
the economy and its beneficial effects.
Some economists have argued that sometimes export instability
can have a favourable impact on the economic growth of the country.
The argument runs as follows. Given a curvilinear consumption
function, the higher the income instability, the lower will be the
average propensity to consume and the greater will be the average
propensity to save. This proposition rests on Friedman’s theory of
consumption function where he regards consumption as a function of
permanent rather than transitory income. A transitory rise in income
will thus leave consumption unchanged and savings will rise. This is
supposed to raise the level of investment and the rate of economic
growth. In other words, instability in export income will favourably
affect the rate of economic growth.
It is important to remember that the above analysis does not take
into account the risk-aversion by the investors in the face of
instability in export earnings. Clearly, given uncertainty and risk-
aversion by the producers, investment may not always be positively
correlated with an increase in instability in export income.

6 A Survey of the Literature

The array of past studies can be broken down into three categories:
those that tried to explain export instability, testing to see whether it
was indeed related to such variables as commodity concentration;
those that tested whether any significant difference could be detected
in the degree of export instability experienced by developing and
developed countries; and those who explored the repercussions of
export instability in domestic economies. A review of past studies
must first include the work of Coppock (1962) who developed a log-
variance index of instability using the total exports of eighty-three
countries for the period 1946 to 1958. Coppock tried to explain the
different indices for countries by means of single and multiple
correlations. He found export earnings instability most closely
 AGRICULTURE AND INTERNATIONAL TRADE 291

related to instability in the volume and price of exports and imports,

and the terms of trade. He also found that instability was negatively
correlated with regional concentration and had a low positive
correlation with commodity concentration.
Michaely (1962) studied the relationship between export prices and
commodity concentration for thirty-six countries for the period 1948
to 1958. His results showed a significant positive relationship between
export price instability and commodity concentration. Massed used a
sample of thirty-six countries and data from 1948 to 1959. He
regressed instability of earnings on commodity concentration. Two
measures of each of the variables were used but no coefficients were
significant at the 5 per cent level. After a geographic concentration
index was added commodity concentration became significant. And
when the percentage of primary products of total exports (% XP) was
added as a variable to an equation including a variable for geographic
concentration, %XP was significant. The coefficient of the geo¬
graphic concentration index in this equation was negative and
insignificant. Massell’s results were inconclusive. His empirical
evidence did not support the hypothesis that concentration of exports
were significantly related to export earnings instability (Massed,
1964, 1970).
The data compiled by Coppock and Michaely was used by
MacBean (1966). His findings suggested that export instability
experienced by LDCs and DCs did not differ greatly, and tried to
explain the variance in instability between countries with three
variables: the ratio of primary products in total exports, and
commodity and geographic concentration indices. A11 these variables
succeeded in explaining only 25 per cent of the total variation of
export instability between countries. It can be shown that Coppock’s
log-variance (LVC) instability index was determined only by the first
and last observations in the sample.
2 11/2

or

or

r = ——r(logVn — log Vt)

n— 1

The log variance index formulated by Coppock was only one of

292 AGRICULTURE AND ECONOMIC DEVELOPMENT

four instability indices in MacBean’s study. Twelve countries

emerged as having unstable exports earnings by all four criteria. But
by further examining each of these countries, MacBean attributed the
instability to very specific local and political factors. Only five
countries exhibited instability due to the type of produce exported.
However these products proved to be primary products which had. a
history of instability. His general conclusion was that:

Such theoretical proposed general factors as specialisation in primary

products or commodity concentration per se may have some slight systematic
tendency to produce export instability but their explanatory value in
particular cases is very small.

MacBean also explored the influence of export instability on

investment in LDCs. He defined the rate of growth of investment as
the dependent variable and the rate of growth of import capacity (the
total value of exports plus net invisibles and net capital transfers,
divided by an index of import prices), the instability of importing
power as merchandise exports, and the rate of change (increase or
decrease) in reserves of gold and foreign exchange over the period as
independent variables. The instability of the importing power of
merchandise exports was insignificant at the 5 per cent, denying a
terms of trade effect. However the rate of growth of import capacity
was positive and significant. Evidence did not support the case that
export instability had a detrimental effect on investment.
Several other authors, including Erb and Schiavo-Campo, point
out the danger of generalising for all periods from evidence of export
earnings instability drawn from one time period. After computing
Coppock’s instability index for countries from 1954 to 1966, Erb and
Schiavo-Campo (1969) found the mean instability index for LDCs
was 13.4 with a standard deviation 6.2. Developed countries had a
means instability index of 6.2 with a standard deviation of 2.2.
Knudsen and Parnes (1975) constructed a transitory income index
of export instability, derived by dividing export income into perma¬
nent income and transitory income components. They proposed that
instabilities in export earnings could not be explained by simple
comparisons with the nature of exports, concentration indices, or
degrees of political instability. Knudsen’s and Parnes’ sample
consisted of twenty-eight countries and data from 1958 to 1968. Their
evidence suggested that there was no general relationship between
instability in export income and the type or concentration of exports.
Knudsen and Parnes also drew in interesting conclusion with regards
 AGRICULTURE AND INTERNATIONAL TRADE 293

to instability and investment. Empirical work revealed that lower

propensities to consume measured under higher levels of instability
resulted in increased aggregate investment, that is, instability stimu¬
lated investment.
Finally we can turn to Hailwood’s study (1979). For the purposes
of his study, Hailwood combined later work of Coppock with a study
done by UNCTAD. A new method had been used by Coppock to
compute an instability index for 109 countries and observations from
1959 to 1971. UNCTAD dealt with data from 147 countries and the
years 1950 to 1970 in five years sub-sets. Hailwood observed that a
marked difference existed in the export instability experienced by
developed and developing countries:

dispersion of export earnings instability of the LDCs was greater than that of
the DCs. Export earnings instability was on average 1.5 times higher for
LDCs with a per capita income below $250 than those with a per capita
income equal to or greater $250.

This conclusion seems to bear out the observation regarding criterion

for disaggregating developed and developing countries. Our review
of some of the literature concerning export instability can now be
concluded. The studies have not led to any definitive consensus on
differing magnitudes of export instability in DCs and LDCs or causes
of effects of instability.

7 Some Measurement Problems

Some other points can also be mentioned in connection with the

measurement of the instability indices. For example, Coppock uses
the ‘percentage increase in GNP per year’ adjusted for price changes.
MacBean, on the other hand, uses the compound annual growth rate
of GDP, 1950-51 to 1957-58 in current prices. But problems arise
because LDCs generally experience higher rates of growth of
population than the DCs and hence the use of the total GDP rates of
growth rather than the respective per capita growth rates introduces
an upward bias into the rates of LDCs. It should also be borne in
mind that since the LDCs usually experience a higher rate of inflation
than the DCs, the use of income data in current prices for measuring
the rates of growth of income implies an over-estimation of the
growth rates of LDCs. Coppock’s work seems to have suffered from
other methodological defects. Coppock uses export instability indices
294 AGRICULTURE AND ECONOMIC DEVELOPMENT

for one period (1946-58) while his data for income growth rates
covers a different period (i.e. 1951-7), (see Gleazakos, 1973).

8 Prebisch’s Hypothesis

Prebisch (1959) was one of the first economists who has drawn
attention to the problem of a secular downward trend in the terms of
trade of the LDCs (where the terms of trade are defined as the ratio
of export to import price paid by the LDCs in their trade with DCs).
Prebisch’s findings implied that the LDCs are losing out in their trade
with DCs and the only way to develop their economies is to resort to
industrialisation, protection and import substitution (for details, see
Ghatak, 1978). Free trade has failed to be a vehicle of economic
growth for the LDCs.
Prebisch’s conclusion has been criticised by many. It has been
shown that the final judgement about a net fall (or rise) in the terms of
trade against the LDCs is critically dependent on the choice of base
periods. Lipsey (1963), for instance, has shown that the terms of trade
has actually moved in favour of the LDCs vis a vis Europe and the
USA. On the other hand, Kindleberger’s study (1956) does not
confirm Lipsey’s findings.
In the 1960s, doubts have been cast by some empirical studies on
the assumed harmful effects of export instability and economic
development of the LDCs. The major conclusions (eg. MacBean,
1967) of these studies can be summarised as follows:

(a) The correlation between instability in export earnings and the

proportion of primary products in the export mix has been found
to be very weak.
(b) The correlation between the degree of primary commodity
concentration in exports and the instability in export income has
been very poor.
(c) The association between instabilities in export earnings and the
geographic concentration of the export market has been very
tenuous.

Studies during the 1970s have not, however, always confirmed the
1960s’ finding. Many recent studies have shown that instabilities in
export incomes for LDCs are significantly higher than for DCs.
Indeed, some of these studies report higher instabilities in export
earnings of LDCs even when the commodity boom of the early 1970s
has been excluded. In some recent empirical work, a high degree of
 AGRICULTURE AND INTERNATIONAL TRADE 295

positive correlation has been observed between export instability and

export concentration. Indeed, Glezakos (1973) concludes:

we feel that the empirical evidence presented here (in contrast to that
presented earlier by Coppock and MacBean) supports the a priori arguments
that export instability is generally larger in the LDCs than in the DCs and this
instability is detrimental to economic growth in the former but not in the
latter.

The adverse effects on trade on economic welfare has been well

articulated by Bhagwati. This phenomenon has been labelled as
‘immiserising growth’ and can be best explained with a diagram. In
Figure 10.5 let us measure industrial goods on the vertical axis and
agricultural goods on the horizontal axis. The domestic production
possibility curve is initially given by TT0, and the terms of trade by
the line AB. Initially the country produces at the point P and
consumes at the point C where the indifference curve it is tangent to
the price line AB. Exports and imports can be measured by PP0 and
P0C respectively.

Industrial
goods

Figure 10.5: Immiserising growth

296 AGRICULTURE AND ECONOMIC DEVELOPMENT

Suppose a LDC has a comparative advantage in the production of

agricultural goods and with trade its production possibility curve
shifts to T1T1. But the terms of trade of the country is now given by
ED. The country now produces at Pt and consumes at Cj as the T 1T'1
curve is tangent to the price line at Pj and the indifference curve is
tangent to the price line at Cj, respectively. Clearly, the country’s
welfare has been reduced as Ct is on a lower indifference curve
than C.
The following conditions must be satisfied for ‘immiserising’
growth to occur:
(a) It is imperative that the economic growth of a country should be
biased in favour of the export industry.
(b) Global demand for the home country exports must be inelastic.
(c) It is important that the economy of the home country should be
closely tied to international trade in the export commodity in
question.
(d) There should be substantial mobility of resources among differ¬
ent sectors.
It is possible to argue that all these conditions may not be satisfied
for the immiserisation effects of trade on growth to occur. Neverthe¬
less, it is difficult to reject the spirit of the model or its internal logical
consistency, particularly when one looks at the primary commodity
exporting LDCs and the decline in their terms of trade vis a vis the
DCs in the last twenty-five years. If it is difficult to reject the
possibility of immiserising growth, then the LDCs may be better-off,
as some argue, by withdrawing productive resources from the primary
agricultural commodity sector and channelling them into the indus¬
trial sector. Hence many, in line with the Prebisch-Singer hypothesis,
would like to pursue ‘inward looking’ import substitution in¬
dustrialisation policies for the long-run development of the LDCs via
protection. However, protection has its benefits as well as costs. (For
a good analysis of the impact of protection on LDCs, see, Bhagwati
and Desai, 1970; Little, Scitovsky and Scott, 1970; Balassa, 1971;
Ghatak, 1978)

9 The Agricultural Self-Sufficiency Argument

It has sometimes been argued that it is important for countries to

achieve self-sufficiency in food production. Such an argument has a
special appeal for the LDCs as many of them are heavily dependent
 AGRICULTURE AND INTERNATIONAL TRADE 297

upon imported food to meet excess home demand. An increase in

domestic food production will obviously mean a fall in food imports
and an easing of the balance of payment constraint. It will also imply
that a developing country can use foreign exchange to import capital
and other essential goods for domestic development. Further, a rise
in domestic food production will raise the level of nutrition where
malnutrition is a severe problem, assuming of course that a greater
volume of food production will be equally available to all. Such a rise
in nutrition level is supposed to raise labour efficiency and producti¬
vity. Some argue that self-sufficiency in food production is necessary
to make a country truly independent of foreign governments
internationalist policy. Hence we hear the slogan: ‘No country should
rely on the world market for its supply of food.’
In order to discuss these issues, it is important to define the term
self-sufficiency. Self-sufficiency (= SS) may be defined as the ratio of
domestic production (= PR) to domestic consumption (= C). Hence,
we have:
SS = PR:C

The SS ratio can be transformed into percentage terms easily.

However, this ratio can be misleading ideas about independence from
imports because of the use of imported inputs in domestic agriculture.
Thus the SS argument ‘involves not only the proportion of the
consumption of a particular product, but also the proportion of the
inputs used in the production of the product which are themselves
domestically produced’ (Ritson, 1977). Further problems arise in the
estimates of ‘overall SS’ and here it is important to find a common
measure to aggregate across commodities. In most studies, market
prices have been used as weights. However, the use of a composite
‘nutritional index’ or ‘calorie intake’ index could be quite fruitful.
The welfare effects of a rise in agricultural terms of trade can be
explained easily. In Figure 10.6, we measure industrial goods (I) on
the horizontal axis and agricultural goods (A) on the vertical axis. PP
is the production possibility curve and with TT j as the terms of trade
(assuming only two products, A and I, the exchange ratio between A
and I is also the terms of trade for that country) the country produces
at K and consumes at C. The country’s welfare is maximised when it
produces ON of industrial goods and OE of agricultural goods. The
country imports HE of food for QN of industrial products.
If the terms of trade rise in favour of A goods (for example, because
of sudden supply shock due to crop failure, war or flood) then the TT 1
will move to T0T2. Welfare clearly falls as the country consumes at J
298 AGRICULTURE AND ECONOMIC DEVELOPMENT

Figure 10.6: The welfare effect of a rise in agricultural terms of trade

(at a lower indifference curve i2). Another way of describing the loss
of welfare is to say that the country now needs to exchange OT3 of I
against OT of A goods (note that the slope of the terms of trade line
TT3 is the same as that of T0T2). Clearing the welfare of the food
importing country diminishes. In the next period, to optimise welfare
the food importing country should raise the production of A goods
(food) from OE to OF with a reduction of food imports from EH to
FG.
It seems that food importing countries should be better off by
introducing policies to stimulate domestic food production if it
expects a rise in global agricultural terms of trade. Imports may be
controlled to allow domestic price to rise, and supply should rise
given a normal supply response curve. However, free-traders argue
 AGRICULTURE AND INTERNATIONAL TRADE 299

that once the effects of the rising global food prices are felt by the
producers and consumers, appropriate adjustments will be made
without resorting to government interventionist policies. On the
other hand, there is no guarantee that the desired adjustments will be
made very quickly and, as such, there remains a case for an
intervention by government to promote desired adjustment.
Where a rising term of trade for food articles has been anticipated
and future benefits have to be set against present costs, government
may also intervene to affect land transfer. The problem here is that
land cannot be transferred easily from one market to the other (rural
and urban) with every turn and twist in the agricultural terms of
trade. It is thus necessary to take a decision regarding optimum land
use in agriculture with respect to a specific terminal period of
planning. Here planning techniques and sometimes cost-benefit
analysis could be used fruitfully (for a discussion of the cost benefit
analysis and planning, see chapter 11).

10 Cartels in Commodity Trade and

Welfare Gains and Losses

Following Viner, we can describe three major conditions for a

successful cartel agreement among the primary commodity exporting
countries:

(i) It is necessary for the cartel to control a very large proportion of

the total commodity supply in the global market. If this condition is
not met, demand curve which an individual cartel member faces
could be elastic even when the final demand from the consumer is
inelastic.
(ii) It is also important to have inelastic demand in the consuming
countries in both the short and long term. It is useful to remember
that higher commodity prices today may lead to the growth of the
substitutes in the long run.
(iii) In both the short and long term, the supply of relevant primary
commodities by the non-member countries should be rather inelastic
in both the short and long term, otherwise, the profits of the members
of the cartel will fall as non-member countries raise their supplies due
to a rise in prices.

Apart from these conditions, it is important to note that, for any

300 AGRICULTURE AND ECONOMIC DEVELOPMENT

cartel to hold, the political will to unite must be there. If any country
wants to follow an independent policy it may be tempted to sell more
than its quota and this may lead to a feud within the cartel and
eventually it may break up.
A policy of trade embargo may also raise welfare if a country’s
trade associates do not take up their choice to alter the terms of trade
back to their own advantage. For LDCs exporting tropical products,
this could be the case. Another way of looking at the situation would
be to accept the exploitation of monopoly power as a just weapon to
encourage development resource flows. Indeed, in the meeting of the
Commonwealth prime ministers and presidents in 1975, the follow¬
ing resolution was adopted: ‘We accept that the emergence of
producer’s associations is a reality born out of historical experience.
In an unequal world it is understandable that such a development
should take place’ (Commonwealth Secretariat, 1975).
It is necessary to point out, at this stage, that cartels and trade
restrictions could lead to welfare gains through market power
exploitations. But such welfare gains can diminish or even evaporate
if non-cartel countries retaliate. For example, Ritson (1977) observes
that during the 1960s, ‘the European Economic Community es¬
tablished minimum import prices for most agricultural products
which were very high relative to global market prices.’ The policy
reduced the level of EEC agricultural imports beneath the level which
would have applied with free trade and this undoubtedly had a
depressing effect on world agricultural prices compared with their
free trade equilibrium level. The result was that some food exporting
countries raised their subsidies on exports of agricultural goods and
this aggravated rather than reversed the initial movement of the terms
of trade.
The ‘inevitability’ of the collapse of a cartel has been questioned
recently (see Osborne, 1976). It has been shown that if members of a
cartel can devise a quota rule, then cheating by some ‘naughty’
members of the cartel may be prevented. The rule is very simple. If
any member of the cartel is found to be cheating, then members which
are not cheating can always retaliate by increasing their production
and profit. This will lead to a fall in the profit of the ‘cheating’
member. In other words, as long as the threat of retaliation persists
against the ‘cheating’ members, self-interest will maintain stability
within the cartel. However, the LDCs have now shown a marked
preference for an International Commodity Agreement (ICA)
scheme rather than for cartels in the interests of both producers and
consumers.
 AGRICULTURE AND INTERNATIONAL TRADE 301

11 Integrated Commodity Agreement (ICA) Schemes

In the United Nations Conference on Trade and Development

(UNCTAD), a Nairobi (1976) (known as UNCTAD iv), a resolution
was adopted for an ‘Integrated Programme for Commodities’ (IPC).
This resolution had three major components: aims, coverage of
commodities, and global measures of the programme.
Two important aims are: the stabilisation of global commodity
markets by ‘avoidance of excessive price fluctuations’, and the rise in
the real income that LDCs obtain from exports of their commodities
by ensuring a remunerative and a just price for the producers in LDCs
that take full account of world inflation and revenue stabilisation
‘around a growing trend’ value of earnings from exports.
As regards the coverage of commodities UNCTAD specified ten
‘core’ commodities: cocoa, coffee, copper, sugar, cotton, jute, rubber,
sisal, tea and tin. This list was supplemented by seven other goods:
bananas, bauxite, beef and veal, iron ore, rice, wheat and wool. The
first ten products are branded as ‘core’ commodities because they
account for about 75-80 per cent of the export earnings of LDCs of
all seventeen goods. Since these products can also be stored for
stockpiling, they are regarded as suitable for promoting a buffer-
stock scheme.
The heart of the IPC programme was the establishment of a
common fund to finance the entire project. Initial calculations show
that it would cost about US$6 billion and such an amount should be
raised by subscriptions from the exporters and importers of these
commodities. It was agreed that the least developed countries should
be exempted from paying the subscriptions. The argument for setting
up a common fund is to share and minimise risks and the principle is
the same as the one followed by any insurance company which seeks
to minimise its risks. Further, it is believed that the establishment of a
common fund will help LDCs to have more bargaining power vis a vis
the global capital markets than could a set of individual funds for
the same goods. ‘It would also require smaller financing than the
aggregate of a set of individual funds because of differences in the
phasing of cycles across commodity markets’ (Behrman, 1979).
It is important to look at the empirical evidence of the rates of
growth and fluctuations in prices and values of the UNCTAD core
and other commodities. Behrman has worked out rates of growth and
fluctuations in prices of such goods for a period that varies between
1953 and 1972 and these figures are cited in Table 10.1. Considerable
evidence of price fluctuations has been confirmed in this table, and
302 AGRICULTURE AND ECONOMIC DEVELOPMENT

Table 10.1: Secular trends in deflated prices for UNCTAD core commodities
UNCATD other commodities, and additional commodities of possible interest
1950-19751

Core commodities
Coffee -0.035 Bauxite 0.019
Cocoa -0.024 Iron ore -0.017
Tea -0.030 Maize 0.029
Sugar -0.0043 Tobacco -0.015
Cotton - 0.038 Lumber -0.005
Rubber -0.058 Hides and skins -0.030
Jute -0.018 Groundnut oil -0.033
Sisal2 — 0.0043 Olive oil -0.010
Copper 0.0043 Coconut oil -0.023
Tin 0.0043 Palm oil -0.028
Linseed oil -0.035
Other commodities oils4
Wheat -0.021 Soybean oil4 -0.015
Rice -0.0083 Cottonseed oil4 -0.016
Bananas -0.037 Palm kernel oil4 -0.006
Beef and veal 0.026 Lead -0.028
Wool -0.041

1 Calculated from UNCATD price indices in United Nations and OECD GDP price deflator.
2 1954-75.
3 Not significantly different from zero at standard 5 per cent level.
4 1954-74.
Source: Behrman, (1979).

clearly it suggests that a number of ICA concluded in the 1960s and

early 1970s have not succeeded in lowering instability. As a cross¬
check, if fluctuations in the World Bank index of primary commodity
terms of trade (based on unit values of LDCs’ exports of thirty-four
non-petroleum primary commodities) are observed between 1954
and 1975, then again Behrman’s finding of large fluctuations can be
confirmed. The figures in Table 10.1 are more instructive as they
reveal a disaggregated picture of price fluctuations of individual
commodities which could be concealed in an aggregated World Bank
index.
In Table 10.1. Behrman’s calculations of the secular trends in
deflated prices for UNCTAD core and other commodities have been
reproduced. The important conclusions which Behrman draws are as
follows:

(a) Negative secular trends have been generally observed for most
products except for copper and tin (non-agricultural products)
beef and veal (dairy products) and maize.
 AGRICULTURE AND INTERNATIONAL TRADE 303

(b) Behrman fails to observe any correlation between the amplitude

of price instability and the size of the secular trends in prices for
seventeen commodities as listed by the UNCTAD.
(c) Somewhat surprisingly, a positive correlation between income
elasticities and secular trends in prices has not been found.
Differential supply shifts (along with demand shifts) and different
supply price elasticities could have been responsible for such a
finding.

