Front Cover
 Title Page
 Table of Contents
 Demographic characteristics of...
 Economic and social characteristics...
 Population growth
 Population in rural planning: Variables,...
 Agricultural labor force, technology,...
 Migration and agricultural...
 Population and food planning
 The family and the farm
 Planning and the rural society
 Population and decentralized agricultural...
 Back Cover

Group Title: FAO economic and social development paper ; 51
Title: Population, society, and agricultural planning
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00085340/00001
 Material Information
Title: Population, society, and agricultural planning
Series Title: FAO economic and social development paper
Uniform Title: Planification agricole, population et société
Physical Description: vii, 163 p. : ill. ; 30 cm.
Language: English
Creator: Marcoux, Alain
Food and Agriculture Organization of the United Nations
Publisher: Food and Agriculture Organization of the United Nations
Place of Publication: Rome
Publication Date: 1987
Subject: Agriculture and state   ( lcsh )
Demography   ( lcsh )
Population   ( lcsh )
Agriculture -- Social aspects   ( lcsh )
Genre: international intergovernmental publication   ( marcgt )
bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Bibliography: p. 152-163.
Statement of Responsibility: by Alain Marcoux.
General Note: Translation of: Planification agricole, population et société.
 Record Information
Bibliographic ID: UF00085340
Volume ID: VID00001
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 17217032
lccn - 88167406
isbn - 9251022259

Table of Contents
    Front Cover
        Front Cover
    Title Page
        Page i
        Page ii
        Page iii
        Page iv
    Table of Contents
        Page v
        Page vi
        Page vii
        Page 1
        Page 2
    Demographic characteristics of a population
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
    Economic and social characteristics of a population
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
    Population growth
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
    Population in rural planning: Variables, estimates, and projections
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
    Agricultural labor force, technology, and employment
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
    Migration and agricultural development
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
    Population and food planning
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
    The family and the farm
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
    Planning and the rural society
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
    Population and decentralized agricultural planning
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
        Page 159
        Page 160
        Page 161
        Page 162
        Page 163
    Back Cover
        Page 164
Full Text



t, NA tA ~ ~

i "Wow-,

Population, society

and agricultural


Alain Marcoux
FAO Policy Analysis Division



Rom- 1987

ISBN 92-5-102225-9

All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted in any form or by any means,
electronic, mechanical, photocopying or otherwise, without the prior
permission of the copyright owner. Applications for such permission,
with a statement of the purpose and extent of the reproduction, should
be addressed to the Director, Publications Division, Food and Agriculture
Organization of the United Nations, Via delle Terme di Caracalla, 00100
Rome, Italy.

FAO 1987

The designations employed and the presentation
of material in this publication do not imply the
expression of any opinion whatsoever on the
part of the Food and Agriculture Organization
of the United Nations concerning the legal
status of any country, territory, city or area or
of its authorities, or concerning the delimitation
of its frontiers or boundaries.

- iii -


Over the last ten years, the Policy Analysis Division of FAO has
carried out various research and training projects covering socio-
demographic aspects of agricultural and rural development. On the one
hand, studies were undertaken, either by FAO itself or by external
specialists, on the interaction of population dynamics and the
development of the agricultural sector. On the other hand,
sensitization seminars and training workshops were organized for
agricultural and rural planners in developing countries. The latter
programme benefited from the preparation of case studies on the
relations between population and agriculture at the level of countries
and small regions.

These activities enabled us to gauge the interest shown by
planners in areas which had been often overlooked in agricultural
development studies and planning. A concrete indication of this
interest was the success of the first manual devoted by FAO to
"Demography for Agricultural Planners", prepared by the late Ken
Baldwin. The said manual, published in 1975, despite a re-edition, was
out of print. It had to be replaced by an up-to-date work which would
draw on new training materials and on the research carried out in other
centres or organizations in the meantime.

This is the purpose of the present document which sets out, for
agricultural development professionals, the demographic concepts
relevant for their work, deals with population dynamics in the
present-day world and with their links with socioeconomic phenomena,
especially in the rural context, examines possible uses of demographic
variables in agricultural planning, particularly in policy studies, and
goes into questions of substance and methods inherent to the
requirements of a type of planning focusing on rural society and its
needs, rather than on economic aggregates. The author, Alain Marcoux,
is a former official of this Division. This English version of the
document incorporates numerous factual and technical updates of the
original French version.

The financial assistance of the United Nations Fund for
Population Activities to the activities which have permitted the
preparation of this work is gratefully acknowledged.

Howard Hjort
Policy Analysis Division





1.2.1 Urban and rural population 4
1.2.2 Population density 6




2.1.1 Definitions 17
2.1.2 Activity rates 18
2.1.3 Employment, unemployment and underemployment 19
2.1.4 World patterns of labour force growth 23




3.1.1 Nature and measurement of fertility 27
3.1.2 Factors affecting fertility 31
3.1.3 The fertility of agricultural households 34




3.5.1 Fertility 44
3.5.2 Mortality 44
3.5.3 Population growth 46
3.5.4 Age structure 48
3.5.5 The growth of urban and rural populations 49

- vi -



4.1.1 The use of the variables
4.1.2 Population projections Calculation of survivors Calculation of births and of surviving children

4.2.1 The use of the variables Projection of the total labour force Estimating the proportion of agricultural
labour force































- vii -



7.1.1 Energy requirements 93
7.1.2 Protein requirements 94

7.2.1 Calculation of energy requirements 95
7.2.2 Calculation of protein requirements 97


7.4.1 Population, income and demand for food 103
7.4.2 Urbanization and demand for food 104






9.1.1 The technical aspect 121
9.1.2 The economic aspect 122
9.1.3 The institutional aspect 123
9.1.4 The political aspect 123

9.2.1 Social groups and economic units 126
9.2.2 The analysis of rural socio-economic systems 128
9.2.3 Social indicators 128

9.3.1 Poverty and inequality in development theories 131
9.3.2 Identification of the population concerned 132





Demographic variables influence development variables and are,
in turn, influenced by them. This was recognized by all participating
countries at the World Population Conference of Bucharest in 1974, and
again in Mexico City in 1984, as stated in the World Population Plan of
Action. But in practice the logical consequence i.e., that
demographic variables ought to be carefully taken into account in
development programmes has certainly not affected planning to the
desirable extent.

In this respect, agriculture is no better off than the other
sectors of the economy despite the importance of the human factor in
agricultural production: the analysis of what is called population
variables does not in most cases cover more than the labour force. This
means that there is scope for broadening and enriching the methods of
agricultural planning from this point of view. Plans, programmes and
projects often omit to take account of the impact of population
dynamics, and, even more frequently, overlook the impact of the actions
planned on socio-demographic variables.

The reasons for this state of affairs are of three kinds. First
of all, the practical importance of the relations between population and
economic variables is not always realized in planning units and other
government agencies. Secondly, the lack of demographic data is often
invoked to justify the neglect of population variables in preparatory
analysis and in plan formulation. Lastly, the lack of planning methods
in which demographic variables are formally integrated is frequently
noted as a matter for regret.

Planning methods, from the preparation of multiannual plans to
the formulation of agricultural projects, could be adapted without
serious difficulties in such a way as to incorporate the relevant
demographic variables. There is no need, it should be added, to
formalize this kind of integration beforehand, and the planners
themselves are in the best position to carry it out as soon as they feel
the need for it.

This is perhaps the decisive factor, and it explains why we
propose to concentrate on showing that there is really such a need.
This need stems from a few simple facts, but which have numerous and
varied consequences, of which we shall provide two examples. At the
level of investment projects, it must be realized that, for the whole
duration of an agricultural project say, a plantation (25-30 years) or
an irrigation scheme (50 or more years), the population concerned will
undergo considerable changes, both as regards its numbers and its
structures, and hence its needs and capacities, and these changes must
be planned for [71] 1/. When it comes to sectoral plans, in order to
prepare the forecasts of nutritional requirements, food demand, agricul-

1/ The numbers in square brackets refer to the Bibliography, pp. 148-159.

-2 -

tural production, employment requirements of the rural population and
migrations between sectors and to the towns, it is essential to be equipped
with a whole series of demographic parameters.

Of course, despite the importance of population data in
agricultural and rural planning, planners do not need to acquire a mastery
of complex demographic issues. However, they should acquire "a sufficient
knowledge of demographic terms and concepts, their meanings and
limitations, in the same way that all competent economists have a working
knowledge of statistics, without necessarily being professional
statisticians" [5]. In this way, they will be in a position to form an
opinion about the implications in their field of action of the trends
described by demographers, and to enter into a dialogue with them to
plan particular studies, for example on project areas, preferably as part
of multidisciplinary teams. In this way, they will be able to formulate
precise requests to demographers for information, calculations and studies,
taking account in the process both of the planning needs which they will be
fully able to identify and of the constraints inherent in demographic work.

Communication between members of difference disciplines working in
national development should be a matter of routine. As was observed by a
historian writing, as it happens, about demography, the social sciences
should serve in turn as auxiliaries of one another. Their overall
achievements would then doubtless be enhanced by the increased realism and
relevance of their approaches.