Under the auspices of the UNCTAD during the 1960s and 1970s,
several commodity cartels have been set up to promote the interests
of the LDCs, for instance, An International Sugar Agreement was
concluded in 1968. An International Wheat Agreement led to the
negotiation of the International Grains Agreement in 1967, which
then led to the establishment of an International Wheat Agreement
alone in 1971. Likewise, an International Tea Agreement was reached
in 1969 and was followed by similar agreements in sisal, cocoa and
henequen in 1970. It is useful to point out that the International
Coffee Agreement collapsed in 1972, but a new agreement on coffee
was reached in 1975. Eckbo (1975) observes that, for a sample of fifty-
one international commodity cartels, the mean duration of the
formal agreements has been 5.4 years and the median has been only
2.5 years - a finding which seems to confirm the view that a large
number of the International Commodity Agreement schemes cannot
be regarded as great successes.
One of the major reasons for such a lack of success has been the
failure to achieve price stabilisation. Law (1975) has shown that in the
case of coffee, the mean annual price changes has been about 50 per
cent greater during the years of agreement (i.e. 1964-72) than for the
pre-agreement years (i.e. 1950-63). In the case of sugar, according to
Law’s calculations, the mean price oscillations has been about 75 per
cent per annum larger for twelve recent years of price agreements
than during the eleven other non-agreement years. The cocoa
agreement failed when prices were pitched too high in 1970. It is
interesting to observe that in the case of rubber, the International
Rubber Agreements actually destabilised the market. It is only in the
case of wheat and tea that commodity agreements have led to an
increase in price stability - but here again, thanks mainly to stockpil¬
ing decisions made outside the agreements by Canada and the USA.
The other objective of UNCTAD IV has been to raise the export
income of LDCs by raising the prices of some selected commodities.
In some cases, these attempts have been successful. For example, in
304 AGRICULTURE AND ECONOMIC DEVELOPMENT

the case of coffee, recent agreements might have transferred funds

from consumers to producers at the rate of $500 to $600 million per
year. In the case of some other commodities, agreements have failed
to raise revenue, except in the very short run. The interesting point
about the commodity agreements that has emerged recently is that,
even for the ‘successful’ commodity agreements, the median duration
is only four years and the mean life is about 6.6 years. Behrman (1979,
pp. 64-5) has neatly summed up the reasons for a ‘successful’ rise in
prices in commodity agreement schemes as follows:

(a) high income elasticities of demand;

(b) high price elasticities of demand;
(c) high concentration of production;
(d) high concentration of foreign trade;
(e) a high share of foreign trade controlled by members of the
commodity agreements;
(f) low cost differences among producers in the agreement; and
(g) less intervention by the government.

The most common causes for the failure of the commodity agreement
schemes are supposed to be intense competition among the members
of the agreement schemes; and strong competition from those who
are not members of the agreement schemes.
The important question that has almost continuously been asked
in connection with the commodity agreement schemes is: Will the
gains in the short run outweigh the costs to be incurred in the long run
because of the growth of the substitutes and a rise in production from
‘fringe’ producers? The answer is probably yes, even when the goods
concerned do not always possess the same attributes which have
accounted for the short-run success of commodity cartels in the past.
As Behrman concludes, after a rigorous and thorough analysis,

even short-run market regulations may have a reasonable probability of

success only if the consumer nations can be persuaded to co-operate in order
to enforce discipline. The growing strength and cohesion of the developing
nations and the more accommodating posture of the developed nations
probably lead to a higher probability of at least short-run success for an
UNCTAD-type proposal than past experience alone might suggest. (Ibid.,
p. 67)

In Table 10.2 Behrman’s simulations of buffer-stock price stabili¬

sation for UNCTAD core commodities and basic foodgrains have
been shown. These simulations suggest quite clearly that the gross
 AGRICULTURE AND INTERNATIONAL TRADE 305

1 Details of the underlying econometric models are given in Behrman (1979). The buffer-stock simulations are identical to the base simulation except that the buffer stock purchases or sells
whatever is necessary to keep the deflated price within the indicated brand-width of the secular trend for 1950—74. The simulations are all for the decade. 1963—72. The initial and final stocks are
so1*
~ 4 2 rr — nh-O'Ot'O 'T'i ~
^ 2
*PC .25 *n
—
©
<«>
3!
<j a
m m <n r- 0 . ^t
I -
* 3
s
Q '2 •5 ^
^ r e ^ ~
■S 5 ^"5 2
£ O' || s© y'i O O m © c/~> r»
^l8 3 *5 . r- .-: CO r^i OS Os
“<3 a -S >5 I 2 I | r*** | o\ ©
R 00 0
fN «n
§ a «
a ~
c* ^ I I

^3 -2
^3 2-1.5 J
I 411 §
I ©lOOMOOtMON ^ ©
O
c:©« 33^ .©^2 2a — 00©—;©oo ©
s§ £ ^ ■§
O *> 5
V‘ CO
Q
2 1 «1*
u SllSal cN^n^t-m© — ©©> ©j
■o — — "fr ©!
I 1-a =--§ * ? v© — «— (N
<N —
© |

valued at the 1962 and 1973 prices, respectively, which overstates the present discounted values of buffer-stock operations.
* .§ -2 £§, §
^ ^ -2 Ov >G
1§> -3sli;
<3
<0 -3^ ^
5

2 Units are: thousands of 60 kg bags for coffee; thousands of long tons for cocoa; millions of pounds for jute.
3 bo r-©©^©u^cs — o
.§£ sd

§»
•S-a
^ c Ti’toor-O'nr'© p
5 5
I ^
•£ -o
a ^
,tg O

OC ~ (N © M MJ\
ri^vor^i^no-
r- — Tt <=> <n v>

s’
-c Vsr fN~p<N©<NCT\fN fN OO ©
* <3
© © 3 iri^da'dod'O ©— vd
£> »> JJ g - -||
.5 J I
-5-°
£ £

2^3 d> “3 sOfN-OO'OOO (N — vo

‘2 -Si 0-00000- © © ©
5© ©.
& |S
<U J- -3
to
II
b
s r--©oor*T©pON<N —■ — so
Tfdoddodd^ ©— »d
5 ~ I I
5
co
c
*0 'c3
O
feb
q :•= •O
o .
,0 1
A <■§ * § | _ g.
H 2 d i o 8 g = I | o.S K 6 ^£■
2 UUh^^wUf- o 3
O I i U w
306 AGRICULTURE AND ECONOMIC DEVELOPMENT

revenue gains to the LDCs from the operation of these specific

programmes may be large, though these revenue gains seem to vary
significantly across commodities. However, this is not a surprising
result given the difference in the basic characteristics of these
commodities.

12 The Compensating Financing Schemes

Another way of stabilising the foreign exchange earnings of LDCs is

to adopt a compensatory financing scheme. This idea gained
considerable attention during the 1960s and it has still retained some
of its appeal. According to this method, a country’s earnings from
foreign exports can be stabilised by setting up a stabilisation fund
from which countries should be entitled to compensatory drawings
for any shortfall in their export income whenever it falls below a
predetermined, mutually agreed level. The definition of export
income could follow the principle which has been adopted by the
International Monetary Fund, i.e. current account receipts in the
balance of payments. Alternatively, such income can be defined
according to the principle followed by the European Economic
Commission under the Stabilisation of Exports scheme (STABEX),
i.e. only commodity export income, or predetermined commodity
export income. However, problems cropped up in the late 1970s when
the benchmarks against which shortfalls in export income are
measured were fixed at a level which was considered too low. Clearly,
under such a rule the developing countries failed to earn substantial
accounts of foreign exchange. In the case of the IMF, the shortfalls in
export income needed for obtaining compensations were too large.
Also, conditions for drawing from the stabilisation fund were very
restrictive. In 1975, some of these severe restrictions were partially
eased. For instance, ‘forecasting restrictions’ (the clause that stipu¬
lated that exports in the post-shortfall year, which had to be forecast,
should not exceed 110 per cent of the nominal level of the two pre¬
shortfall years), have been rejected and the quotas for drawings have
been raised. However, given the limitations of the CFS, UNCTAD is
now advocating an Integrated Programme for Commodities under
which a buffer-stock will be established as an important part for
setting up a New International Economic Order. The problem,
however, is the financing of the buffer-stock scheme which will cost
about US $6 billion.
The other criticisms of the CFSs can be stated briefly. It has been
 AGRICULTURE AND INTERNATIONAL TRADE 307

argued that the commodity agreement schemes will intervene into the
free operation of the market and hence the efficiency in resource
allocation will be adversely affected. Also, an export quota system
will tend to protect high cost producers and result in a sub-optimal
level of production. Further, the costs of stockpiling some agricul¬
tural goods may be a lot larger than the benefits of such stockpiling.
In defence of the CFSs, it needs to be mentioned that the
international markets in agricultural products are not perfect and
there is also some imperfection in the flow of information between the
principal agents of the markets - producers, consumers and specu¬
lators. It is precisely because of this sort of imperfection that price
instability occurs and institutions like the CFS and STABEX are
needed to prevent market failures. Where it is very difficult to
manipulate the two principal instruments for stabilising commodity
export earnings, i.e. buffer-stock and export quota, it is necessary to
use the compensatory finance scheme to aid the LDCs.

References

Balassa, B. (1971), The Structure of Protection in Developing Coun¬

tries, Johns Hopkins University Press, Baltimore.
Behrman, J. (1979), Development, The International Economic Order
and Commodity Agreements, Addison-Wesly.
Bhagwati, J. (1959), ‘Immiserizing growth: a geometrical note’, Rev. of
Econ. Studies, 25, 201-5.
Bhagwati, J. and Desai P. (1970), India: Planning for Industriali¬
zation, Oxford University Press, OECD and Oxford.
Brandt, W. (1980), North-South: A Programme for Survival, Pan.
Coppock, J.D. (1962), International Economic Instability, McGraw-
Hill, New York.
Eckbo, P.L. (1975), ‘OPEC and the experience of previous inter¬
national commodity cartels’, Cambridge, Mass., Energy Laboratory
Working Paper.
Erb, G. and Schiavo-Campo, S. (1969), ‘export instability, level of
economic development and economic size of less developed coun¬
tries’, Bull. Oxford Inst. Econ. & Stats., XXI, November, 263-83.
Ghatak, S. (1978), Development Economics, Longman, London and
New York.
Gleazakos, C. (1973), ‘Export instability and economic growth: a
statistical verification’, Economic Development and Cultural Change,
21, 670-9.
308 AGRICULTURE AND ECONOMIC DEVELOPMENT

Hailwood, P. (1979), Stabilization of International Commodity Mar¬

kets, JAI, CT.
Kindersen, O. and Parnes, A. (1975), Trade, Instability and Economic
Development, Lexington, Mass.
Kindleberger, C.P. (1956), The Terms of Trade: A European Case
Study, New York.
Law, A.D. (1975), International Commodity Agreements. Setting,
Performance and Prospect. Lexinton, Mass.
Lipsey, R.E. (1963), Price and Quality Trends in the Foreign Trade of
the United States, National Bureau of Economic Research, Prince¬
ton, New Jersey.
Little, I. Sktovsky, T. and Scott, M. (1970), Industry and Trade in
some Developing Countries, A Comparative Study, Oxford University
Press, OECD, Paris and Oxford.
Macbean, A.C. (1966), ‘Export Instability and Economic Develop¬
ment, George Allen & Unwin, London.
Massell, B.F. (1964), ‘Export concentration and export earnings’,
Amer. Econ. Rev., 54, 47-63.
Massell, B.F. (1970), ‘Export instability and economic structure’,
Amer. Econ. Rev., 60, 618-30.
Michaely, M. (1962), Concentration in International Trade, North-
Holland, Amsterdam.
Osborne, D.K. (1976), ‘Cartel problems’, Amer. Econ. Rev., 66,
835-44.
Ritson, C. (1977), Agricultural Economics, Theory Policy, Crosby
Lockwood, London,

Notes

1 • Barriers to trade such as the import levy and export subsidy do not apply
to tropical products such as tea, coffee, etc. but they do create barriers
when LDC exports compete with temperate zone crops.
 11 Planning Agricultural
Development

1 Introduction

In this chapter we shall be concerned, first, with the concept of

planning and its potential for accelerating development in LDCs.
Next we discuss alternative macro-planning models with particular
reference to experience in the Soviet Union and China. We then
consider the choice of planning strategy for the agricultural sector.
This is followed by a brief consideration of the commonly used
planning techniques of input-output analysis, linear programming
and activity analysis. Finally, we undertake a detailed examination of
agricultural project planning covering the underlying principles of
project analysis and problems of application in LDCs, including
some peculiarities of agricultural projects.

2 Meaning and Objectives of

Economic Planning in LDCs
Planning, in general, means a deliberate attempt on the part of a
country or its government to follow a specific pattern of economic
growth and structural change in the economy. Frequently, planning
implies intervention by government in the normal working of the
market. However, the degree of intervention very much depends on
government ideology. France, for example, follows a system of
indicative planning whereby the degree of intervention by govern¬
ment is minimal. Indeed, government simply sets some targets for the
private sector, outlines the major directions of economic policy, and
provides information to the market economy so that the desired

309
310 AGRICULTURE AND ECONOMIC DEVELOPMENT

targets can be achieved. Here the state does not control all the means
of production. On the other hand, in Soviet Russia, the state controls
most of the means of production, dictates the pattern of demand and
supply, and owns the ‘commanding heights’ of the economy. As a
consequence, the degree of state intervention in the market mechan¬
ism is the greatest. The market mechanism, in fact, hardly operates to
allocate real resources within the Soviet economy.
In between these two extremes, we have a large number of inter¬
mediate degrees of government intervention in the operation of the
market. However, prices are not suppressed completely. Indeed, in
many countries, prices play important roles in the allocation of re¬
sources. Such countries operate within the framework of a ‘mixed’
economy, where both planning and the market coexist side-by-side.
Many LDCs have mixed economies and, given market imperfections,
government interventions are taken as attempts to rectify these. The
main reasons for the introduction of planning in LDCs are derived
from the Soviet experience which can be stated as follows:
(a) Planning is supposed to speed up the rate of economic growth of
a country. If the market is allowed to operate without any
intervention then it may take LDCs a long time to attain a
reasonable standard of living.
(b) The history of the operation of markets in LDCs does not seem to
suggest that allocation of resources will always be very efficient or
optimal in an unregulated economy. In many cases, markets
operated with considerable distortions and economies of most
LDCs stagnated.
(c) In a market economy, private investors usually try to maxi¬
mise the short-run private profits rather than the long-run
social benefits. As a result, marginal private net benefits and
marginal social (net of costs) benefits will not be equal and
resource allocation will be less than socially optional.
(d) Given the lack of entrepreneurial skill, initiative and leadership in
private business and commerce in many LDCs, it is sometimes
argued that government via planning should provide dynamic
leadership and enterprise to stimulate investment in order to
attain a higher and better standard of living.

Against these arguments in favour of planning, it has been pointed

out that government intervention in the market has considerable
drawbacks. Here we shall briefly state three major criticisms. (More
detailed discussions of planning can be found in Heal (1971); Ellman
(1979) and Healy (1971).)
 PLANNING AGRICULTURAL DEVELOPMENT 311

(1) If the marginal private and social net benefits are not equal to one
another, then government should provide information and other
services (e.g. the creation of social capital like roads, schools,
hospitals) so that they can be equalised. This would, however,
imply a greater role for government in the economy than the
supporters of a pure market economy may favour. But it does not
imply the introduction of planning on a grand scale.
(2) Planning - either for the entire economy or for a particular
sector, such as agriculture - needs considerable information to
match supply with demand at a price which is mutually
acceptable to producers and consumers. In many countries, such
data are very difficult to obtain. Even if information is available,
its nature and quality (particularly for the rural sector) are not
always reliable. Building large input-output or econometric
models for planning the agricultural sector on the basis of rather
dubious information may yield very misleading conclusions.
(3) Many LDCs do not have enough trained manpower to design
and implement planning very effectively. As a result, in a number
of LDCs where planning has been introduced, the implement¬
ation of planning has hardly been very efficient. Moreover, the
use of planning in many LDCs has led to the growth of enormous
and sometimes corrupt bureaucracies. The objective of achieving
a satisfactory balance between demand and supply by planning
(when the market mechanism has been suppressed) has conseq¬
uently not been achieved in many countries. In particular, the
task of disseminating information to obtain a proper coordina¬
tion for an optimal use of resources has remained unfulfilled in a
large number of cases.

3 Macro-Planning: comparison of Soviet

and Chinese Models

Despite these and other criticisms, many LDCs have tried to use
planning partly to achieve a faster rate of economic growth to raise
standards of living and partly to achieve industrialisation via
diversification of a largely agrarian economy. A number of LDCs
have tried to emulate the Soviet planning example in this respect
where agriculture had to pay for the growth of industries. This type of
planning of the agricultural sector has been described as the ‘tribute’
model.
312 AGRICULTURE AND ECONOMIC DEVELOPMENT

3.1 The Soviet ‘Tribute’ model

Just after the Bolshevik Revolution of 1917 Soviet politicians and

planners argued that to achieve a rapid rate of industrial growth,
‘surplus’ resources must be transferred from the agricultural to the
industrial sector. The term ‘surplus’ to this context means the excess
of sectoral production over sectoral consumption. (For a full
discussion of the concept of‘marketable surplus’ and its mobilisation
for economic development, see chapter 4.)
The Soviet leaders decided that agriculture should pay for
industrialisation and the rapid growth of the whole economy, by a
‘tribute’ levied on agriculture to finance the cost. The mechanism for
imposing such a tax is described in Figure 11.1. Let us assume that
there are only two sectors in the economy: agriculture and industry.
Let us also assume that the industrial sector produces only two
commodities: machinery and cloth by using labour and capital
(machinery). Wages of all labourers are assumed to be paid in grains.

Figure 11.1: The tribute model

 PLANNING AGRICULTURAL DEVELOPMENT 313

These grains are produced by labourers in the agricultural sector who

receive cloth in exchange for grains.
The supply of grains is a function of the ‘offer’ curve, i.e. farmers
supply grain for cloth. Such supply of grains also determines the
employment of labour and the wage rate in the industrial sector.
Consider quadrant I in Figure 11.1 which describes the industrial
production possibility curve CM. Assume that line is discrete in the
model. The planner at period 1 can choose a certain combination of
cloth and machinery. He may choose to produce only cloth (= OC)
but no capital i.e. machinery; or he may decide to produce maximum
possible machinery only (=OM) for the industrial sector, but no
cloth at all in period 2. Should the planner decide to produce OM' of
machinery, then the production of cloth will be equal to M'C' which
can be exchanged for grains produced by farmers in the agricultural
sector. The amount of grains which will be available in exchange for
cloth is given by OG in quadrant II as indicated by the offer curve of
farmers OP. Notice that the slope of OP shows the terms of trade (i.e.
price ratio) between grains and cloth. In quadrant III the slope of the
line OW determines wages (measured in grains) for the employment
of labour in the industrial sector. Thus the supply of OG amount of
grains determines the employment of OL amount of labour in
quadrant III.
Once we have all the necessary information about total machin¬
ery, cloth, grains and labour i.e. C', P, G, W and L), we can
determine, in quadrant IV, the production possibility curve like MLt.
It is clear from the above analysis that factors which will constrain
the growth rate are the initial endowments of capital, the terms of
trade and the wage rate. If it is decided by the Central Planning Board
that the growth rate should be raised then, in a command economy,
farmers can be forced to give up more grains for the same amount of
cloth, i.e. a move from a point on the farmers offer curve such as P2 to
a point like P'. Such a ‘tribute’ or tax ( = GG') imposed on the
agricultural sector could then be utilised to employ more units of
labour in the industrial sector. As such, the production possibility
curve shifts to the right to ML2 and a higher rate of economic growth
for the whole economy can be attained. In other words, the policy of
collectivisation is very desirable since it enables the planners to obtain
the necessary ‘surplus’ from the agricultural sector to pay for the
industrialisation of the economy.
A number of criticisms can be levelled at the ‘tribute’ model. First,
while it may be applied to a totalitarian economy where farmers are
forced to give up a part of their product, most LDCs - with the
exception of China and Vietnam - do not have a planned economy.
Indeed, most LDCs have ‘mixed’ economies where both the private
314 AGRICULTURE AND ECONOMIC DEVELOPMENT

and the public sector co-exist and market prices are not entirely
suppressed. Such prices in LDCs still play important roles in
allocating resources. Hence, it is not clear to what extent the transfer
of the ‘tribute’ model in an underdeveloped economy would be
successful.
Second, there is the problem of providing farmers with an adequate
incentive to produce grains when the terms of trade are shifted
against the agricultural sector. It could be pointed out that when the
terms of trade move against agriculture (i.e. when the farmers have to
give up more grain for obtaining the same amount of cloth), then
their incentive to produce more will obviously diminish. A fall in the
production of food grain will also lead to a fall in the amount of
‘tribute’ or surplus that can be raised in the agricultural sector as it is
well known that marketed surplus and production are directly
related. In a free market, a reduction in agricultural price (= Pa) vis a
vis the industrial price (= Pi) (i.e. an adverse movement of (Pa/Pi)
could easily lead to a fall in the production due to the operation of the
substitution and income effects. Figure 11.2 illustrates this.
The budget lines for the farmers are given by AB and CD, in terms
of agricultural output, respectively, prior to and after the transfer of
surplus labour from the agricultural sector. Hence, AC is the additional
per capita of agricultural population after the surplus labour transfer
since in the figure we measure agricultural output per capita. Where
the terms of trade line, AB, is tangent to the indifference curve, the
initial equilibrium is obtained at point E. A relative change in terms of
trade against agricultural sector is reflected by the change in the slope
of CD which is steeper than AB. (e.g. difference between slopes of AB
and CD represents the change in the terms of trade against
agriculture). The new terms of trade line CD passes through E since it
necessary to absorb the whole of the potential savings. Since the
indifference curves cannot intersect, it follows that the new equili¬
brium on line CD will occur at a point like E', and this indicates a fall
in marketed surplus. A more drastic change in the terms of trade
(advocated by Stalin in Soviet Russia in the 1920s and 1930s) may,
however, raise the marketed surplus (Readers can verify this easily by
manipulating the figure.) However, such a policy may involve
significant adverse economic, social and political consequences. (For
a further elaboration of this point see Ghatak, 1978, ch. 6.)
It is also important to remember that the Soviet coercive model for
agricultural planning may largely explain the high costs and low
productivity in Soviet agriculture. A forcible rise in the flow of wage-
goods from the agrarian sector must be viewed against the fall in
 PLANNING AGRICULTURAL DEVELOPMENT 315

Agricultural

Figure 11.2: Terms of trade effect on agricultural surplus

livestock production and a decline in the flow of industrial

goods - including investment goods - to the agricultural sector. In
the long run, a fall in the income and purchasing power of the
peasantry must lead to a fall in productivity in the agricultural sector
due to a fall in investment in the agrarian economy, which in its turn is
very likely to affect the production of agricultural output (the main
determinant of agricultural ‘surplus’) adversely. Indeed, after Stalin’s
death the coercive nature of Soviet agricultural planning was
gradually reduced, chiefly due to its adverse impact on agricultural
production. Yet, despite such relaxation, Russia still needs to import
large quantities of grains - chiefly from the US - even at the beginn¬
ing of 1980s. In the light of Soviet experience, it is not surprising that
following the Chinese Revolution in 1948, Mao decided to follow a
different agricultural planning model.

3.2 The Chinese model

There are important differences between the Chinese and the Soviet
models for agricultural planning. First, in Chinese agricultural
316 AGRICULTURE AND ECONOMIC DEVELOPMENT

planning, massive efforts were made to utilise all available resources

for the growth of the agricultural sector. Second, the Chinese
Communist Party played a very important role in the agrarian
changes introduced after the revolution. Third, planning of Chinese
agriculture has been aimed principally at raising the level of
agricultural production, whereas in the case of the Soviet Russia,
agricultural planning mainly implied raising as much wage-goods as
possible from the agricultural sector. In Mao’s opinion the massive
tax burden imposed on Soviet agriculture was responsible for the fall
of agricultural production.
Mao also advocated the need for a steady growth in the income of
the peasantry for the implementation of a successful collectivisation
programme for agriculture. The relative success of China in imple¬
menting its collectivisation programme has been well summarised by
Ellman (1979). First, in contrast with Soviet experience, grain output
in China did not fall. Secondly, the fall in livestock in China after the
revolution was quite modest, while in the case of Soviet Russia such a
fall was very significant (between 1929-30 in Soviet Russia, the
number of pigs fell by about 47 per cent; in China it fell by about
18 per cent between 1956). Thirdly, the scale of massacre of the
peasantry in Soviet Russia during the 1920s and 1930s was signifi¬
cantly higher than in China during the 1950s. Finally, China did not
deport some of its best farmers to achieve its goal of changing the
rural society through collectivisation.
However, it is useful to note that much of the gain from the
collectivisation programme in China was thrown away during the
period of the Great Leap Forward (1958-62) as both the crop and
livestocks output fell significantly due to the bad management of the
economy. Nevertheless, during the 1950s, substantial progress was
made with irrigation. Further, the system of income distribution was
made much more egalitarian in the rural sector. Capital formation
within the agricultural sector registered a steady rise as resources
were diverted from property income to the accumulation of capital.
Since 1961 (i.e. after the Great Leap Forward), agricultural output
began to rise with the increasing application of chemical fertilisers
and improved seed varieties. A general rise in the proportion of
irrigated land also helped enormously to expand output. Even so,
with population rising steadily, Howe (1978) estimates that the mean
grain consumption in China in per capita terms in the 1970s was only
25 per cent higher than that of Bangladesh. According to another
estimate, Chinese agriculture still suffers from some basic problems
such as a low growth rate, low yield and productivity, and low level of
 PLANNING AGRICULTURAL DEVELOPMENT 317

total production (Padoul, 1975). Indeed, one estimate suggests that

the growth rate of grain output in China has only been two-thirds
that in India during 1952—73 (Bandyopadhyaya, 1976).
It is, however, important to remember that Chinese agricultural
growth estimates are highly sensitive to the base period, as well as the
output figures chosen for the base period (Ellman, 1979). Ellman
claims that the mean annual grain output growth rate is ‘respectable’
3.6 per cent per annum when 1949 is taken as a base; but it goes down
to 2.4 per cent per annum if 1952 is chosen as base. As regards the
official grain production figures, it is useful to point out that many of
these estimates suffer from a number of inaccuracies. Hence, a good
deal of caution is necessary in drawing final conclusions. As Sinha
(1975) points out:

There is substantial qualitative evidence that the level of food consumption

has improved [in China], which in itself indicate that food production has
risen faster than population. However, with the present state of quantitative
information one cannot say whether the levels of food consumption in China
today are better or worse than in the mid-1930. But one can easily say that
because of the egalitarian policies the level of food consumption of the poorer
people is much better now than it was in the 1930s. For the same reason, it can
be said that poorer people in China are eating better than India or that
the Chinese can withstand food scarcity better than Indians, (emphasis
added)

The other point that needs to be emphasised is that the argument for
collectivisation, based on its alleged efficacy of raising net transfers
from agriculture to industry, is wrong (Ellman, 1979). It is difficult to
reach a definitive judgement about the direction of financial and real
resource transfer from agriculture to industry for all the periods in the
Chinese economy since 1949. Some preliminary investigations sug¬
gest that in some periods there has been a steady shift of resources in
favour of the agricultural sector (Paine, 1976).

4 Choice of Planning Strategy for Agriculture

The Soviet and the Chinese experiences are useful in understanding

the nature of problems that might arise in the planning of agriculture
in socialist economies. Most LDCs, however, belong to ‘mixed’
economies. It is, therefore, necessary to adopt those strategies for
planning agriculture in these countries which will be conducive to the
long-run development of the agrarian sector.
318 AGRICULTURE AND ECONOMIC DEVELOPMENT

The formulation of such strategies is of crucial importance for the

growth of output and employment within the LDCs. Some argue that
development plans should aim at the application of the latest in
science and technology and condemn ‘the placing of restrictive
import on farm implements of proven productive performance such
as tractors and power tillers as they are thought to be “labour
saving” and might cause serious unemployment in the rural
economy’ (Hopper, 1968). On the other hand, although it is
important to emphasise the role of modern science and technology in
the agriculture of LDCs, for some it may be ‘patently impossible to
apply the latest science and technology except in a highly selective
manner’ (Johnston, 1972). In this context, it is useful to discuss the
uni-modal and bi-modal strategies for agricultural developments,
bearing in mind that the following aims are particularly relevant:
(a) The contribution to the overall rate of economic growth and
the process of structural transformation.
(b) The realisation of a reasonable rate of rise in agricultural output
at a minimum cost by helping sequences of innovations which
‘exploit the possibilities’ for technical progress most suitable to
a country’s factor endowments.
(c) The realisation of a general increase in rural standard of living
and welfare.
(d) ‘Facilitating the process of social modernisation’, (i.e. controlling
fertility, advancement of rural education and improving the
entrepreneurial abilities, etc.

One of the most important problems that LDCs face is the choice
between a bi-modal or uni-modal strategy. In the first resources are
concentrated within a sub-section of a large and capital-intensive
unit, while under a uni-modal strategy, policy formulators try to
induce a gradual and wide rate of adoption of technical progress, the
nature of which is generally geared to the nature of factor proportions
of the whole sector. Johnston (1972) has put it very clearly:

The essential distinction between the two approaches is that the uni-modal
strategy emphasises sequences of innovations that are highly divisible and
largely scale-neutral. These are innovations that can be used efficiently by
small-scale farmers and adopted progressively. A uni-modal approach does
not mean that all farmers or all agricultural regions would adopt innovations
and expand output at uniform rates....
Although a bi-modal strategy entails a much more rapid adoption of a
wider range of modern technologies, this is necessarily confined to a small
fraction of farm units because of the structure of economies in which
 PLANNING AGRICULTURAL DEVELOPMENT 319

commercial demand is small in relation to a farm labour force that still

represents some 60 to 80 per cent of the working population, (pp. 36-7)

Owen (1971) argues that ‘it may be posited as a basic condition of

economic growth in all countries that most of the available land
resources should be incorporated in the commercial sub-sector.’
However, it should be recognised that, from the point of view of
income distribution and equity, a uni-modal approach may be
preferable to a bi-modal one. Following the analysis of Farell, it can
be shown that ‘total efficiency’ (as the sum of price, technical and
economic efficiency) of a farm under uni-modal strategy will be
greater.
Consider a firm using capital (= K) and labour (= L) to produce a
given amount of output shown by the isoquants in Figures 11.3 to
11.5. A technological frontier is a point of output which can be
produced by the minimum amount of capital and labour. A firm is
320 AGRICULTURE AND ECONOMIC DEVELOPMENT

technically efficient if it has an input combination that lies on the

frontier isoquant (=Ft). Firms which use ‘average’ technology
(i.e. not the most efficient) have input combinations that lie on an
‘average frontier’ (= At).
By using the simple principles of micro-economics we show by
these figures that firms are price-efficient where price lines (PP) are
tangent to the isoquants. Notice that a bi-modal strategy should
involve ‘capital-augmenting’ technical progress within a modernised
sub-sector. A uni-modal strategy should involve a gradual technical
progress and a smooth rise in capital intensity which involves the
whole agricultural sector. These discrepancies are shown in the
figures.
In Figure 11.3, Ft+t. shows a new frontier isoquant after a capital-
augmenting bi-modal strategy has been introduced; whereas in
Figure 11.4, Ft+t. shows the frontier given by the wide use of
improved inputs like seeds and fertiliser.
 PLANNING AGRICULTURAL DEVELOPMENT 321

Given the same relative factor prices (PP), under a bi-modal stra¬
tegy a higher K/L will emerge for those (usually large) firms that have
access to land and capital. As such, a bi-modal strategy implies a con¬
siderable difference in the factor proportions used by ‘best’ and ‘aver¬
age’ firms as shown by slopes of the K/L and K'/L' rays. Now if
resources are concentrated within a small sub-sector to promote bi-
modal strategy, then there will be a general decline in the ability of the
firms outside the sub-sector to use new inputs. It will also imply less
foreign exchange for the firms outside the sub-sector to the extent that
capital has been imported. Further, given the underpricing of capital
(as rates of interest are low and currency is overvalued in the foreign
exchange market) and overvaluation of labour in LDCs (since wages
are higher than marginal productivity of labour), the divergences in
K/L caused by biased technical progress will be accentuated. As such,
if a country adopts a bi-modal strategy the difference between the

Figure 11.5: Biased technical progress under a bi-modal strategy in the presence of
distortions
 322 AGRICULTURE AND ECONOMIC DEVELOPMENT

factor intensities and the technical efficiency of the ‘best’ and

‘average’ firms is likely to widen as the agricultural sector undergoes a
process of structural change. Indeed, given the types of distortions in
the factor and the foreign exchange markets we have described, these
differences are likely to be accentuated even further. The aggravated
divergences in the presence of distortions (underpricing of capital and
overpricing of labour) is shown in Figure 11.5 by the rays, Km/Lm and
Kt/Lt for the modern and traditional sectors, respectively.
A bi-modal approach will also be inadequate to absorb ‘surplus’
labour, and hence the strategy will fail to promote a more egalitarian
system of income distribution. On the other hand, a uni-modal
strategy emphasises divisible and scale-neutral innovations which
generally imply that the factor intensities will be fairly similar for
' both the most efficient and the average firms. Also, under a
uni-modal strategy, most firms within the agricultural sector will
adopt techniques of production which will be compatible with the
economic structure and factor endowments. Such a choice of factor
intensity will minimise the foreign exchange needs of a developing
country. As more firms adopt new technology, a process of‘learning
by doing’ will generate its own momentum and the diffusion of
innovation will be widespread. Needless to say, the uni-modal
strategy offers the possibility of a greater use of labour in
relation to capital-via irrigation and multiple cropping, for
example. If the objective is to increase employment as far as possible
and to achieve a more egalitarian system of income distribution, then
it seems that a uni-modal system should be preferred to a bi-modal
one, even though the latter may generate a high rate of economic
growth.

5 Planning Techniques

5.1 Input-output analysis

The use of input-output (IO) analysis has now become quite

common as a sectoral planning tool in many countries). The
technique was originally devised by Leontief, and is based on the
input-output tables showing the linkage between various sectors
(e.g. agriculture and industry) of the economy and providing the
foundation of the economy-wide general equilibrium model that
focuses on production. In IO analysis the main assumptions are:
 PLANNING AGRICULTURAL DEVELOPMENT 323

(a) inputs are needed in constant proportion to produce output in

each sector;
(b) each activity or sector can produce only one product; and
(c) joint products cannot be produced by any sector.

Note that (a) clearly implies that the returns to scale are constant.
It is necessary to separate prices and quantities if we wish to
develop a model from the IO tables. Assuming a closed economy, i.e.
no exports and imports, the rows of IO tables can be written as
follows:

P,X, = X P.X.j + P.F, -. - (1)

where Xt = production in sector i, P; = price of output in sector i,
= flow of intermediate goods from sector i to sector j, Fj = final
demand for sector i. Let a^ = intermediate needs from sector i per unit
of output of sector j.
As the IO coefficient is fixed, we have:

aij = X,A
Hence, in any period, we have:

Fa.
= PiXu/PjXj

With an IO table for a given year, it is useful to state the units of real
flows in such a way that prices will add up to one.
If we now divide equation (1) by the price, we have:
Xi = ZaijXj + Fi... (2)
In matrix form, this material balance equation is:
X = AX + F... (3)
and hence we can solve for X as follows:
X = (I — A)_1F... (4)

Note that the solution indicates that given the final demand,
production for different sectors can be determined so that supply is
equal to demand in different sectors.
This model can be extended to include foreign trade. Assume that
E = exports, and M = imports. We then have:
X + M = AX + F + E... (5)
324 AGRICULTURE AND ECONOMIC DEVELOPMENT

which gives
X = (1 — A)_1(F + E — M)... (6)

Notice that the IO technique is very useful if planners wish to

promote a balanced growth between agriculture and industry. Notice
also, that many LDCs are net importers of food and as such the use of
equation (6) raises problems as the net final demand turns out to be
negative.
A solution has been suggested by Chenery and Clark who recast
the problem in terms of total demand and supply. (See Chenery and
Clark (1959); and also Dervis, de Melo and Robinson (1982), who
give a very good account of the use of IO technique and a computable
general equilibrium model.) It should be pointed out that IO analysis
is not very helpful in finding the ‘optimal’ technique of production
from a variety of technologies that may be available for producing
a given commodity. In order to identify the ‘optimal’ (most efficient)
technology, it is necessary to use linear programming models.
It is important also to note that the IO analysis is a macro-planning
tool, and its relevance to sectoral planning for agriculture is largely
confined to testing for the feasibility and consistency of sectoral
targets.

5.2 Linear programming

Some planning models have been developed in both the socialist and
the non-socialist countries for improved functioning of agriculture at
farm, local or national levels. These models are generally used to
illustrate the objective function; identify the constraints; evaluate
problems of efficiency in the allocation of resources; and derive policy
implications in the light of such evaluations. It is useful to remember
that some form of planning the agricultural sector is unavoidable in
both socialist and market economies. In the case of market
economies, examples of agricultural planning can be easily cited by
noting the extent to which public policies play a role in determinig
production quotas, support prices, subsidies for output and the
money to be spent on research and development for agriculture,
irrigation, fertiliser use, etc. Indeed, some problems of farm planning
are common to both market and non-market economies as all farms
have to face several constraints - land, capital and labour, for
example-to maximise a given objective function (e.g. profit or
income). In addition, many face institutional constraints - transport,
credit and marketing facilities, say. Farms also face problems
 PLANNING AGRICULTURAL DEVELOPMENT 325

associated with risk and uncertainty - the impact of weather and the
use of new types of seed, pesticide or fertiliser, the effects of new
technology, and so on.