A human population is a group of individuals who share a common
characteristic such as:

(i) residence in a specified territory : the population of a
country, a district, a village;

(ii) the fact of belonging to a national or ethnic group : the
Sudanese population, the Yoruba population;

(iii) sex : the male population, the female population;

(iv) age : the population below age 10, the population over age
60, the population aged 25 to 29;

(v) the practice of a specified activity: the economically
active population, the pupils enrolled;

(vii) the fact of belonging to a household who lives off a
specified economic activity: the agricultural population.

In demographic usage, the term "population" normally refers to
the inhabitants of a given territory, although it may occasionally be
used for part of the inhabitants only, as in the above examples; such
groups are properly called sub-populations 1/.


The basic demographic characteristic of a population is its size,
i.e., the total number of the group. Population counts are a very old
concern of mankind, and the history of numerous civilizations affords
examples of censuses carried out with varying degrees of organization.
The census, or the most complete count of individuals, is still the
essential means of assessing the size of a population.

In census practice a distinction is made between the resident
population of a given area (the people who habitually live in that area)
and the actual population (the people present in the area on census
day). The latter method has the merit of simplicity, since it does not
make it necessary to determine (with the help of criteria which are
bound to be arbitrary) who is and is not a habitual resident, but its

1/ Almost all the definitions of demographic (or associated) concepts
in the first three chapters of the present work are taken from [101]
of the Bibliography. For further details on the analytical indices
and techniques, see [102].


results do not necessarily reflect the "normal situation" especially in
the case of small regions or localities which may be affected by
seasonal or occasional population movements. The former method seems
more fair, but it is less accurate and attempts to use it in order to
enumerate the nationals who have emigrated have generally failed.

The difficulty of carrying out an accurate census is illustrated
by the margins of error recorded, even in developed countries with sound
administrative structures and long experience of censuses. For
instance, several millions of persons are usually omitted by censuses in
the USA. In developing countries, a census which underestimates the
population size by 5 percent is an operation of acceptable quality. In
fact, there is no justification for expecting censuses, and demographic
data in general, to achieve a higher degree of precision than that
reached by other categories of statistics which planners are content
with: in most if not all countries, population estimates are more
accurate than the available estimates of the Gross Domestic Product.

Annex I supplies, inter alia, the size of the population for 148
countries. In 1984, seven countries had over 100 million inhabitants;
in total, they accounted for 57 percent of the world figure. 155 states
and 45 territories contained the remaining 43 percent. This extreme
inequality is the main characteristic of the distribution of the world
population. That population was estimated at 4,842 million in 1985. It
will reach 5 billion in 1987.


The spatial distribution of the population is an important factor
in the study of populations from the economic point of view, since the
location of communities is linked to that of exploitable natural
resources and of working and subsistence opportunities.

One way of studying spatial population distribution is obviously
to examine how the population is spread over the various regions or
administrative units of a country. It is obviously useful to have data
on this distribution and on the characteristics of the population in
each administrative unit when it comes to formulating and managing
development programmes. As this point does not raise any conceptual
problems, we shall not elaborate. It is of greater interest to examine
the implications of two instruments commonly used to assess the spatial
distribution of a population, namely the urban-rural distinction and
population density.

1.2.1 Urban and rural population

The large differences which exist between urban and rural areas
as regards economic and social organization as well as development
problems have made the distinction between urban and rural areas a
universally adopted criterion to characterize populations. The
relevance of this criterion for policy analysis and for the design of
development programmes is undoubted, but the ways in which it is used
have given rise to discussion on various counts.

- 5-

It should first be recalled that the urban-rural distinction is
designed to describe specified areas, and indirectly the population,
depending on whether the place of residence belongs to the urban or to
the rural areas. Secondly, there often is no specific definition of the
rural milieu: one merely takes what is not urban to be rural. As to
the definitions of the urban milieu used at the country level, there are
striking differences between them [103].

The most frequently employed objective criterion is the number of
inhabitants of localities. The threshold adopted in order to decide
whether a locality falls into the urban category is often that of 2,000
inhabitants, but the figure varies considerably, ranging from as low as
200 in some developed countries to 20,000 in some developing ones. A
number of countries identify the urban environment as the total of
localities which possess communal status. As the criteria used by the
administration to assign such a status vary and are often loosely
defined, it is difficult to compare the data prepared on this basis.
Another group of countries does not go beyond the provision of a list of
localities (sometimes reduced to the capital alone) in defining the
urban milieu. This category is akin to the previous one, since the
criteria of selection are not known.

The most fully elaborated formal definitions of the urban
environment use combinations of two or three of the following criteria:
the size of population, the administrative status, the presence of
specified infrastructures and services, non-agricultural economic
activities, and population density. Logically speaking, what is the
value of these criteria? The original problem is to define the urban
and rural areas, or, to put it simply, the town (urbs) an the country
(rus). From our point of view, the essential difference between the
two is the relation of the human habitat to the natural environment. In
a rural environment, the human habitat occupies only a small part of the
space, which is dominated by crops and natural ecosystems such as
forests and pasture land. In an urban environment, the human habitat is
definitely predominant, and leaves only a small part of the space to
vegetation. The conclusion is that the most appropriate objective
criterion for the identification of urban and rural areas actually is
population density.

However, that criterion by itself is insufficient, for it is
difficult to describe a small agglomeration as a town since its size
makes it impossible to distinguish it as a "landscape", in the
geographical sense from its surroundings. It would therefore appear
necessary to combine a minimum density with a minimum size of
population, or, even better, with a minimum area.

We have seen that certain countries grant the urban status to an
agglomeration only if the proportion of its inhabitants living off
agriculture does not exceed a certain percentage. This criterion is
questionable: in fact, one does not see why a town characterized by
its size, density, type of habitat and physical environment would not
be inhabited to a substantial extent by farmers, as in effect is the
case in some regions. It is also illogical to want to use a population


characteristic to characterize space, when the problem to be considered
is just the opposite (characterize population according to the type of
environment it lives in). The use of this criterion denotes a confusion
between the concepts of "rural" and "agricultural", and produces
deformed images of reality.

The classification of localities and their population into urban
and rural has often been accused of being too rough. If we consider the
size and population density of localities, we inevitably note that there
exists a very wide range of situations, which cannot be accounted for by
a simple classification into only two groups. Attempts have been made
to deal with this problem by adding a third category ("semi-urban") to
the classification, or even a fourth one. It is true that such
solutions approximate to the description of what has been called the
"urban-rural continuum" [82]. But at bottom it appears that we must
distinguish between two problems.

When dealing with development problems at the country level, it
is necessary to use a more detailed characterization of space. One will
want to consider for example suburban areas, or dispersed settlement
areas, or middle-sized towns, and so on. At this level, the urban-rural
distinction is therefore insufficient. In dealing with rural areas, it
particularly interesting to differentiate between similar areas by
taking account of their distance to the nearest urban centre [82]: this
is a factor which influences not only marketing (and hence the structure
of agricultural production), but also inter-sectoral trade in general,
ways of life, migration, etc. The levels of development achieved in
rural areas seem to be deeply influenced by the proximity of markets, of
agents of socio-economic change and of socio-political influences which
urban centres represent [64].

When the aim is to establish internationally comparable data, the
current practice is unsatisfactory since we know that the words rural
and urban designate different realities in almost every country. For
example, the United Nations [103] report that, in 1980, the proportion
of the urban population was the same (36 percent) in Gabon as in Ghana.
What conclusion can one derive from this figure on the compared degrees
of urbanization in these countries, when Ghana classifies as urban its
localities with over 5,000 inhabitants whereas the figure for Gabon is
2,000 ? It would help if countries were invited to publish data on the
distribution of their localities by size of population along a standard
scale (for example less than 500 inhabitants; 500 to 1,999; 2,000 to
4,999; etc.), in order to make comparisons meaningful at least among
countries of similar levels of development. Of course, this would not
solve the problem entirely, and it would be difficult to move toward a
standardization of the other criteria of urbanization.

1.2.2 Population density

The intensity of settlement is generally measured by population
density,- i.e., the quotient of the total population by the area of the
territory in which the population lives. This index is usually
expressed as the number of persons per square kilometre or square mile.

-7 -

Various comparative density indices are sometimes computed to take into
account other factors besides surface area. Mention should be made in
particular of:

the density of population per unit of cultivable area;

the density of the agricultural population per unit of
cultivable area 1/.

These indices are sometimes based on the cultivated area instead of the
cultivable area.

Instead of expressing density as the number of persons per unit
of area, one can use the inverse formula, which gives the average area
available per person. This formula, which expresses the same reality,
is probably more meaningful to anyone interested not in the passive
occupation of space, but in its utilization (in particular by
agriculture ). Thus, Annex II shows the ratio of total area and of
cultivable area to total population, for 146 countries, expressed in
hectares per person.