5.3 The ‘cattle-feed’ problem of cost minimisation

Corresponding to the solution of the maximisation problem in the

linear programming analysis, there exists a method to solve a
minimisation problem. For students of agricultural economics, such
a solution can best be illustrated by analysing the minimum-cost of
cattlefeed (the ‘diet’) problem.
Let us assume that the daily feed of an animal should be given by
minimum-cost subject to certain nutritional constraints. To simplify
the problem, we assume that there are only four kinds of raw
materials - oats, barley, sesame flakes and groundnut meal. Let us
further assume that there are two nutritional constraints: (i) the feed
should have at least 20 units of protein; and (ii) the feed should have
at least 5 units of fat.
In Table 11.1, the protein and fat content per unit of each raw
materials and units costs are shown (this example is adapted from van
de Panne). Let us write the system of equations as follows:

12xt -|- 12x2 + 40x3 + 60x4 ^ 20

2xx + 6x2 -I- 12x3 + 2x4 ^ 5 (7)

Let us introduce the ‘slack’ variables (i.e. yt and y2) which are non¬
negative, indicating an excess of protein and fat needs. We can then
transform the inequalities as follows:

\2\l + 12x2 + 40x3 + 60x4 — yi = 20... (8)

2xj + 6x2 -I- 12x3 + 2x4 — y2 = 5... (9)
If the aim is to minimise the costs of the feed, then we can write the

Table 11.1: Protein and fat content and costs of raw materials

Protein Fat Cost per

content content unit

x, Barley 12 2 24
x2 Oats 12 6 30
x3 Sesame flakes 40 12 40
x4 Ground nut meal 60 2 50
326 AGRICULTURE AND ECONOMIC DEVELOPMENT

equation for such an objective function as follows:

F = 24xj + 30x2 + 40x3 + 50x4... (10)

Rearranging equation (8)—(10), we have:

0 = — 24xj — 30x2 — 40x3 — 50x4 + F

— 20 = — 12x3 — 12x2 — 40x3 — 60x4 + y3
— 5= — 2xt — 6x2 — 12x3 — 2x4 + y2... (11)

In Table 11.2 the simplex tableaux for the minimum cost cattlefeed
problem is described. It is obvious that the solution in tableaux is not
feasible as y! and y2 (the basic variables) are negative. We know that
the simplex method needs a feasible solution to begin with; hence we
must find a basic feasible solution. Note that half the sesame flakes
should meet the protein need and have 6 units of fat, which is more
than what we need. We now find the corresponding basic solution in
which the amount of sesame flakes, x3, is not-zero. Thus a non-basic
variable in which the slack variable for protein yl5 is reduced to 0 and
this becomes a basic variable. We find this solution is tableaux 0 by
pivoting on the element in the row of yj and the column of x3. Since
cost-minimisation is our aim, the variable corresponding with the
most positive element in the first row is chosen as the new basic

Table 11.2: Simplex tableaux for minimum-cost cattlefeed problem

Tableau Basic Values X, *2 x3 x4

variables basic
variables

f 0 -24 -30 -40 -50

0 yi -20 -12 -12 -40 -60
y2 - 5 - 2 - 6 -12 - 2
*1 x2 yi x4
f 20 -12 -18 - 1 10
1 3 1
1 *3 2 10 A “ 40 1*
3
y2 1 H -2| 10
16
11

*i x2 yt y2
13 5
f 19| -13 -16i ~ 16 8
13 3 21 1 3
2 *3 32 20 40 320 32
1 1 3 3 1
x4 16 10 ~ 20 160 16

Source: van de Panne (1976).

 PLANNING AGRICULTURAL DEVELOPMENT 327

variable. This is shown by the following equation:

20 + 12x2 + 18x2 + yt — 10x4 = F

It is shown in the first row of tableaux 1. Note that x4 now enters the
basis as y2 departs. We now obtain tableaux 2. It is evident that the
first row does not contain any positive number, and surely we have
obtained the solution for cost minimization! As tableaux 2 shows, the
minimum costs is 19| and the values for x3 and x4 are given by
13/32 and 1/16, respectively.
Sometimes, imputed values can be assigned to the raw materials
used and generally these imputed values are derived from the costs of
the raw materials of the cattlefeed. If markets for such raw materials
are very imperfect - as they could be in many LDCs - then shadow
prices are usually worked out to obtain optimal solutions.
A typical cost-minimisation problem can then be summarised as
follows:
Minimise F = CjXj + C2x2 + • • • + Cnxn... (12)

subject to:

A„Xl + A12x2 + •••+Alnxn^b1

Am,x1+Am2x2 + ---+ AmnxN^bm... (13)

All the x variables are non-negative. (For further discussion of the use
of artificial variables, the saddle point and the dual method, see
Dorfman, Samuelson, Solow, 1959; and Van de Panne (1976).)

5.4 Activity analysis and the choice of production technique

Linear programming is a form of activity analysis in which the

optimum combination of activities is selected subject to the objective
function and one or more linear resource constraints. However, the
technique of activity analysis without resource constraints can also be
used to select the optimum activity from a limited choice of activity
vectors, where the objective function is to minimise total resource
costs. An interesting application of this method of analysis relates to
rice milling in Indonesia (Timmer, 1974).
Five rice milling techniques were simultaneously available, distin¬
guished by different capital/labour ratios. The most labour-intensive
technique, hand pounding, had a zero capital requirement. At the
other extreme, the large bulk rice mill technique was characterised by
a high K/L ratio and a relatively high absolute capital requirement.
328 AGRICULTURE AND ECONOMIC DEVELOPMENT

Figure 11.6: Choice of technique

The K/L ratios of the remaining three milling techniques - the small
rice mill, the large rice mill, and the small bulk rice mill-lay between
the extremes. In Figure 11.6 where required capital and labour inputs
are measured along the two axes, the fixed input coefficients of the
various techniques are signified by the slopes of the vectors HP,
SRM, etc. The objective function is to minimise the costs of milling
relative to a given ‘output’ (strictly, value added) of milled rice.
In the figure this output constraint (arbitrarily determined) is
represented by the linear-segmented isoquant QQ'. A ‘corner’ occurs
on QQ' wherever the isoquant is intersected by an activity vector.
The cost-minimising activity or technique is located at the corner
where the relevant ‘budget’ or ‘isocost’ line is tangent to QQ'. In the
figure the slope of the line BB' reflects the relative costs of labour
and capital, and the point of tangency with QQ' identifies the small
rice mill (SRM) as the cost minimising technique.
Empirical evidence published by Timmer shows that in Indonesia,
where wages, interest rates and rice prices have been determined by
market forces, without undue interference from the government,
small rice mills have indeed become the dominant rice milling
technique, largely at the expense of hand pounding. However, hand
 PLANNING AGRICULTURAL DEVELOPMENT 329

pounding is still practised on farms with surplus labour, and

especially for self-consumption. Despite their ready availability on
the market, large capital-intensive rice mills have been largely ignored
by profit conscious millers, even though large mills offer some
technical advantages such as a higher extraction rate and a lower
proportion of broken grains. But at prevailing prices of both rice and
resources used in rice milling, these technical gains are not worth
pursuing. Moreover, sensitivity analysis, involving variation of the
labour/capital price ratio represented by the slope of the isocost line
BB', indicated substantial stability of the corner solution correspond¬
ing with the choice of the small rice mill. However, as may be readily
perceived from Figure 11.6, large rice mills may well become the
optimal cost-minimising technique at some time in the future as
labour becomes more expensive and the relative cost of capital
declines. In the meantime, adherence to small and relatively labour-
intensive rice mills as the dominant technique must confer very
substantial employment benefits on society. Had millers chosen to
adopt one of the more capital-intensive methods, very large numbers
of jobs in the rice-milling industry would have been lost, with poor
prospects of early alternative employment for a high proportion of
the workers concerned.
This example also serves to illustrate how market price distortions,
or misguided government intervention in the market, can lead to
socially inappropriate technological choices. In the Indonesian case
any number of market imperfections or government interventions
could have combined to alter the K/L price ratio sufficiently to shift
the optimum choice to one of the more capital-intensive techniques of
rice-milling. The social costs of such a shift, in terms of increased
unemployment, could have been very heavy indeed.

6 Agricultural Project Planning

This section falls into three sub-sections. First, the basic principles
of project appraisal are outlined and discussed; second, special
problems of application in LDCs are considered; and finally,
questions of practical application and choice of appraisal methods
are reviewed.
330 AGRICULTURE AND ECONOMIC DEVELOPMENT

6.1 Principles

6.1.1 Objectives of Project Appraisal

Project appraisal is a form of applied welfare economics or cost-
benefit analysis. Its central objective is to estimate the social net
benefits of an investment project on the economy as a whole in
contradistinction to merely estimating its financial prospects as
viewed by private investors.
The earliest examples of cost-benefit analysis were concerned with
the appraisal of major flood control, irrigation and hydro-electric
schemes in the USA, where it was realised that there were many
‘hidden’ costs and benefits associated with such schemes. The US
government accordingly sought a standardised appraisal procedure
for use in separating viable from non-viable projects, and for placing
the viable ones in priority order where these were in competition. The
Food Control Act 1936 laid down that for projects to be authorised
the benefits must exceed the costs ‘to whomsoever they may accrue’
(Eckstein, 1965, p. 2). Cost-benefit analysis represents an attempt to
apply this criterion.

6.1.2 Choice of Investment Criteria

Two of the most comprehensive decision criteria used in project
appraisal are the net present value (NPV), and the internal rate of
return (IRR). A third useful criterion, the benefit-cost ratio (B/C), is
similar to NPV.
Net present value. This is the difference between the PV of total
benefits and the PV of total costs. The following definitions and
notation are borrowed from’Eckstein (1965, ch. hi). Fomally we
have:

B
PVTB = I (14)
t-,a+r)‘

where B = total annual benefits. Similarly:

t y
PVTC = K0 + I (15)
t=.(l + rV'

where K0 = fixed (initial) investment, and V = total annual costs of

operation and maintenance.
For simplicity, we assume that both B and V remain constant for
 PLANNING AGRICULTURAL DEVELOPMENT 331

all t, and that there is no further fixed investment after time t = 0. It

follows that:

NPV = PVTB-PVTC

B V
K+I (16)
(l+r)‘ t=i
d+r)'

Equation (16) expresses the difference between the PVs of two time-
streams, the one representing benefits, the other costs.
In use, the condition NPV > 0 must be satisfied to justify the
project. Competitive projects are ranked by the magnitude of NPV,
subject to the previous condition unless there is an overall capital
constraint. The discount rate is exogenous, i.e. it is part of the input
data.
Internal rate of return. Continuing with the same notation, this is the
discount rate satisfying the condition NPV = K0, where K0 5= 0. Thus,
we have:
B-V
K0=I (1 + i)‘ "* (17)
t=i

where i = internal rate of return:

In use, the expected IRR must reach some positive minimum value
to justify the project. More formally, justification requires satisfaction
of the condition IRR il, where i‘ is the opportunity cost of capital,
the social time preference rate, or other ‘qualifying’ discount rate.
Competitive projects are ranked by the magnitude of IRR, subject to
the previous condition. The discount rate i is endogenous, i.e. it
emerges from the analysis. However, the test discount rate, il, still has
to be determined independently.
Benefit: cost ratio. The components of the calculation are identical
with those of the NPV criterion. But instead of taking PVTB - PVTC
we take PVTB/PVTC. More precisely:
T V i
B
B/C = Y. K+ I
(T+?
t=l (1 + r)' t=l

In use, a project is justified provided B/C ^ 1. Competitive projects are

ranked by the magnitude of B/C subject to the first condition. Unlike
the NPV criterion, this one does not discriminate between large and
small projects.
332 AGRICULTURE AND ECONOMIC DEVELOPMENT

6.1.3 Critical Problems

Two of the most critical problems of project appraisal are the choice
of discount rate, and allowing for risk and uncertainty.
Choice of discount rate. As already noted, a test discount rate is
implicit in all NPV and B/C ratio estimates. It also defines the
minimum acceptable rate or-return for comparison with IRR
estimates. Pragmatically, if the discount rate is set too high a bias is
created in favour or projects with relatively low initial capital
expenditure and with benefits accruing earlier rather than later. So in
agriculture, for example, a rapidly maturing extension project might
be favoured at the expense of an irrigation project with a larger
initial capital requirement and a longer lead time. If the discount
rate is set too low the converse argument applies. Additionally, the
choice of a low rate will tend to result in too many projects being
justified.
Theoretically, the optimum discount rate is the one which equates
the marginal social opportunity cost of capital (SOC) with the
marginal social time preference rate (STPR). Whereas SOC defines
the rate of trade-off between present and future consumption,
according to the level of current saving and investment, STPR shows
the rate at which society collectively trades off present against future
consumption (Pearce and Nash, 1981, ch. 9). Since it is economically
rational to rank investment projects in descending order of their
expected rate of return, an inverse relationship between the levels of
saving/investment and SOC is assumed. The reward for ‘waiting’
goes down with the declining social profitability of ranked projects.
The relationship between the levels of S/I and STPR is assumed to be
direct, reflecting the increased marginal opportunity cost of waiting
at higher levels of S/I. The two functions are shown in Figure 11.7
where, given a socially sub-optimum level of investment Is, the SOC is
q whereas the STPR is lower at i. The condition SOC = STPR defines
the optimum level of investment ID and the discount rate r. Thus,
where the coordinates of SOC and STPR meet at I0 and r, the ‘gap’
between q and i is effectively eliminated.
In the real world a government may be unable to induce as much
investment as I0 due to a ‘savings gap’. The government is more ‘far-
sighted’ than private investors and taxpayers, particularly regarding
investment for the benefit of future generations. Although private
citizens as taxpayers may be prepared to sacrifice more, i.e. accept a
higher cost and lower reward for ‘waiting’, than as voluntary savers,
the total level of saving and investment is nevertheless subject to
political constraints. Thus, if the ‘first-best’ level of saving and
 PLANNING AGRICULTURAL DEVELOPMENT 333

Figure 11.7: Choice of second-best discount rate

investment and the corresponding social equilibrium discount rate

are I0 and r, it is easy to see from Figure 11.7 that a compromise,
‘synthetic’ discount rate may be a ‘second-best’ solution. So, if Is is the
highest feasible level of total saving and investment, q and i,
respectively, represent the upper and lower limits of the synthetic
rate, d (i.e. q > d > i). Regarding d as some weighted average of q
and i, it can be argued that the relative weights depend upon national
economic priorities and political judgement. Being fairly closely
related to the opportunity cost of capital in the private sector, q is
likely to be easier to estimate by observation than i (which is much
more subjective). Thus, in practice, choosing d may boil down to little
more than answering the question: ‘How much less is d than q?’ (We
return to the choice of discount rate for further discussion in section
6.2.2.)
Risk and uncertainty. Substantial disagreement exists in the litera¬
ture on project appraisal about the need to allow for the inherent
riskiness of individual projects. On the one hand, there is the
argument that since the success or failure of a single project does not
334 AGRICULTURE AND ECONOMIC DEVELOPMENT

significantly affect the variance of national income there is no need to

make any adjustment for risk on most projects (e.g. Little and
Mirrlees, 1968, ch. xv). On the other hand, it is argued that
investment criteria must embody some allowance for risk to ensure
that, ceteris paribus, relatively secure projects are favoured over more
risky ones (e.g. Eckstein, 1965 ch. iv). These opposing points of view
indicate that the size of a project, relative to the size of the total
economy in which it is located, may be relevant. Whereas the first
argument seems to hold with particular force for ‘small’ projects in
‘large’ economies the second one may carry weight in relation to
large projects in small economies. As a single project is more likely to
affect regional than national income, the case for risk adjustment
seems to apply, in particular, to projects with an emphasis on regional
development.
Although various ad hoc methods of risk adjustment are available
to project appraisers, virtually all methods are open to serious
criticism (Eckstein, 1965, ch. iv). For example, attempting to limit
risk by imposing an arbitrary upper limit on the life of projects biases
selection in favour of short-lived projects. The effect of adding a
special ‘risk premium’ to the discount rate is similar, in that the
consequent reduction in the NPV of all projects affects longer-term
projects disproportionately. The deduction of an arbitrary flat rate
‘safety allowance’ from the project’s NPV is clearly inappropriate as
a means of adjusting for risks which do grow with time. Because of
the numerous defects of these methods of attempting to adjust for
risk and uncertainty it is better to use the technique of sensitivity
analysis in order to evaluate the problem.
Sensitivity analysis involves appraising and re-appraising the
project on the basis of differing sets of assumptions regarding the
values of critical variables and parameters - such as future prices,
crop yields and supply response co-efficients. By examining the effects
of relatively ‘optimistic’ and ‘pessimistic’ assumptions on the range of
NPV estimates about the mean or ‘expected value’, it may be possible
to judge whether acceptance of the project in its present form entails
an unacceptable degree of risk. If the risk is judged to be too high, it
may be possible to reduce it by redesigning the project, by the
substitution of crops in an agricultural project, for example. But
notice that sensitivity analysis leaves decision-makers to judge how
much risk is acceptable.
A simple hypothetical example will serve to clarify the rationale of
sensitivity analysis. Suppose that the expected NPV of an agricultural
project, based on the best available forecasts of crop yields and prices,
 PLANNING AGRICULTURAL DEVELOPMENT 335

is £ 100,000, but that the forecasting error is such that the risk of the
actual NPV falling below £50,000 is 20 per cent. By altering the crop
mix (to substitute less risky for more risky crops) the project can be
redesigned to yield an expected NPV of £80,000, with the risk of
falling below £50,000 reduced to 10 per cent. Those responsible for
project selection must judge whether they prefer (a) the original
higher-risk version of the project, which also offers a higher expected
NPV, or (b) the redesigned lower-risk but less socially profitable
alternative.
Because of space limitations, this review of the basic principles of
cost-benefit analysis is incomplete. Other aspects, such as the
problems of defining and measuring social benefits and costs,
including external economic effects (externalities), are compre¬
hensively covered by sources such as Pearce and Nash (1981),
Mishan (1972) and Eckstein (1965), to whom interested readers are
referred.

6.2 Application in less developed countries

6.2.1 Reasons for Treating LDCs as a ‘Special Case

The restrictive conditions under which private profit reflects social
profit - such as the existence of perfect competition and full employ¬
ment and the absence of externalities - are never fully satisfied in the
real world, even in developed economies. Project appraisal has
consequently been widely used in economically advanced countries,
particularly in areas such as water resource development, transport
and defence planning.
In turning from developed to developing economies, there are
numerous reasons for expecting market prices to diverge even further
from the values of social costs and benefits (Little and Mirrlees, 1968,
vol. ii, ch. ii: UNIDO, 1972, ch. 2). Some of the more important
reasons for expecting such a divergence in LDCs are:

(1) Price controls and other measures to combat inflation distort

relative prices.
(2) Overvaluation of the currency - often resulting from inflation
with an inflexible exchange rate - makes imports ‘cheap’ and
domestic products ‘dear’ at market prices.
(3) Market prices reflect the existing distribution of income within
the economy, which may be socially sub-optimal. They also fail
to reflect externalities.
336 AGRICULTURE AND ECONOMIC DEVELOPMENT

(4) Market values represent only minimum levels of consumer

satisfaction, i.e. they do not measure consumer surplus.
(5) Wages fail to reflect the real social costs of employment,
especially in agriculture, because of disguised unemployment.
(6) Contemporaneous interest rates vary by more than is necessary
to cover differential risks because capital markets are under¬
developed and imperfect.
(7) A divergence exists between the market rate of interest and the
social rate of discount due to an aggregate ‘savings gap’ (as
already explained in the previous section). The gap tends to be
wider in LDCs due to the greater prevalence of tax evasion or
less efficient collection.
(8) The argument that large projects can have important reper¬
cussions on profits elsewhere in the economy is especially
relevant in LDCs with comparatively undiversified economies
and small national products.
(9) Provided that the demand for its exports is less than perfectly
elastic, an export tax may yield a net social benefit in the country
where it is imposed (an application of optimum tariff theory).
This particularly applies to exporters of primary commodities
without close substitutes, who also enjoy a substantial share of
the total export market.
(10) Protective measures (tariffs, quotas, etc.) are widely employed
by LDCs to foster domestic industry; like currency overvalu¬
ation, this tends to make domestic manufactured products
expensive relative to their world market prices. Inter-industry
competition for resources is distorted by differing degrees of
protection afforded to different industries, so an industry which
is potentially capable of yielding a net social benefit may appear
to be unprofitable due to excessive protection of one of its
intermediate good suppliers - for example cement used by the
building industry, or purchased feeding stuffs used in intensive
poultry or pig production.

For all these reasons, and because the development strategy of

LDCs tends to be heavily reliant on public investment and public
projects, there is an especially strong case for LDCs to use investment
appraisal techniques which show the difference between private (or
commercial) profitability and social (or national economic) profita¬
bility. We now proceed to consider briefly the rationale and
methodology of ‘shadow pricing’ which tends to be an especially
important feature of project appraisal in LDCs.
 PLANNING AGRICULTURAL DEVELOPMENT 337

6.2.2 Shadow Pricing

Since market prices cannot be relied on to measure social benefits and
costs in LDCs, cost-benefit analysis uses real rather than money
prices for project appraisal. ‘Shadow’ or ‘accounting’ prices reflect
the real costs of project inputs and the real benefits of project outputs
to society.
Shadow prices can be broadly defined as the social values of
economic goods and services, as perceived by government on behalf
of the nation. These social values usually diverge from the market
prices of the same goods and services due to market imperfections
(causing resource misallocation), market failures (causing unemploy¬
ment and the existence of externalities) and the pursuit of government
policy objectives such as income redistribution. It is incorrect to think
of shadow prices as being the market prices which would prevail after
the removal of market imperfections and government intervention.
Shadow prices will equate with equilibrium market prices only under
the restrictive assumptions of perfect competition, no unemploy¬
ment, no externalities and a socially optimum income distribution.
Each one of these assumptions tends to be unrealistic in the real world
and the combination of all four is clearly unattainable.
In summary, then, shadow pricing serves the twofold purpose of
correcting for market distortions and market failure, and re-allocating
resources in response to government policy objectives (including
redistributional objectives). Shadow prices are synthetic in the sense
that, unlike market prices, they cannot be observed directly. There¬
fore, the estimation of shadow prices is an important function of
project appraisers. Provided a good or service has an observable
market price, this serves as a starting-point in estimating the shadow
price, despite the need for adjustment. However, in the case of
‘untraded’ goods and services (like family labour in subsistence
agriculture) it can be misleading to regard the shadow price as an
‘adjusted market price’, and an alternative method of estimation
must be used. Methods of estimation can best be clarified by specific
consideration of the shadow prices of labour, investment and foreign
exchange. Our treatment of this topic closely follows the method¬
ology of shadow pricing presented in ‘Guidelines for Project
Evaluation’ (UNIDO, 1972, chs. 14, 15 and 16).
Shadow price of investment. The shadow price (or social opportunity
cost) of investment, P1, may be defined as ‘the ratio of the social value
of investment to the social value of consumption’. In this context
social value signifies ‘the value of the relevant time-stream of
aggregate consumption benefits discounted back to the present at the
338 AGRICULTURE AND ECONOMIC DEVELOPMENT

social rate of discount’ (UNIDO, 1972, p. 341). A premium on

investment typically exists due to government’s inability to induce the
socially optimal level of saving and investment where SOC = STPR,
as discussed above. Thus, whereas unity is a limiting value of P1
consistent with the condition SOC = STPR, the value of the ratio of
the value of investment to the value of consumption is normally
expected to exceed 1. As well as being a function of SOC and STPR,
P1 is also a function of the marginal rate of reinvestment of business
profits. Representing this by s, and SOC and STPR by q and i,
respectively, P1 can be determined from the expression:

i — sq

Clearly, in the limiting case where i = q, P'=l. Moreover, it is

obvious that the magnitude of P1 relates directly both to the size of s
and the size of q — i (given q ^ 1).
Thus, instead of assuming that 1 money unit of investment is worth
the same as 1 money unit of consumption in project analysis, use of
the shadow price of investment enables the differential social values
of investment and consumption to be reflected in the appraisal.
Shadow wage. In an LDC the market wage may fail to measure the
social cost of labour for three reasons. First, in the traditional sector,
labour is allocated and rewarded not in conformity with the ‘rules of
competition’, but according to a customary norm or convention such
as a subsistence wage. Thus, even without visible unemployment,
there is a gap between the MP of labour in the traditional sector (i.e.
its direct opportunity cost there) and the modern sector wage (as paid
by private capitalist employers and government). If there is visible
unemployment in either sector, the argument is the same but
stronger. Since the MP of the unemployed is zero, by definition, the
modern sector wage must obviously exceed the direct opportunity
cost of labour.
Secondly, where increased public sector employment is financed by
taxation there is an income transfer from capitalists to workers which
reduces aggregate investment and expands aggregate consumption
(assuming that workers consume more and save less of their marginal
income than capitalists do). Provided the shadow price of investment,
P1, exceeds 1, this transfer creates additional indirect costs to be added
to the direct opportunity cost of labour in calculating the shadow
wage.
Thirdly, the same income transfer changes the distribution of
 PLANNING AGRICULTURAL DEVELOPMENT 339

workers’ consumption over time. Present consumption increases, but

future consumption declines due to the consequent fall in the current
rate of investment. If, in pursuit of government policy, a special
premium (or weight) is attached to increasing workers’ present
consumption, the present value of workers’ future consumption
forgone for every additional public sector job created in the present
must be included in the calculation of the social cost of labour.
More formally, the shadow wage can be defined at three successive
stages of approximation. At Stage 1, accounting only for direct
opportunity cost:
SW = Ws

where Ws = direct opportunity costs of labour (usually in the

traditional sector). At Stage 2, allowing also for indirect costs,
through the effect on saving;

sw = Ws + SC(P‘ - 1)WC
where Sc = capitalists’ MPS and Wc = capitalist sector wage. At Stage
3, allowing also for the redistribution of workers’ consumption in
time;
SW = Ws + SC(P‘ - 1)WC + m[Ws + (SCVW - 1)WC]

where m = redistributional premium (0 ^ m ^ 1) and Vw = PV of

workers’ future consumption forgone.
The significance of the Stage 2 approximation obviously depends
upon the value of P1 (at the limit where P1 = 1, the indirect cost is zero)
whereas the added cost at Stage 3 depends upon m (at the limit where
m = 0, there is no case for allowing for the redistribution of workers’
consumption in time). Stages 2 and 3 pose difficult problems of
parameter estimation for project analysts and decision-makers and,
in practice, estimation of the shadow wage may not proceed beyond
Stage 1.
Shadow price of foreign exchange. According to one view, the real
price of foreign exchange is no more than the ratio of the free market
price to the ‘official’ price set by government. However, the authors of
Guidelines for Project Evaluation (UNIDO, 1972, ch. 16) argue that
this is unsatisfactory since it is not only countries whose currencies
trade in the free market at a substantial discount who are concerned
about their balance of payments. The basic reason why foreign
currency is worth more than its face value at the official exchange rate
is that the supply is usually rationed by government. There is
consequently a relative scarcity of imported goods, and domestic
340 AGRICULTURE AND ECONOMIC DEVELOPMENT

consumers are prepared to pay more than the landed price of imports
expressed at the official exchange rate. For example, suppose that
tractors of a particular size and specification can be imported at a
landed price of £5000 each. Thus, given an official exchange rate of
£ 1 = RS 20, the landed cost (in rupees) is RS 100,000 per tractor. But
if, at the margin, large-scale farmers are prepared to pay RS 150,000
per tractor, each RS 1.0 worth of foreign exchange (at the official rate)
provides goods worth RS 1.5 in terms of domestic willingness to pay.
In this case, then, the shadow price of foreign exchange is 1.5.
Generalising to any number of commodities, the shadow price of
foreign exchange, PF, can be written:

PiD
PF=If, W

where fj = fraction of (rationed) foreign exchange allocated to

commodity i at the margin by purchasers of imported goods,
P,D = domestic market clearing price of i, and P,L = cif price of i in
domestic currency at official exchange rate.
Choice of numeraire. Consideration of the shadow price of foreign
exchange leads to the projects appraiser’s choice of unit of account or
numeraire. If the practise is followed of choosing the price of freely
convertible currency as the numeraire, as recommended by Little and
Mirrlees (LM) (1968, ch. vm) then the shadow pricing of foreign
exchange is redundant, since all project costs and benefits are valued
at their ‘world prices’. In this case, shadow pricing is transferred from
foreign currency per se to all traded goods and services either
purchased or sold by the project. One difficulty of adopting this
procedure is that of valuing non-traded goods and services - such as
indigenous building materials and power supplies - at world prices.
This is one reason for preferring the UNIDO alternative of valuing all
project goods and services at their domestic prices (market or
shadow) combined with the application of a shadow exchange rate to
imports and exports.

6.2.3 Classification of Projects: the Production and

Distribution Trade Off
Development projects are designed to achieve economic and social
goals. Even with specifically agricultural projects, the objectives differ
considerably. Thus, the emphasis may be on technological inno¬
vation, improvement of the rural infrastructure, the reform of
 PLANNING AGRICULTURAL DEVELOPMENT 341

agricultural institutions - such as the land tenure or marketing

system, the economic and social betterment of either a particular sub¬
section of the rural population (such as farmers with relatively little
land) or a particular geographical region of the country (Benjamin,
1981, ch. 3). A broad distinction can be made between two project
types:
(1) Projects emphasising increased aggregate production where the
distribution of incremental production and income is a secondary
consideration.
(2) Projects which differentially increase production or income in
one or more ‘target groups’ of the rural population where
increased aggregate production of the entire agricultural sector is
of secondary importance.
In a sense, all agricultural projects are designed to relieve rural
poverty through a general improvement of agricultural incomes.
However, projects emphasising increased aggregate production and
income rely on the so-called ‘trickle down effect’ to transfer part of
the benefits to socially disadvantaged sections of the rural popul¬
ation. It may be considered that this indirect method of benefiting the
poor is too uncertain and too slow, and there may be a case for
making the redistribution of income a primary project objective in
order to attack rural poverty more directly and more rapidly.
We have drawn a sharp dichotomy between the objectives of
increasing aggregate production and income redistribution to serve
an analytical argument. In practice, however, governments some¬
times endeavour to design a project to serve multiple objectives.
Because transferring income to the poor inevitably raises aggregate
consumption, this poses the familiar problem of a trade-off between
consumption and investment, or between present and future
consumption.
The orthodox approach to project analysis has been to assume that
the primary project objective is to benefit the national economy by
raising net aggregate production and consumption. The redistribu¬
tive consequences of the projects, or income transfers, are a
secondary consideration which may be ignored altogether. But this
approach is clearly open to the objection that it neglects the
distinction between a gain in national income and a gain in national
welfare.
Suppose that two agricultural projects with identical costs are in
competition. Project X is a technological-type project aimed at
boosting aggregate agricultural output without a major transfer of
342 AGRICULTURE AND ECONOMIC DEVELOPMENT

income to the agricultural sector in the long run. Project Y is a rural

development-type project which is less concerned with expanding the
marketed surplus of agricultural products and more concerned with
improving farm incomes as a means of advancing rural welfare. If the
orthodox approach to project appraisal is applied, emphasising
national project benefits to the neglect of regional or group benefits,
project X must inevitably be preferred to project Y. But if, in pursuit
of its agricultural policy and contrary to the orthodox approach,
government wishes to pursue production and income redistribution
objectives simultaneously, how can such a conflict of ‘project
benefits’ be resolved?
One alternative is to subject the comparison of national economic
benefits to an ad hoc ‘social interpretation’ (Benjamin, 1981, p. 154).
Thus, in our hypothetical example, the government uses subjective
judgement to decide whether the social benefits of transferring
income to project Y farmers outweigh the social costs of forgoing the
additional national aggregate consumption benefits associated with
project X.
Another and more sophisticated method of appraising projects
with multiple objectives is to incorporate different types of benefits in
an appropriately weighted aggregate benefit function (UNIDO,
1972, ch. 3). Aggregation involves the conversion of different mea¬
sures of benefit - additional consumption, employment, etc. - into a
common set of units by establishing equivalences between different
types of benefits. Suppose units of benefit type Bi is considered to
be equivalent to W2 units of B2 and also to W3 units of B3. Then the
aggregate benefit, B, can be written as:

B = W,B1 + W2B2 + W3B3

where Wl5 W2, W3 are weights assigned by the planners.

More generally:

B= i
i=1
WiB,

where W: are the complete set of national weights.

Extending our earlier simple hypothetical example will serve to
illustrate the use of the method. Suppose that, for both projects, Bx
represents national aggregate consumption benefits, whereas B2
represents regional consumption benefits (income transferred to
farmers). Suppose further that with a given test discount rate and
identical costs for both projects we have:

Bt = 1000 and B2 = 0 for project X

 PLANNING AGRICULTURAL DEVELOPMENT 343

and
Bj = 800 and B2 = 800 for project Y.

Suppose also that, for both projects:

W1 = 1.00, W2 = 0.50

This implies that the planners judge 50 pence worth (cents) of money
income redistributed to farmers to have the same social welfare value
as £ 1 ($1) worth of money income available to the nation at large in
the form of additional national aggregate consumption.
Hence, for project X:

Bx = (1.00) (1000) + (0.50) (0)

= 1000
whereas for project Y:

By = (1.00)(800) + (0.50)(800)
= 1200
Clearly, if B2 was the sole criterion of selection, to the exclusion of B2,
project X would be selected. But when Bj and B2 criteria are
combined, using the weights assigned by government to each type of
benefit, project Y is the rational choice. It is easy to see from this
example that even if the net benefits of a project at the national level
were zero (or even negative) the project might still be justified by its
redistribution benefits alone if the weight assigned to these was large
enough.
The assignment of weights to different types of project benefits is
subjective depending on the government’s value-judgements. But a
major difference between the aggregate benefit function and the ad hoc
approach is that, in using the former, the same set of weights can be
used in appraising different projects, so achieving greater consistency
of treatment. It may not always be feasible to express all types of
project benefit in terms of a single unit of account such as money.
Thus, for example, extra employment (which may be valued for its
own sake rather than for its contribution to national income) is
measured in physical terms. However, the problem of aggregating
benefits measured in different units can be circumvented by deriving a
set of equivalences or conversion coefficients with a common base
(see UNIDO, 1972, p. 34)

6.2.4 Project Formulation Process

Economic appraisal is but one aspect of a much broader process
of project formulation. The starting-point of most projects is
344 AGRICULTURE AND ECONOMIC DEVELOPMENT

a set of technical proposals based - in the case of agricultural

projects - largely on the work of agronomists, livestock specialists
and engineers. The technical options must be identified before the
accountants can conduct a financial appraisal and the economists an
economic appraisal. Thus the preparation of a development project is
a complex multi-stage process: it is also a multi-disciplinary activity
in which the economist builds upon the feasible technical options,
whilst the agronomist or engineer must work within the bounds of
economic feasibility.
A 5-stage project cycle is discussed in the literature (Benjamin,
1981, ch. 1). Stage 1 is ‘project concept definition’ in which govern¬
ment accepts development through projects as a feasible means of
promoting development. Stage 2 is ‘identification’ in which the outline
proposals for a project are delineated. Stage 3 is ‘preparation’ in
which proposals identified at the previous stage are expanded and
examined in detail. At this stage the full technical and organisational
details of the project are presented, together with the expected
financial results and the economic justification. Preparation is both
time-consuming and costly. Stage 4 is ‘appraisal’ to check further the
economic viability of the project, including its economic justification
in competition with other projects. Since the external financing of
projects is often conditional upon meeting a standard of economic
justification which is acceptable to the lending or aid agency,
appraisal may be undertaken independently of preparation by an
external agency. Negotiations for external project finance (loan or
grant) are inextricably linked with appraisal. Stage 5 is ‘supervision’
which is concerned with all the details of project implementation,
including the monitoring of progress reports and even modifying
details of the project itself during implementation if this is deemed to
be necessary in response to unforeseen contingencies.
Like other stage theories, this one is open to criticism. It is admitted
that ‘in practice many of the phases overlap’ (ibid., p. 1), as is all but
self-evident with respect to identification, preparation and appraisal.
More interestingly, Benjamin maintains that, in practice, appraisal
tends to be ‘employed post facto to justify a project already selected
and designed’ (ibid., p. 155). This is, of course, contrary to the
economic ideal of using appraisal to discriminate between alternative
projects.
Despite the possibility of disagreement between economists and
other practitioners of project planning about the purpose of
economic appraisal, the proposition that project formulation and
implementation consist of much more than economic appraisal is
 PLANNING AGRICULTURAL DEVELOPMENT 345

beyond dispute. Most projects have a complex technical structure, as

well as being based upon a set of interrelated economic assumptions,
so much so that if the technical assumptions are unreliable, but the
degree of possible error is unknown (which is often the case), little
credence can be given to the results of economic appraisal. This is so
regardless of the reliability of the economic assumptions and the
professional skill of the economists making the appraisal.

6.3 Practical application: appraisal methods

In this section we discuss the application of comprehensive models of

project appraisal to agricultural projects, using the UNIDO model as
an example. We then consider some peculiarities of agricultural
projects and the case for using simpler appraisal technique, with
examples. Finally, we can contrast ex ante appraisal with ex post
evaluation and the feedback from project experience to project
design.

6.3.1 Comprehensive Models: Application of the UNIDO

Method of Appraisal to an Agricultural Project
In this section we describe and summarise the elements of a
case study borrowed from the UNIDO Guidelines for Project
Evaluation (1972, ch. 21) to the authors of which we make due
acknowledgement.
Description of project. This is a case study of the appraisal of an
irrigation project in a particular geographical region of an LDC. The
project is designed to achieve three principal objectives:

(1) The intensification of agricultural production in an arid river

valley by damming the river and constructing a water distri¬
bution network.
(2) Increasing the number of farmers in the valley by the settlement
of farmers from outside the region following redistribution of the
land.
(3) The social betterment of the region’s inhabitants through the
construction of a road network, an urban centre and satellite
villages.

The project would be administered by a statutory water authority

formed by the government specifically to develop the region’s water
resources. Regarding finance, the water authority would apply to the
government and to the World Bank for long-term loans to cover the
346 AGRICULTURE AND ECONOMIC DEVELOPMENT

capital costs of the project. The World Bank would be asked to cover
the foreign exchange component of the total investment and the
home government the remainder.
This case study exemplifies the appraisal of net benefits at the levels
of: (i) the national economy, (ii) the region, and (iii) a particular
‘target’ social group (small farmers).
Methodology.We briefly consider the following aspects.
(a) framework of government social objectives,
(b) data requirements,
(c) appraisal procedure (in 3 stages),
(d) presentation of results,
(e) interpretation of results, and
(f) summary view of the method.

(a) Statement of the government's social objectives. Consistency

demands that detailed project proposals should be prefaced by an
explicit statement of the government’s social objectives. In the
present case those are:
(1) Increased aggregate consumption to raise the average standard
of living in the country as a whole (national welfare objective).
(2) Redistribution of income to the region in which the project is
located (regional welfare objective).
(3) Redistribution of income to small farmers (group welfare
objective).
(4) Creation of new employment opportunities.
(5) Provision or improvement of basic welfare facilities, such as
housing.
(6) An improved balance of payments.

Analysis of the project actually revolves around its expected contri¬

bution to realising the first three objectives since, in the words of the
authors, ‘the last three items do not so much represent separable
objectives as observations on the limited ability of market
prices - the wage rate, the price of social services and the foreign
exchange rate - to reflect true social benefits and costs with respect to
the aggregate consumption objective’. Thus the realisation of objec¬
tives (4)-(6) is sought indirectly by substituting shadow prices for
market prices, where the latter are not thought to reflect true social
values.
(b) Data requirements. Two basic types of data are required. First,
data are needed to construct a table of resource flows generated by
 PLANNING AGRICULTURAL DEVELOPMENT 347

the project. This table shows the expected monetary values at market
prices of all project benefits and costs, grouped and totalled by
major categories on a year-by-year basis for the life of the project.
Also shown in the table are the expected monetary transfers between
social classes, or between citizens and government. At this stage of
the analysis the expected values of all future resource flows remain
undiscounted. The form of presentation is exemplified by Table 11.3
reproduced from UNIDO (1973). The bare contents of the table
conceal the amount of detailed work needed to prepare it. First, a
vast quantity of data on input availabilities, feasible input mixes,
product choices, and product mixes, physical input-output co¬
efficients and prices, must be collected. Second, these basic data
must be analysed and processed in order to produce estimates of
the expected levels of output (or benefit) by time period, together
with the corresponding costs. In order to suit the form of the
subsequent analysis, a detailed breakdown of each year’s costs is
needed, the main distinctions being between capital and current (or
operating) costs, and costs borne by the principal ‘actors’ of
government, farmers and taxpayers.
As far as economic appraisal per se is concerned, the table of
resource flows might be regarded as technical data which the
economist responsible for appraisal must accept without question.
This view probably exaggerates the division of labour between the
economist and specialists from other disciplines who are also
involved in project analysis. The important point is that, as one of the
most basic documents used in the process of appraisal, the results of
appraisal are to an important degree predetermined by the table of
resource flows. Most important of all, the reliability of the final result
is critically dependent on the quality of the information available to
the analysts responsible for preparing this table and their skill in
using it to produce trustworthy estimates.
The second basic data requirement is a set of ‘national para¬
meters' for use in evaluating the project within the context of the
government’s social objectives. Examples include the social rate of
discount, the shadow prices of investment, labour and foreign
exchange, and the weights assigned by government to dissimilar
project benefits such as increased per capita consumption and income
redistribution (regional and social). We have already discussed the
welfare economics rationale of using subjectively determined weights
in project analysis. The set of national parameters may also include
certain more objectively based coefficients, such as the marginal
 348 AGRICULTURE AND ECONOMIC DEVELOPMENT

O04©00~*v004r0©©© o oo o Os
o o © o
vo p O oo 8 voos©TfcoTtoqpcqpp
i-id'-ooot^^odNO'OO
o X
vd o
, P oo
o os
I - (N (N 1 (N
m On m

O © 04 © OO — vO 04 m © © © © oo Os
o o o oc
VO VO 00 ©vopppppppppp1'! 8 X , s
— On (N fN cd^f^-Nrioovdocn — vd o
1
o
<N
Os
n On ro

5
rf O O
vo oo oo
goon. o^^r-tNcc^'cio
OMOOOO'-^'OCOOO o 00
o X 1
O Os
P OO
Tf — © vd o o oo
OO ON OC (N
m 00(N Tf ^ 1 04

00 00 O O poovqr^vq^gpOO
O © 04 ___ „
__ _ o oo o o
04 vq vq oo ©vOOs©pXOspppp©> o X 1 p p
Oo c/S os «n 04 -OOOC^TfVO^O vd o
1
o o-
04
© r- 04

04 04 © ©
—« p poo
§ vo os
0 04 oc m 04 OO 04 — ©
h M O - ^ 8
oo
p
1
s
S vd
<n O' (N ri O — O m 04 Tfr vd o 1 o
O' no rs 04