The figures of Annex II reveal wide differences between

certain countries have less than a quarter of hectare per
inhabitant: Bangladesh, Barbados, Malta, Mauritius,

others have over 50 hectares per inhabitant: Australia,
Bostwana, Libya, Mauritania, Mongolia, Namibia.

This index does not in itself indicate any degree of economic
advantage or disadvantage. To realize that this is so, we need merely
point out the presence in each of these two groups of countries, both
those with a low and a high density, of "rich" and "poor" countries as
measured by per caput GDP. The amount of cultivable land per inhabitant
is more significant in this regard.

The factor linking the two indices is the proportion of arable
land; in terms of density, the relation would obviously be written as

Population Population Cultivable area
-----------= --------------- x ---------------
Total area Cultivable area Total area

1/ The agricultural population includes all persons depending for
their livelihood on agriculture, and includes everyone whose main
economic activity is agriculture, together with all their
non-working dependents [71].

- 8 -

Let us examine the variations from country to country in the
ratio of cultivable area to total area. For the world as a whole, the
proportion of arable land is 10.3 percent. At the country level, there
is a huge diversity:

the proportion may be over 30 percent, as is the case in 22
countries, including: 12 European countries (Bulgaria,
Czechoslovakia, Denmark, the Democratic Republic of Germany,
the Federal Republic of Germany, France, Hungary, Italy,
Poland, Portugal, Romania and Spain); 5 Asian countries
(Bangladesh, Cyprus, India, Thailand and Turkey); 3 African
countries (Burundi, the Comoros and Nigeria), and Mauritius;
and one American country (Barbados);

at the other extreme, the proportion may fall below 3
percent, as is the case in 33 countries including: 16
African countries (Algeria, Angola, Botswana, Chad, Congo,
Djibouti, Egypt, Gabon, Liberia, Libya, Mali, Mauritania,
Namibia, Niger, Somalia and Zaire) 1/; 8 Asian countries
(Bhutan, the Democratic Republic of Yemen, Kuwait, Mongolia,
Oman, Saudi Arabia, Singapore, and the United Arab
Emirates); 6 American countries (the Bahamas, Belize,
Bolivia, Guyana, Peru and Suriname); 2 Oceanian countries
(New Zealand and Papua-New Guinea); and one European country

The countries most richly endowed with arable land per inhabitant
(over 0.80 ha per inhabitant) are:

5 countries in the temperate zones with a large area and
either a relatively generous proportion of arable land (the
United States), or a low population density (Argentina,
Australia, Canada and the USSR);

and 5 African countries, either sparsely populated
(Botswana, Central African Republic and Zambia) or with a
relatively high proportion of arable land (Cameroon and

The less richly endowed countries (with less than 0.05 ha per
inhabitant) are:

4 small, largely desert countries: Djibouti, Kuwait, Oman
and the United Arab Emirates;

and 6 island countries whose land is either largely
uncultivable or highly urbanized: the Bahamas, Iceland,
Japan, Malta, Papua-New Guinea and Singapore.

1/ It can be seen that Africa is not well endowed in this respect.
For the region as a whole, the proportion of arable land is only
5.5 percent (and 10 percent if the countries bordering the Sahara
are excluded).


However, even the density of population to arable land is only of
mediocre interest for a diagnosis of the agricultural situation and
prospects. For the productive potentials of equal land areas may be
widely different, depending on the nature of the soil, the climate and,
of course, the level of technology employed. In this perspective, the
density could be expressed as a relation between actual population and
the maximum potential density or population carrying capacity of the

An FAO project was aimed precisely at determining the productive
potential of the land in developing countries [99]. The study was
conducted by agro-ecological zone, taking into account the nature of the
soil and climatic variables (temperature, humidity). For each area, the
maximum potential production (corresponding to the range of crops best
suited to the agro-climatic conditions) was worked out under three
technological hypotheses regarding the level of inputs. Deductions were
made to take account of non-agricultural land uses, cash crop
cultivation and fallow. On the basis of the maximum production
(expressed in calories) for each area, the study assessed the population
which could be fed with this production, using standard calorie
requirements per caput. By relating this population-carrying capacity
to the actual population (in 1975) or to the projected figure (in 2000),
the study identified critical areas where the resources might be
insufficient to meet food requirements.

It is clear that the ratio of the maximum possible population
carrying capacity (under given technological assumptions) to the actual
population is a better indicator than the simple ratio of the quantity
of land (even if arable) to the population. The example given below
brings out the differences in the results which may be obtained by the
use of one or the other of these indices (Table 1).

Table 1- Ratio of arable land to population and of
population-carrying capacity to population
in 1975 for seven Asian countries

Country Ratio of arable Ratio of population-carrying
land to population capacity to population at
(ha per person) the levels of inputs:

Low I Medium High

Pakistan I 0.28 1.1 1.2 1 1.4
India 0.27 0.8 1.8 1 3.5
Laos 1 0.26 2.5 8.4 I 22.5

Bangladesh 0.12 0.5 1.3 1 2.1
Vietnam 0.12 0.9 2.7 1 4.9

Brunei 0.08 3.3 1 28.9 I 34.6
Bhutan 0.07 1.3 I 3.8 1 6.9

Sources: [84 85].

- 10 -

In order to establish the above table, we extracted from the
studies cited in reference the data for three groups of countries with
similar ratios of arable land to population. It will be seen that the
differences in population carrying capacity, despite the similar
densities, are far from negligible. Even in the low-inputs hypothesis,
there is a 100 or even 200 percent difference in carrying capacity
between countries with the same density. The effect of higher levels of
technology on land productivity leads to even greater differences, of as
much as 400 percent in the third group of countries, and of 1 500
percent in the first group.


A population is normally divided fairly equally between males and
females. The usual index employed to describe this distribution is the
sex ratio, i.e., the ratio of the number of males to the number of
females (usually expressed as an index value, e.g. the number of males
per 100 females). For instance in Benin, the 1979 census gave the
following results:

Males: 1,600,512
Females: 1,737,728

The sex ratio was therefore:

SR = --------- x 100 = 92.1

The sex distribution may diverge markedly from equality in two
kinds of situation:

(i) after a war which has caused serious military losses, and
hence a shortage of men;

(ii) when emigration or immigration occurs on a sufficient scale
and affects mainly the members of one or the other sex; in
most cases, these movements involve males, with the result
that there is an excess of males in the immigration areas
and an excess of females in the sending areas.

Besides migration, the factors affecting the sex distribution of
a population are, on the one hand, the sex ratio at birth, and on the
other, mortality differences between sexes. In every population, more
boys than girls are born: the sex ratio at birth usually is comprised
between 101 and 105. As regards mortality, we must distinguish between
two cases. In developed countries, mortality (even in peace time) is
higher for males than for females. Thus, despite the excess of boys at
birth, we find more females in the population of these countries (see
the following section, where the example of Switzerland illustrates this
point). In developing countries, an excess of males in the total
population has sometimes been registered for instance, in about a
third of the African countries and higher female mortality, either
observed (especially among young children and women of reproductive age)

- 11 -

or assumed, has often been advanced as an explanation for this fact.
However, there may also have been an under-enumeration of the female
population in censuses. There are grounds for believing that this is
the case in several countries which "export" male labour, but which have
a sex ratio of over 100 (see Table 2).

Table 2 The

range of the sex ratio in a few developing countries

Country Year Sex ratio

Bahrain 1981 140.3
Kuwait 1976 120.4
Libya 1984 115.5
Tunisia 1975 101.8
Chile 1983 98.1
Senegal 1976 96.8
Burundi 1983 94.0
Malawi 1977 93.0
Swaziland 1975 91.1
Botswana 1981 89.0

Source: [104]

In practice, the sex ratios at country level are greatly
influenced by the degree of migration, be it immigration (see Bahrain,
Kuwait and Libya) or emigration (see Botswana, Malawi, Swaziland). The
same is true within countries; for example, when rural-urban migration
is predominantly male, we find situations such as that of Burundi, where
the sex ratios in Cameroun, where the urban and rural sex ratios were
the following in 1983:

121.6 in the urban areas
and 92.2 in the rural areas.

The opposite case is found in Ethiopia or in the Philippines for
instance, and more widely in Latin America, for example in Chile, where
the sex ratios were the following in 1983:

94.6 in the urban areas
and 116.5 in the rural areas.


Age distribution, or "age structure is a key characteristic of
any population. Demographically speaking, the levels of mortality,
fertility and nuptiality are strongly influenced by age. From the

- 12 -

economic standpoint, the same holds for the nature and intensity of
economic activity. The distribution of population between economically
active persons and dependents is therefore largely a function of age

The age structure of a population is assessed by recording the
age (or, better still, the date of birth) of the persons covered by a
census or by a sample survey. These data are fairly accurate in
developed countries, but in countries with no or defective civil
registration, many people do not know their date of birth or age with
precision. In these circumstances, the results are distorted, and in
particular there is a tendency to round the age upwards or downwards
(especially to multiples of 10). There are methods for measuring the
probable extent of incorrect declarations and to smooth the age
distributions affected by such irregularities.