<3 vo VO O O go 04 -
vO OS I
ooOsooOsrfooOO
rf p p p 04 p p © 8
oo
p
m
p
vo
os r- 04 oo
00 OS On 04 Tt —: O dooo^^ONOn vd o 11 S
O •o]
0 - — 04 — — 04
Table 11.3: Benefit, cost and transfer flows by year (at market prices) (million pesetas)

o o o
■ oo
SO <N O OO vO
vOOsOppppp
oo
8 X P
OO 00 i 8
os vd 04 rf-^o^-o^todeNTtodordr^ O o
m
1 O
04
vO — 04

o o o o o
© p cq p p
oi rf wd os oi
i/l NO M ^

O O Os O O
O^h^O
0C OS —^ rn Tt
oo VO n 1C

Source: UNIDO (1972) Guidelines for Project Evaluation, Table 21.19.

o o o o
<N 8 P (S| Tf >C
uortcnoN^
r- 04 — — 04

O O O O
ro h h ro
O’—Os 04
04 —< 04

CD
3 0>
8 3 tsP O 3
v> O oc
E o
o « -9
ce
O o
o O
"O
8 <u 15 3
o , 03 <L> <D <L>
■c w> 2p j3< 9"'
03
W) * oc s- - oc c
3
a 3 O
C 33 cd c 5
2 w, 3 -3 •3 3 g
3 —
-2 =3 IS 3 X 3 O.X g
o c/5
O
a
3 3 C/5
O
X 3 3 cS
X)
I^ U O
8
3
O o p
15(M S-g O -T3 cd o o 'C 3 c S‘5 00 US -o
0C 3 cS
I" y TP -O ?n
c 33 ^3 H> GO cd w §>* .a &5 .. 3 c
+—■ OC
2 M « •X M *> P '5 »- 3 ^ •« — 3 0^
3 <L» 3
2 6 a <d i: ^ 2 3
o ‘c/5 t % rp« § 8 2 S £
c/5 3 v O o £ a3 ■-
c/5
3 o .y a 3 15
C p o
oc o Uh Q u. CLlI £ :s ^ n- E.|>c
O C a)
03 ^ ^
< X u o Uh < U - 06
Ph DC jr-._.
QU, j vTQ ST Jw Q fc C/5 Q U-
C i i I I I I I I I I I § —
cq v Cj n n m nn?’t'trf^fi’»TOinioS'0'0'oP k °° e^c.
 PLANNING AGRICULTURAL DEVELOPMENT 349

opportunity cost of capital in the private sector (q), the marginal rate
of reinvestment (s), and the marginal propensity to save (MPS)
amongst different categories of savers.
An example set of national parameters is shown in Table 11.4,
reproduced from the same source as Table 11.3. Adequately reliable
data at an acceptable cost is again critical. Highly inaccurate or
unrealistic national parameters, applied at a later stage of the
analysis, must inevitably undermine the credibility of the final results,
(c) Appraisal procedure. In this case study appraisal of the project
is based on three criteria: (i) net national aggregate consumption
benefits, (ii) regional redistribution benefits, and (iii) social group
redistribution benefits. In explaining the method of appraisal it is
convenient to deal with each criterion separately before proceeding
to discuss overall assessment of the project from values inserted in
an aggregate benefit function.
(i) Appraisal of net aggregate consumption benefits. This part of the
analysis is best explained in three successive stages. In Stage 1, all the
component elements of the net aggregate consumption function are
valued at market prices, on the simplifying assumption that these
adequately reflect social opportunity costs. Thus, we have the
identity:

MC = (1) + (2) — (3) — (4) — (5) — (6) — (7)... (a)

Table 11.4: Values of general parameters

(1) Foreign exchange premium (p = + 1.0

(2) Unskilled labour premium A = — 1.0
(3) Extension worker premium AE= + 1.0
(4) Marginal rate of return on investment
in private sector q = 0.20
(5) Marginal rate of reinvestment of profits s = 0.20
(6) Social rate of discount d = 0.05; 0.075; 0.010
(7) Associated social price of investment P' = 16.00; 4.57; 2.67
(8) Marginal propensities to save:
(8-a) Farmers SF = 0.2
(8-b) Unskilled workers SL = 0.0
(8-c) Government SG= 1.0
(8-d) Taxed public ST = 0.8
(9) Marginal propensity to (re-) spend in region y = 0.2
(10) Weights on objectives:
(10-a) Aggregate consumption (f= 1.00
(10-b) Redistribution to region
(10-c) Redistribution to small farmers
CF=100
0 = 0.50

Source: Adapted from UNIDO Guidelines for Project Evaluation, Table 21.20
350 AGRICULTURE AND ECONOMIC DEVELOPMENT

where MC = net aggregate consumption at market prices, and

(1), (2), etc. are resource flows specified and quantified in Table
11.3.
In Stage 2 the shadow prices of labour and foreign exchange are
substituted for their market prices, except that no allowance is made
for the premium on investment (on the simplifying assumption of a
zero savings gap). The identity of components of the net aggregate
consumption function now becomes:
SC = (1) + <p{ 1-F) + (2) - (3) - A(3-L) - cp(3-F)
— (4) — A(4-L) — ip(4-F) — (5)
— A[(5-LF) + (5-LH)]
- (p( 5-F) - (6) - AE(6-S) - </>( 6-F) - (7)

where (1), (1-F), (2), etc. are resource flows from Table 11.3,
(p = shadow price of foreign exchange, X — shadow of wage of
unskilled labour, and XE = shadow wage of agricultural extension
personnel.
Re-arranging terms:
SC - MC + <pFE + XL + AEE- • • (b)

where FE = (1-F) — (3-F) — (4-F) — (5-F) — (6-F),

L = - (3-L) - (4-L) - (5-LF) - (5-LH) and E = - (6-S).
Note that whereas the shadow wage of unskilled labour, X, is expected
to be at a discount compared with the market wage rate, the shadow
wage of extension personnel, AE, is expected to be at a premium. All
shadow prices are expressed relative to a market price of unity.
In Stage 3 the substitution of shadow prices for market prices is
completed by also bringing in the shadow price of investment.
Assuming that the marginal propensity to save differs amongst social
classes, as well as between citizens and government, assessment at this
stage entails the allocation of all SC benefit and cost flows amongst
four groups of gainers or losers, as follows:
SC = SCF + SCL + SCG + SCT
where the superscripts F, L, G and T identify four classes of net
consumption benefit recipients, farmers, unskilled workers, govern¬
ment and taxpayers. Moreover:

SCF = (1) + (2) — (5) — (6-S) — (7) + (8)

- (9) - (10) - A(5-LF) - AE(6-S)
 PLANNING AGRICULTURAL DEVELOPMENT 351

SCL = - A[(3-L) + (4-L) + (5-LH)]

SCG = - (3-F) - (4) - (6-D) - (6-F)
~ (8) + (9) + (10) + <p(FE)
SCT = - (3-L) - (3-S) - (3-D)
Then, if the average farmer saves a proportion SF of his marginal
gains, the (adjusted) social value of the net consumption benefit going
to farmers is:

CF = [SFP' + (1-SF)]SCF
where P1 = shadow price of investment.
Similarly:

CL = [SFP, + (1-SL)]SCL
CG = [SGP' + (1-SG)]SCG
CT = [STP, + (1-ST)]SCT
Thus, with full adjustment for divergencies between the market prices
and social opportunity costs of labour, foreign exchange and
investment (= current consumption foregone), we have:
C = Cr + CL + CcrCT- (c)
which can be re-written as:

C = SC + (P'-1)[SFSCF + SLSCL + SGSCG + STSCT]

(ii) Appraisal of regional distribution benefits. There are two steps in
the assessment procedure. First, we assess the net aggregate con¬
sumption benefits redistributed to the region of the project in any
given year. Thus:

DR = (1) + (2) + (3-L) + (3-S) + (4-L) + (4-S)

-(5) + (5-LF) + (5-LH) + (6-S)
— (7) -F (8) — (9) — (10)

We note that the project benefits (1) and (2), which were previously
counted as benefits to the nation, are now treated as benefits to the
region. Of even more interest, some items previously treated as costs
to the nation, including the costs of all labour employed by the
project - (3-L), (3-S), (4-L), (4-S), etc. - are now counted as benefits
to the region. Second, we adjust DR for the marginal propensity to
respend in the region (= regional income multiplier).
352 AGRICULTURE AND ECONOMIC DEVELOPMENT

(d)

where y — marginal propensity to respend and 1/1 — y is the regional

income multiplier.
(iii) Appraisal of social group redistribution benefits. Small farmers
are the target group to whose improved welfare redistribution is
directed. The total value of net consumption benefits provided to
small farmers by the project is given by:
RSF = A[(l) - (5) + (5-LF) - (9) - (10)] + B(2)
— C(7)... (e)
where A = small farmers’ aggregate share of the project’s cultivated
area, B = number of small farmers as a proportion of the total
number affected by the project, and C = small farmers’ aggregate
share of the land taken over by the project.
Small farmers enjoy a share of the agricultural output (1) and
employment benefits (5-LF), as well as part of the benefits of
housing and social services (2). But they also bear a share of the non¬
labour agricultural costs, and the agricultural disbenefits of the
transfer of farm land to non-agricultural uses (houses, roads, etc.) by
the project.
Estimates of the project net benefits accruing to the nation, the
region and small farmers during each year of the project’s life are now
complete. The final stage of the appraisal procedure is to discount
and sum all annual net benefits in order to find the NPV of the project
as follows:

(1) Take the values of the social rate of discount and other national
parameters (assumed constant over time for simplicity) from
Table 11.4.
(2) Use the SRD to find the NPV of each time unit of benefit, cost
or transfer shown in Table 11.3, and sum to find the NPVs of
the corresponding time flows.
(3) Substitute the resulting NPVs together with the appropriate
national parameters where required, in equations (a)-(e).

The results of following this procedure are summarised in

Table 11.5.
(d) Presentation of results. The results are shown at three different
social rates of discount as well as separately for each of the three
 PLANNING AGRICULTURAL DEVELOPMENT 353

separate appraisal criteria of national, regional and social group

redistribution benefits. Looking first at aggregate consumption (na¬
tional) benefits, alone, the results show that with full adjustment for
divergencies between market prices and the social opportunity costs
of labour, foreign exchange and investment (appraisal criterion C)
the project is ‘justified’ only at the lowest of the three test social rates
of discount (5 per cent). The effects of full adjustment are brought out
by comparing the ‘no adjustment’ criterion MC, and the partial adjust¬
ment parameter SC with the corresponding values of criterion C.
Generally, SC > MC > C in this case for two reasons. First, market
prices are judged to overvalue unskilled labour and foreign exchange.
Second, the project redistributes income in favour of ‘low savers’
despite a relatively high P1. Turning now from aggregate consump¬
tion to redistribution benefits (regional and social group), the project
is justified on both redistributional critieria at all three SRDs. It is
apparent, then, that if the test SRD is judged to be at either of the two
higher rates the project cannot be justified economically unless the
benefits of redistribution (regional and group) are judged to out¬
weigh the disbenefits of the expected loss of net aggregate consump¬
tion benefits. This brings us back to the aggregate benefit function, 6,
and the weighting of its components.
The weighted sums of project benefits, or the aggregate benefit, B,
are shown in the bottom line of Table 11.5. Although there is a strong

Table 11.5: Net present value of regional irrigation project in year 0 (million
pesetas)

Social rate of discount

Item Equation 5% 7j% 10%

Aggregate consumption1
MC a 572 219 39
SC b 1030 506 234
C c 467 - 164 -276
Regional redistribution
rM d 1701 1158 855
Social group redistribution
RSF e 887 555 374
Aggregate benefits
B (2) 1336 403 125

1 MC, SC and C are defined by equations (a), (b) and (c) pp. 349-51.
2 B = CQf + RM0RM + RSF(rF, with weights on objectives 0°, 0RM and 0RSF taken from
Table 11.4.
Source: Adapted from UNIDO (1972), Guidelines for Project Evaluation, Table 21.22.
354 AGRICULTURE AND ECONOMIC DEVELOPMENT

inverse relationship between B and SRD, the project is justified by a

narrow margin even at an SRD of 10 per cent. Thus, in the case of this
project, national disbenefits at higher rates of SRD are judged to be
outweighed by positive redistributional benefits.
(e) Interpretation of results. The method of appraisal we have
outlined so far is to find the NPV of the project as a whole, subject to
given values of the national parameters, i.e. the SRD and the weights
on objectives. An alternative approach is to find combinations of
national parameter values at which the project NPV is zero. Recalling
that the condition NPV = 0 defines the internal rate of return, the
approach is to find combinations of the IRR and other national
parameters which ‘switch’ the project from being viable to being
nonviable (Pearce and Nash, 1981, p. 174).
More formally, define:

NPV = £I (1 + d)J
j '

where is net benefit type i in year j, 9I} is the weight given to B(j, and
d is the social rate of discount.
The first approach is to find NPV given d and the sets 0;j and B^.
The alternative approach is to substitute 0 for NPV on the LHS of the
above identity. Then, given the B;j, solve for either 0y or d ( = internal
rate of return) on the RHS. A supposed advantage of this approach
is that decision-makers can learn to judge, by experience, whether the
internal rate of return implied by a given set of weights (or the set of
weights implied by a given IRR) is acceptable. For practical reasons,
the size of the weight set that can be conveniently handled by this
method is only quite small. The ‘switching values’ of the SRD and
other national parameters can also be presented graphically for ease
of interpretation (UNIDO, 1972, p. 352).
Even if the project passes the test of economic viability at accepted
levels of the national parameters, this does not necessarily imply that
it should be undertaken in the precise form of the original proposal.
Relevant alternative forms need to be explored. So, for example, the
present case study project could be made to yield a higher expected
level of net national aggregate consumption benefits by means of two
specific changes in project design, as follows:

(1) changing the land distribution pattern in favour of large farmers

(and to the deteriment of small farmers), and
(2) changing the cropping pattern to increase the ratio of cash crops
to subsistence crops.
 PLANNING AGRICULTURAL DEVELOPMENT 355

The same changes in project design would also yield an improve¬

ment in regional redistribution benefits. The cost of project redesign
in this case would be borne by small farmers (the target social group
in the original project design) who would inevitably suffer a
substantial loss of redistribution benefits. However, despite a lower
RSF (possibly negative), because C and RM are both higher, the overall
net benefits of the redesigned project could still be better than the
original, depending upon the relative magnitudes of (weighted)
positive and negative effects.
The foregoing discussion is limited to only one of many possible
project redesign options. The total number of options which it is
feasible to compare is limited by practical considerations. (Our
discussion of project risk and uncertainty (section 6.1.3) included
some more general comments on forms of ‘sensitivity analysis’.)
(g) UNIDO method of project appraisal: summary view. Compared
with simpler or more ad hoc methods of project appraisal, compre¬
hensive methods, like the UNIDO model, offer at least two potential
advantages:

(1) their greater emphasis on the distinction between the financial

and economic merits of projects through the use of relatively
sophisticated methods of shadow pricing; and
(2) their ability to combine measures of economic (or aggregate
income) and social (or redistribution) benefit in the appraisal
procedure systematically.
However, comprehensive methods are more costly than simpler ones
in terms of their technical complexity. Also despite their sophistic¬
ation, the application of comprehensive methods remains substanti¬
ally reliant on subjective judgements by those responsible for
selecting projects.
The UNIDO method is remarkable not only for its technical
complexity and its dependence on political value-judgements, but
also for its pragmatism. The method’s techniques of shadow pricing
as described in Section 2.2 typify its technical complexity. The
dependence on political value judgements is a consequence of the
method of assigning weights to objectives. Although those responsible
for selecting projects may learn to achieve a measure of consistency in
the choice of ‘switching values’, there is no truly objective method of
weighting competing types of project benefits, such as (a) augmenting,
and (b) redistributing the national product.
One of the most pragmatic features of the UNIDO method is its
choice of unit of account, or numeraire. Whereas a rival method of
356 AGRICULTURE AND ECONOMIC DEVELOPMENT

comprehensive appraisal (the Little-Mirrlees method) effectively

values all project costs and benefits at ‘world prices’, the UNIDO
choice of numeraire is the more realistic one of net aggregate
consumption benefits at domestic prices (either the market price or
the equivalent shadow price, as appropriate).
Comprehensive methods of project analysis, such as the UNIDO
method, have been the subject of substantial criticism on both
theoretical and practical grounds. The use of distributive weighting
systems has been a particularly heavily criticised aspect. Although a
number of more disaggregated approaches to project appraisal have
been advanced by the critics, none of these has been proved to be
indisputably superior to CBA. Pearce and Nash (1981), chapters 1
and 3, contains an excellent discussion of the issues involved.

6.3.2 Pecularities of Agricultural Projects:

The Case for Simpler Methods of Appraisal
None of the main methods of project appraisal has been developed
specifically for application to agricultural projects. So, are agricul¬
tural projects sufficiently peculiar or different from other projects to
render these methods unsuitable for use in agriculture? Certainly
there is some substance in the argument that agricultural projects
‘flaunt their difficulties’ (Hirschman, 1968, ch. 5).
Good agricultural planning depends on a reliable data base, but
this is generally lacking in LDCs. Reliable production and price data
for agricultural inputs and products are hard to come by even in the
modern agricultural sector. In the traditional sector the problem of
inadequate data is usually far worse. All these problems add to the
difficulty of making reliable estimates of agricultural project costs
and benefits in the future, as required for the purpose of constructing
the most basic of project planning documents, the table of resource
flows.
Critics of the use of sophisticated methods of appraisal (such as
UNIDO and LM) for selecting agricultural projects in LDCs
maintain that, regardless of the strength of the a priori case for
refinements such as shadow pricing and estimating the income
redistribution effects of projects, there is no point in making these fine
adjustments in practice unless the unadjusted, undistributed financial
flows are adequately reliable. As far as agricultural projects are
concerned, scarce planning skills are better concentrated upon
improving the quality and reliability of physical production esti¬
mates, cost estimates and market price forecasts than being diverted
to less urgent and fundamental planning activities of the kind
 PLANNING AGRICULTURAL DEVELOPMENT 357

highlighted by UNIDO and similar advanced appraisal methods.

There is the further argument that because of their technical
complexity, the application of advanced methods of appraisal may be
beyond the capacity of agricultural project planners in LDCs.
Alternative methods of appraising agricultural projects. These and
other criticisms of economically sophisticated methods of appraisal
have led to the advocacy of simpler methods for use with agricultural
projects (Carruthers and Clayton, 1977).1 The so-called decision
matrix method exemplifies this approach. The advantage of this
method is not that the number of criteria used in appraising the
project is reduced, but that the individual criteria are less complex
and they are considered together without formality. The main criteria
are:

(1) An expected economic internal rate of return (with limited

shadow pricing) but hedged with details of the forecast error
distribution.
(2) The expected financial internal rates of return (similarly hedged)
(i) to farmers (or other private beneficiaries) and
(ii) to government.
(3) Additional employment created per unit of investment.
(4) Share of project income accruing to the poorest quartile (or other
predetermined proportion) of the population.
(5) Is the project located in a special development area? (Yes or No).
The cost of the project in terms of foreign exchange is an additional
criterion applied to nearly all projects. On shadow prices, Carruthers
and Clayton recommend ‘a simple pragmatic approach, with prices
adjusted in the right direction to the approximate order of magni¬
tude.’ Thus the shadow wage will usually be below the level of the
money wage, whereas the shadow price of foreign exchange will
normally be above the level of the official exchange rate. The actual
amount of the discount (shadow wage) or premium (shadow
exchange rate), which can be no more than a rough approximation,
can be left to judgement and experience.
As already indicated, when using the decision matrix approach, the
relative weighting of the various criteria is left to the informal
judgement of politicians advised by planners.
Other alternatives to the UNIDO method of project selection are
more narrowly based on a single major appraisal criterion, though
not to the complete exclusion of other criteria. Thus, programming
models are used to maximise a single priority objective (such as
national per capita income) subject to minimum performance levels in
358 AGRICULTURE AND ECONOMIC DEVELOPMENT

associated objectives (such as minimum wages in a particular group).

Even simpler, is the earlier mentioned practitioner’s approach, which
is heavily biased towards selection on the basis of a single
criterion — such as the economic internal rate of return - but subject
to a measure of informal ‘social interpretation’ (Benjamin, 1981,
ch. 11). Part iv of Benjamin’s book contains a number of case studies
of agricultural projects using this approach.
In the end, the choice of method for appraising agricultural
projects is likely to be influenced by many considerations - not least,
the character of government’s programme of economic and social
development, the state of the agricultural information base, and the
amount and quality of specialised project planning expertise at
government’s disposal. It is clear that, in principle, economically
sophisticated methods, such as the UNIDO model, are superior to
simpler but cruder methods. Whether the more advanced methods
are superior in practice depends on the ability of planners and
politicians to apply and use them correctly. Their ability should
improve with experience and as better resources for planning become
available. In fairness to the economists who have developed the more
advanced appraisal methods it may be added that, despite their
apparent preoccupation with the ‘fine tuning’ of project benefits and
costs, they are not unaware of other equally important aspects of the
planning process. For example, Square and van der Tak (1976),
remark that

The judicious use of shadow prices is an important means of assessing the

economic merits of a project... but it is not a substitute for careful analysis of
its technical features, investment and operating costs, organisational arrange¬
ments, market prospects, financial results and many other considerations
relevant to the outcome of the project, (p. 14).