Table 3 -

Sex and age structures of the population of Ivory Coast (1975):
numbers and proportional distribution

Age Numbers (in '000s) Distribution (per 1,000)
Groups I

Males I Females Males | Females

0-4 625.1 622.9 93 93
5-9 542.7 514.0 81 77
10-14 372.9 319.8 56 48
15-19 309.5 330.5 46 49
20-24 316.6 297.2 47 44
25-29 295.7 297.0 44 44
30-34 229.6 216.8 34 32
35-39 201.3 180.6 30 27
40-44 168.1 128.0 25 19
45-49 133.0 98.5 20 15
50-54 94.1 74.3 14 11
55-59 72.1 46.9 11 7
60-64 51.9 40.5 8 6
65-69 28.6 21.4 4 3
70-74 1 21.4 18.6 I 3 3
75 & + 22.1 22.5 I 3 3

Total 3 484.8 3 229.3 519 481

6 714.0 1 1000
1 I I I

Source: [104]

- 13 -

The age structure is usually presented by five year age groups,
for each sex, as in Tables 3 and 4 which show two typical structures -
one for a developed country and one for a developing one. It will be
noted that the classification is based on complete years, i.e., on age
at last birthday. For instance, the "5-9 years" age group includes
those who had reached at least their fifth birthday but not their tenth

Let us offer a few brief comments on these sex and age
distributions. In the Ivory Coast, an excess of males is found in
almost all age groups. It is difficult, without making a detailed
study, to ascribe this either to mortality differences by sex, or to
female under-enumeration. In any case, the presence of a large and
mainly male immigrant population produces a particularly high sex ratio
between ages 35 and 65.

Table 4 Sex and age structure of the population of Switzerland (1979):
numbers and proportional distribution

Age Numbers (in '000s) Distribution (per 1000)

Males I Females Males I Females

0-4 184.3 175.4 29 28
5-9 218.8 208.0 35 33
10-14 253.9 242.4 40 38
15-19 251.2 243.4 40 39
20-24 237.7 237.8 38 38
25-29 240.3 239.8 38 38
30-34 254.8 247.2 40 30
S35-39 226.4 220.4 36 35
S40-44 201.1 198.9 32 32
45-49 188.5 194.3 30 31
50-54 172.4 187.2 27 30
55-59 160.6 175.5 26 28
60-64 132.1 148.9 21 24
65-69 127.1 158.9 20 25
70-74 100.4 141.8 16 23
75 & + 116.3 211.8 18 23

Total 3 065.9 I 3 231.7 486 I 514

S6 297.6 1 000
1 I I I

Source: [104]

- 14 -

In Switzerland, we find an excess of males in all the age groups
above 45. This is a normal situation (because of the sex ratio at
birth) up to the age of 15 to 20, but after that, higher male mortality
should produce an excess of females. This factor is masked in the
present case by the existence of a largely male immigrant population.
Thereafter, the further up we go in the age groups, the greater is the
excess of females resulting from higher male mortality.

As regards age structures, the tables bring out the differences
between the two main types of structure, i.e. between a "young" and an
"aging" population. In order to compare the two populations presented
here in this perspective, we must refer to the proportional distribution
(see the right side of the tables) which relates the figures to a common
base (1 000 persons). It will then be seen that the share of the
younger age groups is much greater in the Ivory Coast than in
Switzerland, and that the share of the older age groups is much greater
in Switzerland than in the Ivory Coast. These differences are readily
observable in the double histogram traditionally used to represent age
structures: the population pyramid (Graph 1).

What are the reasons for these differences in the shape of the
pyramids ? The layman often supposes that a broad-based and rapidly
tapering population pyramid, like those found in developing countries,
reflects a high mortality. It is not at all so. It is essentially a
high fertility which leads to this type of age structure. Conversely,
aging population structures in developed countries are caused not by low
mortality, but by the fertility decline. Mortality plays only a minor
role in the formation of age structures [105].

Within a country, age structures may obviously vary from one
region to another, as a result of the differences in the level of
fertility and as a result of migration. A typical example is the
difference between age structures in urban and rural areas. Table 5
gives two instances of this contrast.

Table 5 Age distribution of urban and rural populations:
Zambia (1974) and the Dominican Republic (1970)

Zambia Dominican Republic

Age Urban I Rural Age Urban I Rural
Groups (U) I (R) U/R Groups (U) I (R) U/R
0-14 517 523 0.99 0-14 443 497 0.89

15-49 448 372 1.20 15-49 467 408 1.14

50 & + 35 105 0.33 50 & + 90 95 0.95

Total 1000 1 1000 1.00 Total 1000 I 1000 1.00
Source: I I I I I I [104]
Source: [104]

- 15 -

Graph 1 Population pyramids of Ivory Coast (1975)
and Switzerland (1979)



I10 9 8 7 i 4 3 2 1
109 87654321001

2 3 4 5 6 7 8 9 10

Size of the age groups as percentages of the total population




I o I I I I o I I
4 3 2 1 00 1 2 3 4
Size of the age groups as

percentages of the total population




- 16 -

A frequent pattern of rural and urban age distribution is one with
a higher proportion of adults in the urban areas (because of the
presence of immigrants) and, as a corollary, a higher proportion of
children and elderly people in the rural areas. Such structures may
reflect different situations as regards migration; this is the case in
the two examples in Table 5. In Zambia, the difference between the
urban and the rural age structures is especially due to male migration
to the towns. In the Dominican Republic, the difference is caused by
female migration to towns. A simple way of showing this is to calculate
the urban and rural sex ratios, by broad age group (Table 6). In
Zambia, the sex ratios are particularly high from age 15 upwards in
urban areas. In the Dominican Republic, on the contrary, they are
especially low at these same ages in the urban areas.

One could also, of course, identify rural areas with a flow of
adult immigrants and hence with age structures similar to those in towns
as shown in Table 4. Thus, resettlement areas usually show highly
distorted age structures with few children, many adults and no elderly
persons. Because of this, their demographic evolution is rather
peculiar 1/.

Table 6 Sex ratios by broad age group, in the urban and rural
areas: Zambia (1974) and the Dominican Republic (1970)

Zambia Dominican Republic

Sex ratio Sex ratio
I I ___________I I _______________
IAge groups I Age groups
Urban Rural Urban Rural

0-14 98.9 101.1 0-14 97.0 104.3

15-49 108.8 76.1 15-49 85.0 106.2

50 & + 252.3 118.0 50 & + 85.2 129.1

Source: [104]

1/ See section 6.4.




The main economic characteristic in a population is activity,
whereby a distinction is generally made between the economically active
population, also referred to as the labour force, and the economically
inactive population.

2.1.1 Definitions

The economically active population comprises all persons of
either sex who supply the labour for the production of goods and
services. It therefore includes employers and wage-earners as well as
unpaid family workers. In the same way, it includes persons without a
job, but who are seeking it, as well as employed persons. It generally
excludes housewives, whose activity is not regarded as being of a
professional nature.

Statisticians and planners are well aware that the contours of
the labour force are blurred because of the difficulty of laying down
criteria for determining who is active and who is not, and because of
the ways in which these criteria are applied in statistical operations,
censuses and surveys. Let us illustrate the main problems arising in
this connection, especially in the agricultural sector.

In accordance with international recommendations [106], the
preconditions for inclusion of the persons enumerated in the
economically active population are:

(i) to be above a specified age;
(ii) to have a job or to be seeking work;
(iii) in the case of an unpaid family worker, to work at
least a third of the normal working time.

The minimum age is of necessity somewhat arbitrarily fixed. For
a long time it was put at 15 in most countries. But there has been a
growing trend in the developing countries to fix it at 10 in view of the
early age of entry in activity, especially in the agricultural sector of
these countries. In any case, this limit is more a statistical
convenience, setting limits to the work of the enumerator, than the
reflection of a clearcut situation.

It is not easy to apply the criterion of search for work for the
unemployed. Some people are not looking actively for work, but would be
willing to accept it if they felt that the situation of the labour
market was favourable. These persons are regarded as inactive according
to the usual criteria, but they do represent a potential economic
contribution. This may be the case in particular for many housewives.

- 18 -

The case of family workers, especially in agriculture, raises the
problem of women's work, as well as of children's work. The application
of criterion (iii) above to the agricultural sector may exclude from the
active population a large number of young people and women who divide
their working time between study or domestic work and work on the farm.
This is why FAO [86] recommends that the total work put in by each of
these two categories be evaluated and, if it turns out to be
significant, should be compared with the total work put in by full-time
agricultural workers grouped by age and sex. In this way, it would be
possible to have a more complete picture of activity in the agricultural
sector and to aggregate the data at the international level.