6.3.3 Ex ante and ex post Evaluation

Another aspect of project planning is ‘learning by experience’. A
useful terminological distinction may be made between ex ante
appraisal and ex post evaluation of projects. (But note that this
distinctive usage of the terms ‘appraisal’ and ‘evaluation’ is not
universal: they are often used synonymously.) The point is that pre¬
implementation appraisal and post-implementation evaluation are
complementary with evaluation providing feed-back information.
Carruthers and Clayton (1977) distinguish two distinctive types of
feedback:
(1) Feedback to the project itself, using comparisons between the
 PLANNING AGRICULTURAL DEVELOPMENT 359

initial objectives, predicted performance and the actual achieve¬

ment of objectives and performances as a yardstick.
(2) Feedback to the planning process by comparing project achieve¬
ments with the goals of current policy (which may diverge from
past goals due to shifting priorities).
Whereas some mistakes in project design can be learned by
experience, others might be avoided by more careful preparation. For
instance, an agricultural project may fail because the planners have
overlooked important institutional constraints. Witness the example
of an agricultural project in West Africa which assumed that on
peasant family farms the labour of husbands and wives would
substitute freely. The customary division of both land and labour
between the sexes, and the fact that wives are not obliged to work on
their husband’s land, were both overlooked (Dey, 1982). This
example serves to warn against the error of assuming that models of
economic behaviour, particularly western models, are freely transfer¬
able amongst different countries and cultures.

6.4 Project analysis and policy reform measures

Project analysis is open to two general classes of criticism (Stewart,

1978). So called first-order criticisms question its very raison d’etre by
arguing that instead of tinkering with problems of price distortion,
unemployment and externalities, by attempting to make compensat¬
ing adjustments in the appraisal of investment projects, governments
should attack the problems head-on with policies to remove the
causes of market imperfections. Second-order criticisms are con¬
cerned with technical details of the methods used whilst accepting the
principle of attempting to correct for the distorting effect of market
imperfections. Having already examined the methodology of project
appraisal in some detail, we do not propose to further pursue the
second-order criticisms at this stage, but the first-order criticisms
deserve brief amplification. Stewart places these in four major
categories.
First, why cannot governments achieve their economic and social
objectives by means of direct action on prices to bring them into line
with social values, instead of merely adjusting prices as part of the
appraisal procedure? Appropriate price, tax and tariff adjustments by
government would be more efficient than applying elaborate rules to
correct for ‘incorrect’ prices, which merely diverts attention from the
need to remove market distortions.
360 AGRICULTURE AND ECONOMIC DEVELOPMENT

Government’s apparent inability to enforce its will might be

interpreted as casting doubt on whether its publicly stated policy
objectives (forming the basis of project appraisal) are its real
objectives. If the real objectives are different (and remain concealed)
then project appraisal is no more than political propaganda.
Second, if government cannot achieve success by direct market
intervention, perhaps because its policy fails to command an
adequate public consensus, is indirect action - using the artificial
device of project appraisal - any more likely to be successful?
Third, there are the linked questions of whether the rules of project
selection specified by appraisal methodology are compatible with
how governments reach decisions in the real world and, even if they
are, whether the government’s value-judgements reflect society’s
social values. Regarding the rules of decision-making, it is unrealistic
to assume either that economic decisions are made purely in the light
of economic criteria, or that the politicians who make decisions are
motivated only by altruism. Short-term political advantage may
loom larger than longer-term goals of economic and social welfare in
deciding where and when projects are approved, and politicians with
national responsibilities may still be influenced by regional loyalties.
Regarding the value-judgements which are such a prominent feature
of comprehensive methods of project appraisal and selection, it is
generally assumed that government judgements reflect social consen¬
sus values. But to the extent that reality is represented not by national
unity and social consensus but by a plural society with conflicting
interest, what purport to be social decisions are really government
decisions (reflecting the values of the ruling party and its supporters).
Fourth, is the criticism that, because of its static approach to
resource allocation, project analysis is incapable of being adapted to
dynamic adjustments necessitated by uncertainty. Comparisons of ex
post evaluation with ex ante appraisal show that major divergencies
frequently occur between expectations and actual fulfillment due to
unforeseen (and often unforeseeable) events.2 This, say the critics,
makes nonsense of appraisal. Moreover, due to its pre-occupation
with improving static resource allocation, project appraisal misses
the contributions of ‘learning by doing’, and improved management,
to long-term growth and development.
What are the counter-arguments in response to these criticisms of
project analysis and appraisal? On the question of the choice between
direct and indirect measures to achieve government’s development
objectives, no government is omnipotent: an elected government
must rule by persuasion. Because of the ‘isolation paradox’, private
 PLANNING AGRICULTURAL DEVELOPMENT 361

citizens and firms tend to be more myopic than the government. So,
for example, it is extremely difficult for government to close the
‘savings gap’ by direct intervention (i.e. fiscal and monetary policy).
On the question of why indirect measures can succeed where direct
intervention has failed, it may be argued that, in a mixed economy,
public sector investment and employment policy can be used to steer
the economy in the direction desired by government. Public sector
policy affects the private sector too, though indirectly, because of
competition.
In reality, the criticism that project appraisal mis-specifies the
nature of government’s decision-making process, and that appraisal
depends on value-judgements is an argument about the justification
for planning. In a laissez-faire economy public investment projects
would be ruled out ex hypothesi. In the real world virtually all LDC
governments have rejected laissez-faire in favour of national
economic planning, backed by public sector investment projects.
Whilst the decision-making rules used in project planning models
might be politically naive, they do at least attempt to rationalise the
decision-making process in pursuit of government plans for
economic and social development. In the final analysis it is for
government to decide whether to accept or reject the project analyst’s
advice on how it should reach its decisions. As to the legitimacy of
value-judgements, because of intangible costs and benefits, no
method of project selection can be entirely objective. Project analysis
merely attempts to systematise and improve the consistency of value-
judgements.
Finally, in response to the criticism that project appraisal neglects
dynamic factors affecting economic growth and development, it may
be claimed that a good project plan anticipates the possibility of
change by subjecting the results of appraisal to sensitivity analysis, by
avoiding unnecessary plan rigidities and by provision for regular
project monitoring and ex post evaluation.
It appears to us that the balance of the argument is in favour of
systematised project planning rather than leaving public investment
plans to the ad hoc decisions of government. At present, few LDCs
have had sufficient experience of project appraisal and
evaluation - especially in agriculture - to have advanced beyond a
rather low level of performance. However, with better basic data
and greater skill born of experience, their proficiency should
improve. Moreover, the discipline which project planning imposes
on its practitioners is a valuable means of gaining a better insight
into how the economy works.
362 AGRICULTURE AND ECONOMIC DEVELOPMENT

References

Bandyopadhyaya, K. (1976), Agricultural Development in China and

India, Allied Publishers, New Delhi.
Benjamin, M.P. (1981), Investment Projects in Agriculture, Longman,
Harlow.
Carruthers, I.D. and Clayton, E.S. (1977), ‘Ex-post evaluation of
agricultural projects - its implications for planning’, J. Agric. Econ.,
XXVIII (3).
Dervis, K., de Melo, J. and Robinson, S. (1982), General Equilibrium
Models for Development Policy, Cambridge University Press,
Cambridge.
Dey, J. (1982), ‘Development planning in The Gambia: the gap
between planners’ and farmers’ perceptions, expectations and objec¬
tives’, World Development 10 (5).
Eckstein, O. (1965), Water Resource Development: the Economics of
Project Evaluation, Harvard, University Press, Cambridge, Mass.
Ellman, M. (1979), Socialist Planning, Cambridge University Press,
Cambridge.
Ghatak, S. (1978), Development Economics, Longman, London and
New York.
Heal, G. (1973), The Theory of Economic Planning, North-Holland,
Amsterdam.
Healey, D.T., ‘Development Policy: New Thinking about an Inter¬
pretation’, J. Econ. Lit., September, 1972.
Hirschman, A.O. (1968), Development Projects Observed, Brookings
Institution, Washington, DC.
Hopper, W.D. (1968), ‘Investment in agriculture: the essentials for
pay-offs’, in Strategy for Conquest of Hunger, Proceedings Rocke¬
feller Foundation, New York.
Howe, C. (1978), China’s Economy, Cambridge University Press,
Cambridge.
Johnston, B. with Page, J.M. and Warr, P. (1972), ‘Criteria for the
design of agricultural development strategies’, Food Research Studies
in Agricultural Economics, Trade and Development, vol. XI
pp.27-58.
Joshi, V. (1972), ‘The rationale and relevance of the Little-Mirrlees
Criterion’, Bui. Oxford Inst. Econ. & Stats., 34 (1).
Little, I.M.D. and Mirrlees, J.A. (1968), Manual of Industrial Project
Analysis in Developing Countries, vol. II, OECD Development
Centre, Paris.
Mishan, E.J. (1972), Elements of Cost-Benefit Analysis, Allen &
Unwin, London.
Owen, W.F. (1971), Two Rural Sectors: their Characteristics and
 PLANNING AGRICULTURAL DEVELOPMENT 363

Roles in the Development Process, International Development Re¬

search Center, Occasional Paper 1, Indiana University, Bloomington.
Padoul, G. (1975), ‘China 1974: problems not models’, New Left
Review, 89, January-February.
Paine, S. (1976), ‘Balanced development: Maoist conception and
Chinese practice’, World Development, vol. 4, no. 4.
Panne, van de C. (1976), Linear Programming and Related Techni¬
ques, 2nd edn, North-Holland, Amsterdam.
Pearce, D.W. and Nash C.A. (1981), The Social Appraisal of Projects,
Macmillan, London.
Sinha, R.P. (1975), ‘Chinese agriculture: a quantitative look’, J. Dev.
Studies, 11, 3, pp. 202-23.
Squire, L.A. and van der Tak, H.G. (1975), Economic Analysis of
Projects, Johns Hopkins University Press, Baltimore, for the World
Bank.
Stewart, F. (1978), ‘Social cost-benefit analysis in practice: some
reflections in the light of case studies using Little-Mirrlees techni¬
ques’, World Development, 6 (2).
Stewart, F. and Streeten, p. (1972), Little-Mirrlees methods and
project appraisal’, Bull. Oxford Inst. Econ. c£ Stats., 34 (1).
Timmer, C.P. (1974), ‘Choice of technique in rice milling in Java’,
Research and Training Network, The Agricultural Development
Council, September, pp. 1-10.
Symposium on the Little-Mirrlees manual of industrial project
analysis in developing countries (1972), Bull. Oxford Inst, of Econ.
Stat., February.
UNIDO (1972), Guidelines for Project Evaluation, United Nations,
New York.

Notes
1. A substantial number of purely theoretical limitations of comprehensive
methods of project appraisal has been discussed in the literature: see,
for example, the ‘Symposium on the Little-Mirrlees manual of industrial
project analysis in developing countries’, Bull. Oxford Inst, of Econ.
Stat., February 1972. However, these general criticisms lie outside the
purview of this book since none of them relates specifically to agricultural
project appraisal.
2. Apart from ‘random shocks’ such as the natural disasters (e.g. earthquakes,
droughts and floods) and sudden shifts in the external terms of trade or
access to overseas markets, two of the major reasons for divergencies
between ex ante expectations and ex post performance are (a) externalities
such as linkage effects (see, for example, Stewart and Streeten, 1972;
and Joshi, 1972); and (b) large investment projects can cause actual
prices and costs to vary substantially from their expected values.
12 Agricultural
Development:
an Overview

The dominant feature of the LDCs compared with the rest of the
world is their relative poverty, the amelioration of which is a central
objective of economic development. For reasons discussed at greater
length in chapter 2, poverty is also characteristic of traditional
agriculture, the predominant form of agriculture in most LDCs.
Inadequate access to land and capital for the majority of farmers,
combined with technological backwardness or even stagnation, are
two of the principal reasons for agricultural poverty. It follows that in
order to ameliorate agricultural poverty through raising farm output
and income, some relaxation of resource constraints is needed
combined with a higher rate of technological innovation. However,
due to the risks and uncertainties of agriculture, particularly in
physically underdeveloped economies in tropical environments, poor
peasant farmers understandably tend to be risk-averse. Hence policy
measures to reduce the risks and uncertainties of agricultural
production and marketing are also important for accelerating
agricultural development.
Development theory postulates a progression from the type of
economy where primary production, including agriculture, is domi¬
nant, through a stage where the majority of GNP and gainful
employment derives from secondary industry (i.e. manufacturing) to
a final stage in which tertiary or service industries predominate.
Moreover, it is virtually axiomatic that, to sustain economic develop¬
ment in the long run, most LDCs must undertake sectoral
diversification to reduce their dependence on agriculture. But, as
discussed at length in chapter 3, the feasibility of sectoral diversific-

364
 AGRICULTURAL DEVELOPMENT: AN OVERVIEW 365

ation is usually critically dependent on a growing marketable surplus

of agricultural products. However, due to the influence of labour of
surplus development models, the importance of the agricultural
sector as a source of redundant labour waiting to be mobilised for
‘costless’ industrial development has been exaggerated. Furthermore,
due to the combination in many LDCs of continuing rapid popul¬
ation growth, with severe constraints on the growth of employment
outside agriculture, the onset of a decline in the absolute size of the
agricultural sector labour force is likely to be delayed until a relatively
late stage of development. In the interim, the creation of additional
employment within agriculture - by encouraging farmers to adopt
appropriately labour-intensive methods of increasing output, for
example - is a policy priority.
Agricultural rent and agricultural production surplus are closely
analagous concepts. The amount of rent (or surplus) available either
for investment in agriculture itself, or to be taxed away to fund non-
agricultural investment depends upon agricultural productivity,
consumption within the agricultural sector, and possibly also on the
land tenure system. For the reasons discussed in chapter 4, share
tenancy is not necessarily inefficient. But in LDCs, the market in
agricultural land is frequently flawed by unequal bargaining power
between landowners and tenants, as well as by other market
imperfections. Although dual economy models have been useful in
helping to conceptualise the meaning of agricultural surplus, failure
to integrate the theory of capital accumulation with the theory of
choice of technique helps to explain their limited predictive power. A
further shortcoming of such models is that they neglect analysis of the
forces generating productivity gains and development within the
agricultural sector, as discussed in chapter 5.
In chapter 6 we refer to the substantial body of literature which
exists on the efficiency of peasant agriculture. One view is that,
despite their poverty, peasant farmers are allocatively efficient in the
neo-classical mode. An opposing view is that, because of poverty,
risk-minimisation takes precedence over profit-maximisation.
Moreover, despite the fruits of learning from experience, gaps in the
spread of knowledge can result in substantial technical inefficiency.
Those adhering to the ‘poor but efficient’ viewpoint conclude that
technological advance is virtually the sole means of raising output
and incomes in peasant agriculture. But opponents of this view
believe that significant increases in output can be achieved through
the more efficient use of existing technology. This may be feasible by
providing farmers with better safeguards against extreme fluctu-
366 AGRICULTURE AND ECONOMIC DEVELOPMENT

ations in prices and yields and by improving both the quantity and the
quality of agricultural extension services. We find the evidence in
favour of holding the view that peasant producers are allocatively
efficient reasonably convincing, but subject to expected economic
returns being discounted in some degree to allow for risk and
uncertainty. But the evidence supporting the view that peasant
agriculture is also technically efficient is much less convincing. We
consequently support the view that important though technological
advance may be in hastening agricultural development, this is not the
only means of improving agricultural output and incomes: progress
can also be achieved by improving production incentives, as well as
by encouraging farmers to make better use of present techniques.
Despite the foregoing caveat, more research is needed to hasten
socially beneficial agricultural technical change in LDCs. Due to the
unfavourable labour/land ratio in many countries, research pro¬
grammes need to emphasise biological innovations to promote land-
augmenting technical change. Governments have a vital role to play
in establishing and funding national research institutions and in
ensuring that the objectives of research programmes are not biased in
favour of private vested interests. The scope for conducting research
at a purely national level in poor countries is greatly limited by
financial and other constraints such as skilled manpower. One means
of surmounting this hurdle is international co-operation in promot¬
ing and funding agricultural research to benefit LDCs. Such research
has been notably successful in recent years, as exemplified by the
dramatic yield increases obtained from new varieties of wheat and
rice developed by CIMMYT and IRRI. Substantial scope exists for
extending international co-operation in agricultural research to other
crops and livestock products, as well as to regions of the world that
have not yet been significantly affected by the Green Revolution.
The role of machinery investment, particularly field mechani¬
sation, in agricultural development is controversial. On the one hand,
extra machine-power can help to raise productivity and reduce the
drudgery of hand labour for workers remaining in agriculture. On the
other hand, forms of mechanisation which are potentially labour-
displacing threaten to aggravate unemployment. However, because
of inter-sectoral income and employment linkages, the displacement
of labour from agriculture does not necessarily imply an overall
reduction of employment, particularly in countries with a domestic
farm machinery manufacturing capacity. Serious doubt exists con¬
cerning the feasibility of ‘selective mechanisation’ as a practical
means of avoiding labour displacement by tractors and other farm
 AGRICULTURAL DEVELOPMENT: AN OVERVIEW 367

machinery. Not only would it be difficult to identify machinery types

that are invariably labour-displacing (or land-augmenting), it would
be equally difficult to control the use of any particular machinery
type. But governments can influence the rate of farm mechanisation
unselectively, for better or for worse, by controlling the distribution
of agricultural machinery and the terms on which it is acquired by
farmers. Due to the predominance of imported farm machinery in
LDCs, the government’s import policy is critically important. In the
past, governments have probably tended to encourage rather than
discourage farm machinery imports, but without necessarily re¬
cognising and evaluating the full economic and social consequences.
The major conclusion on farm mechanisation policy in LDCs is that
although the substitution of capital for labour in agriculture is
inevitable in the long of run, the rate at which mechanisation occurs is
for public decision. Governments may choose to encourage or
discourage mechanisation unselectively by applying specific taxes or
subsidies at appropriate rates. Numerous non-specific government
policies, such as the exchange rate, also influence the rate of farm
mechanisation.
As discussed in chapter 7, there is little or no convincing evidence
of genuine perverse supply response in LDC agriculture, despite the
plausibility of the supposition that as their income rises peasant
farmers might choose to consume more leisure as well as extra
material goods. But, as expected, cash crops usually exhibit a more
pronounced supply response than do food crops grown primarily for
subsistence consumption. Moreover, due to farmers’ crop substi¬
tution opportunities, it is easier to raise the output of individual crops
by raising their price than to induce an overall increase in agricul¬
tural output by improving the agricultural sector’s terms of trade.
Non-price variables such as resource endowments and the choice of
production techniques are equally, if not more, important than prices
in determining levels of aggregate agricultural output in LDCs.
Gross inequality in the distribution of land gives rise to the demand
for land reform which is often regarded as a precondition of
significant agricultural development. But, for reasons discussed
in chapter 8, mere redistribution of landownership is insufficient
to transform agriculture. To achieve such a transformation, a
comprehensive ‘agrarian reform’ is needed which, as well as land
redistribution, also includes a full array of extension, credit and
marketing services. The demise of the large private landlord some¬
times leaves a vacuum in the provision of some or all of these services,
and new institutions, such as service co-operatives, are needed to fill
368 AGRICULTURE AND ECONOMIC DEVELOPMENT

the gap. Very small-scale farmers and landless labourers usually

derive few direct benefits from land reform. Other measures are
needed to improve their livelihood and incomes, including better
non-agricultural employment opportunities. More generally, in some
LDCs the population density is so high that it is quite unrealistic to
see land reform as a panacea' for abolishing rural poverty and
unemployment, simply because insufficient land is available to
provide every rural household with an adequate-sized plot. Again, it
is necessary to look to the general economy, rather than to agriculture
per se, for a resolution of the problem. Alternatives to land reform,
such as co-operative farming and tenancy reform, may help to raise
agricultural output and incomes, but fail to assuage ‘land hunger’.
More countries have enacted land reform legislation than have
actually implemented successful land reforms. Thus strong govern¬
ment direction and leadership are needed not only to enact the
legislation but also to carry it out.
The root of the ‘credit problem’ in LDC agriculture is farm
tenants’ lack of direct access to organised money and commodity
markets which, in any case, tend to be rather poorly developed in the
countries concerned. Credit policy reforms include land reform (to
improve the credit standing of small farmers), expansion of insti¬
tutional credit agencies (to compete with private moneylenders) and
the raising of deposit interest rates (to encourage more saving). The
role of credit in agricultural development is especially important in
facilitating the adoption of more appropriate technology to raise
productivity. This also includes the substitution of capital for labour,
especially in the long of run, with the more rapid growth of non-
agricultural employment opportunities. There is also much scope for
linking agricultural credit and marketing programmes to mobilise a
larger agricultural surplus.
In non-subsistence economies, agricultural marketing services link
food consumers with agricultural producers. The development and
improvement of these services therefore form an important compo¬
nent of overall agricultural development. Some reforms in agricul¬
tural marketing, such as infrastructural development and the provi¬
sion of better market information, are pre-eminently a government
responsibility. But since some marketing functions tend to be
performed more efficiently in the private sector, a good case exists for
collaboration between the government and private enterprise in
improving the range and quality of marketing services.
A fully convincing theoretical explanation of the relationship
between population growth and food supplies in LDCs has yet to be
 AGRICULTURAL DEVELOPMENT: AN OVERVIEW 369