The recording of female activity raises not only the problem of
definitions but also that of their application in the field. Thus, in
Tunisia, during the 1956 census, all women belonging to agricultural
households and of working age were classified as economically active:
there were more than 300,000 of them. In the 1966 census, on the
contrary, women were classified as inactive, unless they were in paid
employment. The apparent number of active women fell, as a result, to
below 7,000. In the 1975 census, female activity was better covered
(and the result was a very substantial apparent increase in female
labour force which some observers later commented as if it was real),
but in fact it was again under-recorded, to the extent that a subsequent
multiround survey on employment, using concepts more appropriate to the
rural economy, estimated the female labour force at five times the
number which would have been arrived at using census definitions [26].
Examples illustrating the same point can easily be found in other
countries 1/. In this field, employment surveys usually give results
which are much more realistic and useful to planners than those of
censuses, provided that an adequate effort has been made to define
operational concepts.

Another problem is that of the length of the reference period by
which employment is determined. Agricultural work is irregular, varying
markedly in intensity from one month to the next, and is subject to the
vagaries of the weather. So, the agricultural statistician cannot be
interested in measuring work and unemployment over short periods. He is
rather concerned with quantifying the volume of work to be carried out
on the farm, the distribution of that work among sources of labour, its
relation to the size and type of the farm, etc. [86]. The desired
information should therefore cover the whole year, which implies
multiround statistical operations.

2.1.2 Activity rates

The size of the labour force depends not only on the age
distribution of the population, but also on social considerations such
as the extent of schooling, the usual age for retirement, the degree to
which women can engage in economic activity, etc. [71]. Owing to the
influence of these factors, the proportion of the economically active
population in the total population, or crude activity rate (CAR) varies
from one country to another. In particular, since male activity patterns

1/ See the following section.

- 19 -

are very similar from country to country, the level of the CAR is
heavily influenced by the extent of female participation in the labour
force, or on the way in which this participation is recorded by
statistics. For instance, the CAR in Burkina Faso was estimated at 23
percent in 1975 by the census, while the rate estimated in Burundi by
the 1980 employment survey was exactly double: 46 percent. The reason
for this amazing difference is simply that the female activity rate was
considerably underestimated in the former case: the published figure
was 2 percent, against 41 percent in the latter case.

The CAR, which varies little in the short run, may be useful for
certain planning purposes. If more detailed information is needed, the
rates have to be established by sex and age group, and the factors
causing their possible variations have to be studied. Table 7 shows
such rates for Cameroon. These are fairly representative of the levels
which may be found in developing countries. The activity rates of the
young people are relatively high because schooling ends at an early age.
Those of the elderly people are also distinctly higher than in developed
countries, because of the absence of institutional systems for the
social protection of this category. Between the ages of 25 and 60, it
is almost universal practice (at least for men) to engage in economic

The interpretation of such data must take account of the
definitions of economic activity adopted, and even of the methods of
data collection employed. Moreover, the simple classification of the
population between active and inactive persons does not permit to give
account of the varying intensity of the work or of the proportion of
time devoted to it (at least beyond the minimum proportion which is used
to define activity). When dealing with particular populations, for
example in areas affected by specific projects, it is necessary to
collect more precise information on these subjects.

The interpretation of female activity rates is bound to be
somewhat particular, for most women who are classified as housewives,
and therefore as inactive, are actually occupied in productive
activities in the home, even if the goods and services provided by them
are not included in the GDP. So, the apparent economic advantage
derived by a country from a high level of female economic participation
is "largely illusory if it results merely from women producing in paid
employment the same kinds of services and goods which, under a different
organization of the economy, they would produce without pay within the
home" [71].

2.1.3 Employment, unemployment and underemployment

As has already been pointed out, the active population includes
both employed and unemployed workers. It may also include a substantial
proportion of those workers who (because of a local or temporary
situation) only have a reduced activity and could be defined as
partially unemployed, or underemployed. Let us examine the concepts
used to identify and study these three categories of the active

- 20 -

Table 7 Active population and activity rates
Cameroon, 1976 census

by sex and age:

I Males Females

Age Total I Active population I Total I Active population
Group IPopulation ---------I-----------I Population I----------I----------
I I ('000) I Number I % ('000) Number I %
S10-14 423.2 96.7 22.8 379.2 77.8 20.5

15-19 335.3 153.5 45.8 352.6 124.9 35.4
20-24 252.8 204.8 81.0 297.7 132.2 44.4

25-29 222.4 206.2 92.7 272.9 129.4 47.4

30-34 189.0 180.4 95.5 232.9 115.9 49.8

35-39 192.6 185.2 96.1 223.9 124.9 55.8

40-44 162.2 155.3 95.8 179.8 101.8 56.7

S45-49 145.7 138.8 95.3 149.1 94.5 63.4

50-54 117.0 109.3 93.4 119.1 69.7 58.5

55-59 90.8 81.9 90.2 87.2 53.8 61.6

60-64 75.4 63.4 84.0 75.2 35.9 47.7

65-69 42.3 32.8 77.7 40.8 18.6 45.2

70-74 34.9 23.6 67.7 38.4 11.6 30.2

75 & + 51.2 24.2 47.2 52.2 11.0 21.0

STotal 1(2760.3) 1 1656.2 I 60.0 I (2914.5) I 1101.7 I 37.8
population I I I I I

Source: [107]

- 21 -

The standard employment statistics, such as those furnished by
censuses and surveys, distinguish between two categories of labour only:
the employed and the unemployed. The result is that the category of the
underemployed (yet so important from the standpoint of economic and
social policies) is not measured, let alone characterized. As for the
criteria used in order to identify the members of the active population
and to divide them into employed and unemployed, they vary from country
to country, and even over time in a given country.

There are substantial differences even in the definitions of the
active population [107]. Generally, data on the active population do
not include students, women exclusively engaged in household work,
retired persons, rentiers, and dependent persons. But there are
variations between countries as regards the classification of the armed
forces, pensioners of institutions, people seeking their first job,
seasonal workers and part-time workers. In certain countries, these
groups are at least in part included in the active population, whereas,
in others, they are regarded as inactive.

An important source of differences, especially in the case of
agriculture, is the treatment of unpaid family workers. The extent to
which these are included in the active population differs very markedly
between countries; in one case (Ivory Coast) they are totally excluded.
But it is the measurement of female economic participation which is most
strongly affected by practices in census and survey taking. In most
developing countries, a high number of women work, without being
remunerated, on the family farm, but there are substantial differences
between countries in the criteria used in order to determine the extent
to which this category of female workers should be included in the
active population. In the non-agricultural sectors, apprentices may be
formally included (as in Cameroon), but they are often treated as unpaid
family helpers, and hence often excluded from the active population.

Lastly, contrary to common practice, some countries exclude the
unemployed from the active population (Burkina Faso, Jamaica,

It should be added that the reference period used to determine the
individual status as regards economic activity is another source of
differences. In certain countries, the data refer to the actual
situation of each person on the day of the census or survey, or for a
short specified period, such as the week immediately preceding the date
of the data collection. In other countries, the data refer to the
habitual situation of each person.

As to unemployment, a standard definition was put forward by the
8th International Conference of Labour Statistics 1/. In practice,
however, national definitions of unemployment often ignore this
recommendation. One of the reasons for this is undoubtedly the fact

1/ Geneva, 1954. See [106-107].

- 22 -

that the standard definition is suited for contexts in which employment
is codified 1/, and therefore appears unsuited for the economies of the
developing countries, specially for the agricultural sector and even the
informal urban sector.

At the bottom, as has been pointed out by a number of economists
in studies of employment in developing countries, it is the very idea of
unemployment in the sense of the state of a person who is without a
job and seeking work which is unsuited for activities which are
informal or irregular by their very nature, such as agriculture. For
example, censuses habitually define as unemployed a person who has
worked for less than specified time during a specified period preceding
the operation (less than a week during the month preceding the census,
etc.). It is clear that, in this framework, "unemployment" in
agriculture will vary very widely depending on the season in which the
census is taken. In fact, the statistics will thus record not only the
few really unemployed persons and the most seriously underemployed ones,
but will also include other underemployed ones in the group of
"employed" persons. It is therefore necessary (and statisticians are
increasingly concerned by this problem) to identify measurement methods
for underemployment as such.

However, when dealing with the agricultural population, we must
be on our guard against the implications of the usual concepts. In
pre-industrial societies or in those moving toward industrialization, a
considerable part of the time which is not devoted to marketable
production is needed for social production and reproduction. If we
study the ways in which a peasant uses his time, "it will be quickly
realized that the activity which is termed productive is not the main
one. It is a serious mistake to consider that non-productive activity
is less important, less interesting and less dignified" [34]. We must
regard as one of the characteristics of the traditional way of life the
fact that a certain part of the time has to be employed in activities
linked (upstream or downstream) with production as such, particularly in
activities of social relations which are practically compulsory in this

The classical definitions of underemployment, therefore, do not
seem satisfactory, and there are grounds for envisaging a dualistic
approach to employment and underemployment: in market-oriented
societies, on the one hand, and in subsistence societies, on the other.
Certain methodological precautions would therefore be indicated in the
study of employment policies: identifying sub-populations which differ
from the standpoint of the patterns of participation in economic and
social activities; working out the patterns of time allocation for each
of these and assessing the social (by sex, age and status) and technical
division of labour; etc. [34].