developed. The classical or Malthusian model has been found

wanting in several respects, not least by the historical experience of
what are now termed the developed countries. But, as discussed in
chapter 9, contra-Malthusian theories, like Ester Boserup’s, suffer
from the same defect of assuming that the relationship between food
and population is unidirectional: a better and more comprehensive
theory might need to take account of feedback effects. Theoretical
reasoning backed by substantial empirical evidence point to the
conclusion that hunger is primarily a function of poverty rather than
a consequence of inadequate aggregate food supplies. The distri¬
bution of food supplies in LDCs roughly coincides with the
distribution of income. Although increasing aggregate food produc¬
tion and curbing excessive population growth may help to ameliorate
malnutrition in the long term, income redistribution and special
programmes to raise per capita food consumption within target
groups appear to be the only viable short-term instruments for
relieving hunger in the Third World.
Few economies are entirely self-sufficient and external trade can
play an important role in LDC development, both through the
working of the principle of comparative advantage and because of the
facilities afforded by trade for entering export markets and importing
capital and scarce technical skills. In many LDCs agricultural export
earnings form a major proportion of the country’s total export
earnings. Moreover, agricultural export prices and earnings are
prone to substantial fluctuation due to a combination of factors
affecting both supply and demand. The problem of agricultural
export instability is exacerbated by illiberal trade policies in devel¬
oped countries with respect to agricultural imports from LDCs.
Commodity price stabilisation and the removal of trade barriers both
require international action. Experience with international commod¬
ity agreements suggests that, like producer cartels generally, ICAs are
inherently unstable, so that any temporary success in raising or even
stabilishing prices is likely to be shortlived. Schemes of international
compensatory finance to ‘top up’ abnormally low primary product
export revenues, though not beyond criticism, may offer a more
promising solution to commodity instability. These issues are
discussed at greater length in chapter 10.
To plan or not to plan? Despite the association of macro-economic
planning with socialism, most present-day LDCs, regardless of
political ideology, make development plans, including sectoral plans
for the development of agriculture. The debate about whether
governments ought to undertake planning, rather than leaving the
370 AGRICULTURE AND ECONOMIC DEVELOPMENT

course of economic growth and development to the interplay of free

market forces, therefore seems somewhat arid. Economists are better
employed in helping governments to plan more successfully. Na¬
tional economic planning originated in the Soviet Union between the
two world wars and, in the context of agricultural planning, it is
instructive to contrast the ‘Soviet tribute model’, as imposed in
Russian agriculture during the 1930s, with the more recent ‘Chinese
model’. We make this comparison in the first part of chapter 11.
Apart from setting aggregate targets which are consistent with the
overall plan frame, and formulating matching policies of implement¬
ation, the core of planning for the development of agriculture in
LDCs consists of project appraisal and project selection. This is the
subject of the second part of chapter 11. A vast and complex
literature on project appraisal in LDCs has accumulated in recent
years. The objective of project appraisal is to measure total net social
benefit. However, a particularly contentious issue concerns the
project analyst’s ability to measure divergencies between private and
social benefits or costs. Even though casual observation may confirm
that because of numerous market imperfections in LDCs, prices of
both factors and products are commonly distorted, the planning
benefits of having reliable estimates of the underlying ‘shadow prices’
may fail to match their costs of preparation in terms of scarce
planning expertise. The debate on this issue continues. Due to the
nature of biological processes, as well as for other reasons, agricul¬
tural projects are notoriously difficult to appraise with confidence.
This has led some authorities to advocate the use of simpler and less
complex techniques of appraisal for agricultural projects. In practice,
the choice of method for agricultural projects is likely to be influenced
by numerous considerations, including the amount and quality of
project planning expertise at the disposal of the government. The
precise method of appraisal seems less important than that the
selection of competing projects should be based upon the application
of some reasonably uniform, consistent and economically rational
procedure. As might be expected, the record of project planning in
LDCs consists of mixture of successes and failures. The ex post
evaluation of projects has an important contribution to make to
improved planning performance in the future.
For the foreseeable future, the economic and social welfare of
much of the world’s population inhabiting the developing countries
will continue to depend upon progress in agriculture. For that reason,
social equity demands the apportionment of adequate resources of
finance and skilled manpower to LDC agriculture. Ensuring that
 AGRICULTURAL DEVELOPMENT: AN OVERVIEW 371

adequate resources are allotted to agriculture also serves the interests

of food consumers everywhere, assuming that adequate security of
supply is sought. However, realism compels us to recognise that
agriculture is only one of many claimants on such limited national or
international funds as are available for development. Moreover, in
the long run, the relative importance of agriculture is bound to
diminish in virtually all countries. Thus, at some stage of develop¬
ment, it is equally inevitable that agriculture must yield capital and
manpower to the industrial sector. But finding the optimum alloca¬
tion of investment between industry and agriculture during a
specific planning period is one of the most difficult policy and
planning problems confronting LDC governments. Present models
of economic development are regretably of little assistance in solving
this problem.
The transformation of LDC agriculture requires brain-power as
well as adequate capital and manpower. LDC agricultural policy
should therefore include educational measures to improve
agriculture’s generally poor public image and social status in order to
attract larger numbers of the most able and ambitious people into
entering careers associated with agriculture. The ‘dignity of farming’
needs to be restored not only for this reason, but also to raise the self-
respect of countless numbers of people in LDCs who have no effective
choice of career except on the land.
Index

activity analysis, 327-9 landless labourers, 24, 224, 270,

agrarian reform 368
defined, 220 opportunity cost of, 8-9
agri-business sector, 37 productivity, 8-10, 17-18, 33,
agricultural capital 36, 147, 253-6
market imperfections, 16, 20 surplus, 50, 329, 365
—co-operation, 141, 242-3, 367 transfer to industry, 50 et seq.
—credit turning point in growth, 49-50
sources of, 16 —land
transactions costs of supplying, distribution of, 18-21, 221, 367
16 effect of technological change on
policy, 368 price of, 146
—development, financial policies market imperfections, 20, 365
for, 235-7 productivity of, 17, 139, 222,
—exports 253-6
and foreign exchange systems of land use, 257-8
contribution, 67-8 use intensity and labour
—goods productivity, 258-60
trade in, 279 —marketing
—import substitution, 68 government participation in,
—inputs 247-9
supply bottlenecks, 16-17 infrastructural deficiency in, 240
—information and extension reform of, in LDCs, 241 et seq.
barriers to spread of, 12-17 —markets
reasons for concentration in barriers to spread of
government provision of, information in, 22
13-14 imperfections of, in LDCs,
supply costs, 14 239-41
—labour narrowness in LDCs, 22
diminishing returns to, 18, 263 —mechanisation
displacement of, 12 policy in LDCs, 166 et seq., 367

372
 INDEX 373

role in agricultural development, 40-1, 68-9

366-7 benefits of, 41, 66-9
—prices —yields
policy, 245 effects of weather, 21
stabilisation of, 244-7 problems of forecasting, 21
—production surplus, 32, 45, 221, agriculture
254, 365, 368 biological basis of, 21
—productivity competitive structure of, 22
and rent, 365 factor market distortions, 18-21
total factor, 139 poverty in, 13, 17, 364
—project planning, 329 et seq. productivity of, 36, 43
—projects resources structure of, 31
alternatives to UNIDO appraisal semi-subsistence, 6
method, 357-8 subsistence, 6, 42, 145
peculiarities of, 356 systems of, and relative factor
—research and information policy, security, 7
164-6 traditional, 4-18, 252
—sector agro-industries
declining relative importance in relative importance in LDCs, 37
the economy, 30-1, 37, 43 appropriate technology, 44, 165
share of GDP growth, 27-32
terms of trade, 239, 245, 260, balanced growth, 108
367 Bangladesh
—self-sufficiency, 296-9 agricultural unemployment in,
—surplus, 82-4 60-2
—technology benefit-cost rates, 330-1
appropriate/inappropriate, 11-12 Berry and Soligo model, 55-7
as production shift variable in bi-modal strategy, 318-22
contra-Malthusian model, 257 Boserup, E.,
backwardness of, in traditional contra-Malthusian model, 147,
agriculture, 10-11, 364 256 et seq., 275-8
barriers to adoption, 11-17 buffer stocks and buffer funds,
choice of, 10, 327-9 245-7
international transfer of, 11,
154, 261, 366 capital, 227
new, adoption of, 14-17, 40, scarcity of, 228
145, 165 social opportunity cost of, 332-3
new, generation of, 147 et seq., captial contribution
257, 264 equity of, 43-4
stagnation in, 10 means of transfer, 44-8
—trade cartels, 299-300
and, comparative advantage, CGIAR, 154
374 INDEX

Chakrabarti, S.K., 34-5 synthetic, 333

China use in project appraisal, 352-4
guaranteed employment in, 273 disguised unemployment, 98
population control in, 263 and shadow wage rate, 336
taxation of agriculture, 48 defined, 52
use of surplus agricultural empirical evidence for, 58-63
manpower, 63 Nurkse’s model, 52
Chinese model, 315-17 Sen’s model, 52 et seq.
Choice of technique, 327-9, 365, dual economy models, 97, 105,
367 119-21
CIMMYT, 154, 366 assumptions of, 97-8
Cobb-Douglas production Fei-Ranis model, 105-12
function, 119 Jorgenson model, 112-14
cobweb model, 174-77 Kelley, Williamson, Cheetham
stability of the equilibrium, 177 model, 114-19
stable, 177 Lewis model, 97-105
unstable, 177 remarks on, 119-21
collective farming, 142, 218 dualism
commodity trade economic, 5, 12, 42, 44, 365
concentration in, 280 labour market, 8-10, 17
cartels in, 299-300
Commonwealth, 300
comparative advantage ecological disequilibrium, 262-3
static versus dynamic, 68-9 economic growth
compensating financing schemes, agriculture’s contributions to, 26
306-7 et seq.
competition and poverty, 271
in agricultural markets, 240-1 economic location theory
contra-Malthusian model of and classical model of economic
economic development, 253, development, 255-6
256 et seq., 369 economic rent, 75-6
co-operative farming, 271, 225, 368 origin of, 76
cost benefit analysis, 330, 335 Edel, M., 35
costs efficiency, in agriculture
of adopting new agricultural and farm size, 136 et seq.
technology, 15-16, 261 and land reform, 222
credit concepts of, defined, 123-4
price of, in agriculture, 16 norms of, 136
Cummings, 197 Engel’s Law, 6, 31
European Economic Community,
discount rate 306
choice of, 331, 332-3 exchange entitlements
 INDEX 375

failure, as cause of famine, 266 literacy of, 12

et seq. weak bargaining position of, 17,
expected price, 183-5 240-1
Cagan’s model of, 183-4 feudalism
Nerlove’s model of, 183-4 characteristics of, 18
export instability and economic defined 25nl
growth, 289 financial policies, 235-7
financial surplus, 75, 82, 85
factor contributions food
capital, 43-8 income elasticity of demand,
labour, 49-66 31-2
factor prices maldistribution of supplies,
distortion of, in LDC 266-9
agriculture, 138 marketable surplus of, 218, 220,
family farm, 5 221, 222
family formation per capita consumption, 266,
economic motivations for, in 272, 274
peasant agriculture, 270 price elasticity of demand, 33-4
family labour redistribution of supplies, 271-5
characteristics of, 7-8 stamps, 274
family size foreign exchange contribution,
correlation with farm size, 5 66-9
famine fragmentation of farms, 5
cause of, 267-8
‘boom’ and ‘slump’, 268 geographic concentration in trade,
farm size 280
and farm income, 18 Green revolution
and land productivity, 138-40, and differential efficiency of
222 large and small farms, 140
and land-use intensity, 18 distribution of benefits, 146-7
policy, 140 growth models, 97
farm unions
and farm size, 18
improvement of, as project Helleiner, G.K.
planning objective, 342 land surplus economy model,
level of, compared with non¬ 65-6
farm sector, 13, 17 Hirschman, A.O., 38-9
methods of raising, 12 human capital, 43, 63-4
farmers, in traditional agriculture human fertility
access to information, 12 and economic growth, 269-70
as price-takers, 22 high rate in low-income
economic goals of, 131, 259-60 agriculture, 18
376 INDEX

rural-urban differential in, Bardhan’s theory, 20

269-70 intermediate technology, 44, 165
HYVs, 43 Intermediate Technology Group,
12
income internal rate of return (IRR),
redistribution as planning 330-1
objective, 341-3 International Coffee Agreement,
income-sharing 303
on family farms, 8 International Grains Agreement,
inflation 303
and food supplies, 33-6 International Monetary Fund, 306
imports International Sugar Agreement,
of food, 32, 67-8, 261, 264 303
India International Tea Agreement, 303
designed agricultural International Wheat Agreement,
unemployment in, 58-60 303
distribution of cultivation rights inter-sectoral terms of trade
in, 19 manipulation of, to extract
effect of agricultural technical agricultural surplus, 46
change on employment, 160-4 investment criteria
‘fair-price’ shops in, 274 choice of, 330-1
famine in Bengal, 268 IRRI, 154, 366
farming efficiency in, 125-6 Israel
landless rural households in, 19 land settlement in, 225
taxation of agriculture, 47
input-output analysis Krishna, R., 161-4, 191-3
use in estimating indirect Kuznets curve hypothesis, 273
employment effects of Kuznets, Simon, 26-8, 43, 64
technological change, 162
use in estimating inter-industry labour-displacing technical change
linkage coefficients, 39 defined, 158
institutional change labour reward system
theory of induced, 152-4 in present agriculture, 7-8
institutionalists, 330 labour surplus, 75, 82, 85
interest rate, 231-4 land-augmenting technical change
inter-industry linkage defined, 158
defined, 37-8 land reform
employment linkages, 39, 65 and ‘agrarian reform’, 218-20,
income generation linkages, 39, 367
41-2 and distribution of benefits of
production linkages, 37-9 technological progress, 146
inter-locking factor markets, benefits and costs of, 221-4, 226
 INDEX 377

effect on production and marketable surplus, 40, 44, 46,

productivity, 139, 223 342, 365
effect on size of agricultural marketed surplus, 185, 189-200
surplus, 45 estimation of, 191
limitations of, 224 marketing boards
meaning of, 218-20 fiscal function of, 244
motivations for, 18-19, 220-1, functions and limitations of,
367 242-3
redistributive effects, 20 marketing co-operatives
land settlement, 225 functions and limitations of,
land tenure system, 45, 217-18, 242-3
365 meta-production function, 148 et
leisure and agricultural output, seq.
109-11 model
leisure Chinese, 315-17
marginal utility of, in Fei-Ranis, 105-12
agriculture, 10 Jorgenson,112-14
leisure preference Lewis, 97-105
of peasant farmers, 56-8, 258-9, moneylenders, 328, 329
261
loanable funds, 327 national parameters
Lewis, W.A., use in project appraisal, 347-9
model of economic development, net present value (NPV), 330-1,
54 352, 354
New International Economic
malnutrition Order, 306
ameliorisation of, 269 et seq. Nurkse, R., 52
and distribution of food nutritional education, 274
supplies, 266-7, 369
definition, 265 oppurtunity costs
link with low productivity and of domestic food supplies, 32
agricultural poverty, 17, 63, of food imports, 32
369 ordinary lease, 87
measurement of, 266 owner-occupier, 86
Malthus, theory of population
and demographic features of Parikh, 197
traditional agriculture, 18 peasant farmer
Malthusian model of economic objective function of, 6, 15
development, 253 et seq., 369 planning techniques, 322-9
marginal propensity to consume activity analysis of, 327-9
in agriculture, 42 ‘cattle feed’ cost minimisation,
market contribution, 39-43 325-7
378 INDEX

input-output analysis, 322-4 protein-energy malnutrition

linear programming, 324-5 (PEM), 272
population density
and land reform, 368 quasi-rent, 77-8
and land use intensity, 257 et
seq. Ranis-Fei model, 57
population growth regional income multiplier
and food supplies, 252 et seq., application in project appraisal,
368-9 352
control of, 262, 269-70 rent, 76-82
cultural restraints on, 262, 264, and agricultural productivity,
high rate in traditional 365
agriculture, 18, 252, 365 and inter-sectoral capital
‘poor and efficient’ hypothesis, 125 transfers, 45
et seq., 365 and land reform, 221-2
Prebisch’s hypothesis, 294-6 and location theory, 255-6
price stabilisation, 286-9 and security of land, 7
price stabilisation schemes distortion caused by market
buffer stocks and buffer funds, imperfections, 20
245-7 influence of technological
limitations in reducing change, 146
agricultural price uncertainty, rental market, 79-82
23 resource flows
primary commodity prices table of, used in project
fluctuations in, 281-2 appraisal, 346-7
product contribution, 27-39 resources structure
profit maximisation of peasant agriculture, 7-8
as farming goal, 6 risk and uncertainty
project appraisal and adoption of new
and policy reform, 359-61 technology, 11, 14-16
and project evaluation, 358 and project appraisal, 333-5
criticisms of comprehensive and resource allocation
methods of, 356-8 efficiency, 366
defined, 330 nature and causes of, in
Little-Mirrlees (LM) method, agriculture, 21-3
340, 356 respecting prices, 22
objectives of, 330, 370 respecting yields, 21
UNIDO method of, 345 et seq. risk aversion, 6, 17, 22, 131-2,
project cycle, 344 264, 364
project evaluation, 358-9, 370 Ricardian corn rent, 78-9
property rights rural interest rate determination,
in land, 7 231-5
 INDEX 379

rural money markets, 228 subsidies

functions of 288-31 on food prices, 36
rural-urban migration supply response, 172-98
theory of, 8 labour market, 185
policy, 70 normal, 179
Russia perverse, 179
taxation of agriculture, 47-8 surplus labour, 98-9, 105, 112
switching values (of project
appraisal parameters), 354
savings gap, 332, 336, 361
selective mechanisation taxation
defined, 166 of farmers and landowners, 46-7
feasibility as policy instrument, land tax, 47
167-8, 366-7 technological change in agriculture
semi-feudal agriculture, 329 biological and mechanical, 151
characteristics of, 329 et seq.
Sen, A.K., distribution of effects, 144-7
model of disguised effects on employment, direct
unemployment, 52 et seq. and indirect, 160 et seq.
sensitivity analysis, 329, 334-5, 361 embodied and disembodied, 144
shadow pricing factor bias in, 155 et seq.
costs of, 356, 370 land-augmenting and labour-
of foreign exchange, 339-40, 350 displacing, 145, 158 et seq.,
of investment, 337-8, 350 366, 367
of wages, 338-9, 350 nature of, 142 et seq.
principles of, 337 theory of induced, 148 et seq.
share tenancy, 86, 226, 365 technological stagnation
economic case against, 89-90 consequences of, 17-18
Marshallian theory of, 89, 91-4 reasons for 11-17
theory of, 88 tenancy reform, 226, 368
Soviet planning, 311 time preference
‘tribute’ model, 312-15 social rate of, 332-3
STABEX, 306-7 Todaro’s rural-urban migration
stabilisation model, 8, 64
of commodity prices, 369 trade
of food prices, 36 commodity concentration in, 280
starvation features of, in agricultural
risks of, in traditional goods, 280-2
agriculture, 21 geographic concentration in, 281
strategy in agricultural goods, 279
bi-modal, 318-22 policies in developed countries,
uni-modal, 318-22 282-6
380 INDEX

traditional agriculture wages

definition, 10 and abundance of labour, 7
subsistence, 8
UNCTAD, 301-6 weather
UNIDO method of project effects on agricultural yields,
appraisal 21
appraisal procedure, 349-52 working hours
data requirements, 346 customary, in peasant
presentation and interpretation agriculture, 52-3, 258-9
of results, 352-5 work sharing
urbanisation effect, 31 on family farms, 8
 cr c
University of Waterloo

Z3 D
<’ <’
(D (D
“1
If)
in

cr^- a
CT
-i O 3 a 3
O) —ft
3
£ Q)
O) i-r
t~r fD
0) -i
O O
O O