1/ For instance, formal reference is made to the labour contract and to
other juridical characteristics of the status of the individual as
regards employment.

- 23 -

In practice, it is not always possible to apply such refinements.
However, it is highly desirable to concentrate on the division of labour
and on the way in which it can be affected by economic transformations
which are under way or in the offing. Let us consider an illustrative
case [51]. In a region of Tanzania, where it was not possible to detect
any underemployment in terms of time use, people were still able to
introduce tea growing into their agricultural system without on that
account neglecting food crops (or the compulsory social activities), and
all that despite the labour-demanding nature of the new crop. The
adjustments made in the utilization of labour in this case probably
included an increased degree of participation by children. This example
also reminds us that the time-budget approach ignores productivity
aspects: in our case, there was underemployment in the sense that a
more remunerative activity could be integrated into the agricultural
system. This consideration supports the idea that agricultural
underemployment can be identified only by measuring "under-income".
Although such an approach would not be free from arbitrary norms, one
could can at least avoid the pitfalls inherent to the interpretation of
data on the time spent by people at what statistics narrowly define as
economic activity.

2.1.4 World patterns of labour force growth

The labour force at the world level currently grows less rapidly
than the total population, so that the global CAR tends to decline: the
ILO has estimated this rate at 41 percent for 1975, against 44 percent
for 1950. This change is due to a fall in the male activity rates below
age 25 and over age 55, a fall which is not completely offset by the
increase in female activity rates. One consequence of this dual
evolution in activity rates by sex is of course an increase in the
proportion of women in the total active population: from 31 percent in
1950 to 35 percent in 1975.

It is estimated that, between 1950 and 1975, the activity rate of
males under 25 dropped from 63 percent to 49 percent and that the rate
for males over 55 fell from 75 percent to 61 percent. For men between
ages 25 and 55, the rate remained practically unchanged. Both in
developed and developing countries, there seems to be a rise in the
average age of entry into the labour force and a lowering of the average
age of retirement; these developments are related to the extension of
education and to the expansion of wage labour (in the secondary or
tertiary sector) and of wage regulations.

The CAR is, on an average, lower in the developing countries than
in the developed ones. The activity rates of younger and elderly men
are higher in developing countries because of the lesser extension of
schooling and of institutionalized retirement. As against this, women's
activity rates are (or seem) lower in developing countries. Moreover,
developing countries have an unfavourable age structure, with a smaller
proportion of adults than developed countries (see sections 1.4 and

- 24 -

Because of that structure, the active population has increased
much less rapidly than the total population since 1950, so that the
overall CAR has declined. The dependency ratio of developing countries,
measured by the number of inactive perons to every 100 active persons,
rose from 134 in 1950 to 153 in 1975. During the same period, this
ratio remained practically unchanged in the developed countries.

In view of the data reliability problems pointed out earlier, it
is difficult to assess the real trends in female activity in developing
countries. On the basis of the crude data, the differences between
countries seem to be vast: women form 50 percent or more of the total
labour force in certain countries of tropical Africa against less than 5
percent in certain countries of North Africa or the Middle East. It is
also difficult to assess the changes over time, since the methods and
quality of data collection are not stable.

The evolution of agricultural employment has a marked influence
on the employment opportunities for women when these form a relatively
small part of the non-agricultural labour force, which is the case in
most developing countries (with the exception of Latin America). Thus,
the fall in the share of agriculture in total employment (see section
5.2) contributed, other things being equal, to the shrinking of female
employment which has been observed in most Asian and African countries


The concept of agricultural population (see footnote, p. 7) can
be based either on the criterion of economic activity or on that of
income [86]. When the criterion of economic activity is adopted,
reference can be made either to the sector of activity or to the
profession. In the former case, the agricultural population is taken as
including all those economically active in the agricultural sector and
their dependents. In the second case, the term covers all those
exercising an agricultural profession and their dependents. It will be
seen that these two definitions should yield fairly similar results, for
it is rare that people meeting the requirements of one of these
definitions do not meet those of the other.

The only problems linked with the application of these
definitions arise over the reference period (the present activity should
be recorded, but also the habitual one), the identification of the main
activity for people with several activities (either on the basis of the
time spent or of the income obtained) and the identification of the
dependents of every active person.

When the criterion used is income, it is considered that the
agricultural population includes all those deriving their income mainly
from agriculture and their dependents. There are few cases of persons
belonging to the agricultural population as defined by income but not as
defined by activity (absentee landowners with no other activity). There
are also few instances of persons belonging to the agricultural

- 25 -

population as defined by activity but not as defined by income,
although, in a number of regions, non-agricultural activities and
emigrants' remittances are of increasing importance for rural

It is sometimes easier, seeing the problems posed by the
classification of individuals in relation to the agricultural
population, to use the concept of farm population [71 106], which
comprises the farmers and the members of cooperative, collective and
community farms, as well as the members of their families and unrelated
persons living on the farm, whether they work on the farm or not. But
this concept, which is useful and applicable especially in developed
countries, has given rise, even in these countries, to a number of
criticisms which bear on the difficulty of defining the farm and the
fact that quite a few farmers, as well as a large part of the
agriculture labour force, do not live on the farm 1/.


We propose to deal briefly here with certain characteristics which
are fairly commonly used to classify persons in presenting the results
of censuses or of other statistical operations regarding population.

The United Nations Demographic Yearbooks regard as social
characteristics such variables as: nationality; the ethnic group;
colour; the language spoken (or dialect); and religion.

The educational characteristics always receive special attention
in censuses and surveys. The population is often classified by its
degree of education. Persons above a certain age (often 10) who can
read and write constitute the "literate" population, as opposed to the
illiterate. Statistics based on the degree of education may include a
classification by length of study and diplomas obtained.

Educational characteristics are very important for planning,
especially in decentralized planning and in the case of programmes
involving the introduction of new practices and techniques, as the level
and type of education have an influence on the capacity for economic
change within a population. Hence the need to take these variables into
account in economic surveys (as also in demographic surveys, given the
influence of the level of education on demographic behaviour as well).

It is often necessary in a planning exercise, especially at a
decentralized level, to distinguish within a population different socio-
economic categories which carry out specific functions in the economic
and social system and, as such, are unequally or variously affected by
development programmes because their collective or individual strategies
are different. In terms of purely social characteristics, this would be
true, for example, of ethnic groups or castes.

1/ On the measurement and quantitative importance of the agricultural
population, see [83].

26 -

As regards what are called socio-economic characteristics,
reference should be made in particular to the status of persons in
relation to their economic activity, which in fact reflects their degree
of autonomy: employers, wager earners, independent workers etc. In
agricultural planning, the distinction is particularly important, for it
is obvious that landless labourers, smallholders (whether they
occasionally hire out their labour or not) and large-scale farmers (who
often resort to wage labour) do not have the same motivations, interests
and behaviors, and that neither the agricultural population nor the
agricultural labour force can be regarded as homogeneous groups.



Population growth is the change in the size of population over
time as a result of births, deaths and migration.


3.1.1 Nature and measurement of fertility

Demographic studies of fertility deal with certain phenomena
connected with human childbearing. The term fertility (or its
equivalent natality) refer to the frequency of births within

The crude birth rate (CBR) is the simplest fertility index. In
its usual, annual form, it is obtained by relating the annual number of
births 1/ to the average size of the population 2/ during the year:

Crude birth rate = Number of live births
Average population

The rate is generally expressed in births per 1,000 inhabitants.
Let us take an example. In Mauritius, in 1978 there were 24,234 live
births. In the same year, the average population was estimated at
894,471. We therefore have:

CBR = 24,234 = 0.0270

which can be expressed as "27.0 per 1,000" 3/.

1/ The rate is more exactly based on live births, meaning that
stillbirths are not included.

2/ This means the half-sum half of the population estimates at the
beginning and at the end of the year. The population at mid-year
can be used instead, if population growth during the year is fairly

3/ The expression "X per 1,000" is equivalent to "X divided by 1,000".
This is why it is incorrect to say that the crude birth rate is
equal to the ratio of the number of births to the size of the
population "multiplied by 1,000", as some authors do.

- 28 -

Annex I gives estimates of the average CBR during the 1980-1985
period, for 148 countries. It will be seen that the lowest birth rates
are of the order of 10 per 1,000, while the highest are of the order of
50 per 1,000. The lowest rates are found in Europe: the Federal
Republic of Germany, Sweden, Luxemburg and Switzerland have CBRs of
between 10 and 11 per 1,000. The highest are found in Africa:
Mauritania, Botswana, Malawi, Niger and Kenya have CBRs of between 50
and 55 per 1,000.

It is clear that developing countries have distinctly higher
birth rates than developed countries. The level of 20 per 1,000 is a
rough dividing line: only a few developing countries (Barbados, Cuba,
Uruguay, China, Cyprus and Singapore) have a lower birth rate, while
even fewer developed countries (Albania and Ireland) have a higher birth

In the range between 20 and 30 per 1,000, we can identify
countries in a transitional stage, i.e., in which fertility has begun to
decline and is on its way to developed countries' levels. This category
includes Mauritius, Costa Rica, Guyana, Jamaica, Panama, Trinidad and
Tobago, Lebanon, Sri Lanka, Thailand, and especially Cape Verde,
Argentina, Chile, the Republic of Korea and Israel.

The crude birth rate is a simple but imprecise instrument of
measurement. For instance, phenomena such as migration, which in the
short run influence population size (the denominator of the rate) but
not the number of births, may lead to variations in the crude birth rate
which are not indicative of fertility changes. Moreover, the
denominator of the CBR includes persons not directly concerned with
reproduction, children and the elderly for instance. This is why it is
better to use indices which relate the number of births observed during
a period in a group of women of childbearing age, to the size of that

The rates obtained by relating legitimate births to the number of
married women are called marital fertility rates; non-marital fertility
rates relate the number of illegitimate births to the number of single,
divorced and widowed women. I/ When no distinction is made as regards
the legitimacy of births or the women's marital status, one deals with
overall fertility rates.

The general fertility rate (GFR) relates the total number of
births to all women of reproductive age regardless of marital status.
Rates based on a narrower age range (usually, five-year age groups of
women) are called age-specific fertility rates (ASFR). Let us
illustrate the calculation of the general fertility rate with the example

1/ In some societies this sharp distinction is not consistent with
social practices and values. In such cases, one will want either to
study phenomena in terms of overall fertility, or to study the
fertility of women in the different types of union, defining the
appropriate indices to do so.

- 29 -

of Mauritius. The childbearing age
to 49; the number of women aged 10
288,277. Therefore:

GFR = 24,234

groups, in this case, extend from 10
to 49 in 1978 was estimated at

= 0.0841

or 84.1 per 1,000.

Table 8 shows how the age-specific fertility rates are
calculated, on another example.

Table 8 Age-specific fertility rates:

Thailand, 1983

Age Number of Number of Fertility
groups women births rate
= a = b b / a
(thousands) (thousands) (per 1,000)
1 I 1 I

15-19 I 2834.4 133.5 47.1
20-24 1 2385.1 342.5 143.6
25-29 ] 1994.2 275.2 138.0
30-34 1708.9 148.5 86.9
35-39 I 1390.6 73.7 53.0
1 40-49 I 1056.2 35.7 33.8
S 45-49 1 886.4 23.4 26.4

15-49 12255.8 1032.5 84.2

Source: [104]

The general fertility rate appears at the bottom of the table.
The pattern of the series of ASFRs, with a peak between ages 20 and 30,
follows the most common model. However, the rates in Table 8 ought not
to be regarded as reflecting the average situation in developing
countries; these rates are relatively low in relation to that average,
for the decline of fertility is relatively advanced in Thailand.

The GFR is also not an ideal index. Table 8 can help understand
why. It is seen, in effect, that the level of the GFR depends in part
on the age-specific fertility rates and in part on the distribution of
the female population between age groups. For example, if the
age-specific rates remain unchanged, but the proportion of women of
between 20 and 30 diminishes, the result will be a decline in the
general fertility rate, which will obviously not reflect a real trend in

- 30 -

This is why, on the basis of the age-specific fertility rates,
one calculates a synthetic index which is not affected by the age
distribution of the female population: the total fertility rate (TFR),
computed by the summation of the series of age-specific fertility rates.
Total fertility for a given period represents the number of children
that would be born per woman, in a group of women experiencing no
mortality and subject, during their reproductive years, to the
age-specific fertility rates observed for the said period. In the case
of Thailand, the total fertility rate for 1983 would be:

TFR = 5 x (0.00471 + 0.1436 + 0.1380 + ... + 0.0264) = 2.644

(since the rates used here measure the average annual fertility over
five years of reproductive life, each of them must be multiplied by five
to depict total fertility).

The reader should also be aware of the existence of an index of
the same nature as the total fertility rate, the gross reproduction rate
(GRR), which is derived by multiplying the TFR by the proportion of
female births. The GRR has often been used in the past, but the TFR is
preferred at present as the summary index of period fertility.

Some idea of the range of levels of fertility world-wide at the
present time will be given by the fact that total fertility is of the
order of 2 or less in most industrialized countries, and of the order of
8 in some developing countries. The lowest levels imply that the
population is barely able to reproduce itself, while the highest levels
imply (if account is taken of the prevailing mortality) an approximate
trebling of the population from one generation to another 1/.

Annex I gives estimates of total fertility in 148 countries for
the period 1980-85. For the reasons stated above, the hierarchy of the
levels of fertility measured by this index differs slightly from the one
yielded by birth rates. But, broadly speaking, the picture of fertility
world-wide as thus supplied is similar. The highest levels are found in
Africa, the only region in which some countries have a TFR of over 7,
and where no one country has a TFR of less than 2.6. The lowest levels
are encountered in Europe, where eleven countries have a TFR lower than
1.8, and where almost all countries are, from this point of view, below
"replacement level" 2/. On this scale, countries "in transition" are
characterized by GRRs of less than 2 and, where the transition is more
advanced, of less than 1.5.

I/ That is, in the average interval between generations of mothers and
their children (the average age of mothers at the birth of their
children) which is generally of between 25 and 30 years.

2/ A TFR of about 2.15 (or a GRR of 1.05) is needed in developed
countries in order to ensure the bare replacement of generations.
The amount in excess of GRR = 1 makes up for the losses due to the
mortality of mothers before the end of the childbearing period, so
that each woman is replaced by one daughter in these conditions.

- 31 -

3.1.2 Factors affecting fertility

Fertility is influenced by many kinds of factors, ranging from
biological to social and economic. This incidentally makes it difficult
to interpret its trends. In this connection, mention should be made of:

(a) Physiological factors. Fertility depends in the first place on
physiological reproductive capacity, that is, fecundity. The incidence
of sterility (whether male or female) obviously has negative
consequences on the actual level of reproduction, that is to say, on
fertility 1/.

(b) Nuptiality. This is one of the fertility determinants which are
linked to individual behaviour. The prevailing customs in a society as
regards marriage (or, more generally speaking, entry into union) affect
the level of fertility, especially through the age at marriage and the
forms of union.

Women who marry late (let us say after 25) tend to have fewer
children than those who marry early. Late marriages reduce the rates of
population increase because they cause, on the one hand, a shortening of
the reproductive period "at risk" of pregnancy, and, on the other hand,
a lengthening of the interval between generations. The postponement of
age at marriage has been an important factor in slowing down population
increase in industrialized countries since the 19th century, and in a
number of developing countries, especially in Asia, during the last
thirty years [109].

As regards the forms of union, it may be said (restricting our
description to the more common ones) that the highest fertility levels
have been found in monogamous populations practicing early marriage and
able to cover their vital needs to a reasonable extent. Polygamy is not
the most fertile regime, contrary to a popular belief which overlooks
the characteristics of this regime, and in particular the predominance
of older men on the "marriage market". In populations in which it has
been possible to effect a comparison, women in polygamous unions have
fewer children than those of equal age in monogamous unions. The
fertility of consensual unions, which are frequent in Latin America and
the Caribbean, is lower than the fertility of marriages in the same

(b) Birth control practices. Contraception and abortion obviously
have a direct impact on fertility, and their effect on population growth
is significant when these practices are sufficiently widespread. This
is the case in developed countries and in a considerable number of
developing ones, so that, at world level, the use of modern methods of
birth control has become an essential determinant of determining family

1/ Sterility can be very high when it is the consequence of
epidemiological factors generall diseases). This is the case in
large parts of Central Africa, the countries most seriously affected
by this problem being Gabon, Congo, Zaire, the Central African
Republic and Cameroon (over 17 percent of the women are childless at

- 32 -

size. All social groups and all individuals are not of course equally
inclined to use these methods, or indeed to limit the size of their
families. We can see here the interaction of various factors affecting
fertility: socio-economic factors, which we discuss below, rather have
an indirect effect, in the sense that they partly condition the couples'
attitudes as regards family size, and hence the age at marriage and the
use of birth control methods.

(d) Standard of living. The correlation between standards of living
and fertility is usually a negative one. The birth rate in poor
countries is considerably higher than in rich countries. Differences of
the same kind may be observed within countries between rich and poor
regions. At the household level, studies on the relation between income
and fertility often have not used clear enough definitions of income to
permit reaching precise conclusions. On the theoretical level, a
distinction is made between two effects of income level on the "demand"
for children:

an income effect, which tends to increase this demand (as it
does for all economic goods) when income rises;

a price effect, or substitution effect, which tends to
reduce this demand (more sharply) by raising the relative
costs of children (it costs more to maintain and educate
them when the expectations and requirements rise along with
the economic and social level of the parents); this
prompts families to "substitute" other goods for them.

When considering the negative relationship between income and
fertility, it is difficult to sort out what is ascribable to the income
factor and what is due to other factors related to income, such as
employment and the level of education.

(e) Employment. Here we shall cover only the employment of children
and of women, since available studies show that male employment has
little if any independent influence on fertility. When there are
opportunities for the children to assist the family in ensuring its
economic maintenance, parents desire a larger family. It has often been
contended that the number of children is also regarded as a factor
giving the parents some security for their old age 1/.

Both economists and sociologists have elaborated theories on the
relation between female employment and fertility. Economists point out
that when a woman works, she increases the household income. This tends
to increase fertility (see hereabove the "income effect"). However,

1/ This idea is supported by the results of various surveys carried out
in Asia (see [111], pp. 28-31). However, some authors have
questioned it in a number of cases [112].

- 33 -

this also tends to increase the opportunity cost of children, and this
may reduce fertility. The opportunity cost is obviously lower for women
whose wages are low, or who can have their children taken care of at
little or no expense; let us note that this is the case in most rural

According to the sociologists, women in a position to fulfill
various roles, including work outside the home, will be less inclined to
devote their whole attention to housework and to children, and their
fertility will be lower. In the case of women with "good" professional
positions (and hence more prestige, power and money), this effect will
be reinforced. Sociologists also describe an inverse trend in the case
of women whose work and family roles are compatible, as occurs when work
is on a family farm holding. In the latter case, fertility is not
negatively affected; the contrary may actually be true (see section

Empirical data generally fit in with these theories. They
suggest that, in developing countries, in view of the present structure
of the female labour market, women's participation in the active
population would doubtless be insufficient to bring down fertility.

(f) Education. A general conclusion of studies on fertility is that
there is an inverse relationship between fertility and the level of
education. This correlation is "generally greater than between
fertility and any other variable but education" [113]. According to
most of these studies, the influence of the woman's level of education
is more marked than that of her husband's. It has been suggested [114]
that the level of education could be considered as a variable which
reflects the costs of bringing up children (all the higher because the
parents wish to give their children a level of education at least as
good as their own) as well as the attitudes towards birth control and
knowledge of its methods. Generally speaking, it would seem that
primary education is a threshold: there are rather sharp differences in
fertility between uneducated women and those with primary education.
With secondary and higher education, the differences in fertility tend
to diminish. It should be noted that one of the ways in which education
affects fertility is the direct incidence on the age at marriage, which,
in turn, affects fertility (see above).

(g) Cultural factors. This is an extremely complex field. One
indication of the importance of such factors is that, in a country's
population, the different ethnic groups often have different levels of
fertility. For example, the fertility of Asian communities in Africa
declined during the fifties, whereas that of native populations remained
practically stable. Differences in the level of education or standard
of living can be invoked in this case a new example of factor
interaction. But, even in Asia, the areas where Chinese culture
predominates experienced more rapid declines than others. In this
field, any explanations must be largely speculative, but no doubt
anthropological research on the relations between social organizations
and demographic regimes will provide interesting insights in the future.

- 34 -

Religion is an important aspect of cultural systems. In
developed countries, the influence of religion on fertility is
determined in particular by the doctrines bearing on birth control.
Protestantism and Judaism have relatively liberal views on this matter.
The influence of these two confessions in developing countries is
limited: in those parts of the world, the dominant religions seem to
foster high fertility or rather, societies tend to interpret the
religious principles in this sense. It is, then, the degree of
attachment to those principles which indirectly influences fertility.

(h) The environment. It has often been noted that fertility is
higher in rural environments than in the urban ones 1/. But this is
because the factors cited above (standard of living, level of education,
age at marriage, etc.) differ sharply between the urban and the rural
populations rather than because of specific influences ascribable to the
urban and rural environments themselves. In this connection, the proper
approach is to wonder why the behaviour of urban and rural households
differs. In fact, it is interesting for rural specialists and for
planners to examine the behaviour of farming households in order to try
and understand how economic and demographic patterns of behaviour are
linked and how they influence each other in this context.

3.1.3 The fertility of agricultural households

Agricultural households are the basic economic and demographic
units of the agricultural sector within which decisions are taken:

on production (allocation of labour, land and other
factors) and on the choice of economic activities;

on consumption (allocation of production to subsistence
needs and to market sales; purchases of goods and

on saving and investment;

on demographic behaviour: nuptiality, fertility,

These decisions are interdependent. In particular, decisions on
demographic behaviour cannot really be analyzed out of the context of
the management of human resources as part of the household's economic
and social strategy. We shall have the opportunity of examining
migrations in this perspective. For the moment, let us look at
fertility, that is, at reproductive behaviour.

1/ Only a few examples of the opposite situation (surveys in Egypt,
Liberia and Zaire) are known. Sometimes, fertility is higher in the
urban areas because sexual interdictions are less strictly observed
there than in the villages [115].

- 35 -

The high level of fertility in most agricultural populations has
attracted the attention of analysts and given rise to long arguments,
especially because of the interest which has developed during the last
20 years or so for policies designed to slow down population growth in
developing countries. For the countries seeking to restrain this
growth, the numerical predominance of their rural and agricultural
population and the persistence of high fertility in this population,
obviously create problems, and the search for the causes of this high
fertility is a condition for designing policies to curb it.

Economic theories of fertility at the family level are based
either on the study of the costs and benefits brought about by children,
or on the study of the net intergenerational flow of resources which
they give rise to (the differences between the two approaches are
small). In this perspective, one may distinguish broadly between two
types of society. In one of these, children, in the course of their
life, provide their parents with more resources than they receive from
them. Families have therefore no interest in limiting their fertility,
which accordingly remains high. In the other type, the net resource
flow benefits the children. It is therefore to the families' advantage
to keep fertility low. When the cumulative resource flow changes
direction, a major incentive to have children disappears, and fertility
declines. This reversal of values, which would trigger the fertility
"transition", seems to be linked to the development of mass education,
to the shift of employment opportunities from family production to wage
labour, and to the cultural influence of the industrialized countries
which spreads their model of small family as well. The role of these
developments is in effect to equalize individual consumption within the
family and to weaken the individual's moral obligations to the
traditional family, the kinship system in the broad sense and the local
or tribal community [117].

As regards the fertility of agricultural populations, two
conflicting theories exist. For some authors, high fertility is not
rational for rural societies, and fertility would decline if families
had access to the knowledge and modern methods of fertility control.
Mueller, for example, contends that the the children's work contribution
is not substantial enough to prevent them from being an economic burden
on peasant societies. The productivity of that contribution is low, and
"increases in the productivity of agricultural investments may be
raising the opportunity cost of supporting children; at the same time,
land fragmentation and longer schooling lower the value of children as
productive agents" [116]. For some others, on the contrary, the observed
reproductive behaviour is a rational response to the logics of
traditional agricultural systems. In those systems, especially in those
incorporating fallow periods, the bulk of agricultural work relies on
women's and children's work. As women are responsible for food
production, they rely intensively on the help of their children, whence
a strong motivation to have many of them.

- 36 -

It thus appears that the wife has no less valid reasons to desire
a large family than has the male farmer who, in view of periods of peak
demand for agricultural work, is clearly better off with a large number
of children [118].

On the other hand, when growing land scarcity brings about the
stage of intensive agriculture, this incentive to have large families no
longer exists, for the land base of a family cannot be extended any
more, while the unit productivity of labour tends to decline. In other
terms, the economic value of children also declines. It has also been
noted that land shortages also bring about a rise in age at marriage (as
young men have to wait longer before they can settle on their own farm),
and this is a factor of fertility decline.

In fact, it does seem that access to land, more than the very
nature of agrarian societies, is an important determinant of fertility
in rural populations. It has been noted, for example, that, the more
plentiful the supply of land, the higher the fertility in the population
living off it 1/. At the household level, there usually is a
correlation between the size of the holding and fertility. How should
these data be interpreted ? Is it the availability of land which
fosters a high fertility (because the existence of land reserves
increases the "economic value" of children) ? Or is the reverse true,
i.e., is it the existence of a large family which makes it necessary to
extend the holding ? For we must not forget that correlations are not
conclusive in themselves: they indicate factual connections, but do not
provide explanations to these. A time perspective is therefore
essential; and in fact it supplies us with interesting data.

In certain rural areas in Hungary, the serfs' families began to
practise birth control as early as the second half of the 18th century
in order to avoid having to divide their holdings between many children
(and hence incur a fall in standard of living as well as in social
status) at a time when land was becoming scarce. At the same time,
landless workers maintained a high level of fertility: for them,
children undoubtedly constituted a positive economic value, since they
could always find work while parents did not have to be concerned with
the preservation of a holding. The differences in behaviour persisted
beyond the moment when agricultural employment reached its peak as
industry developed. Birth control then became general among independent
farmers, while landless workers took to emigrating to the towns and to
foreign countries [65]. These differences show that, in a peasant
society, status, and especially access to the land, is an important
factor in determining demographic behaviour. This point is corroborated
by the following example.

1/ See [119]. Fertility was higher for example in the 18th and 19th
centuries among European settlers in America than in Europe itself.
Observations on the same lines have been made nowadays in Asia
(India, Philippines, Thailand, etc.) and in Latin America.

- 37 -

In Kerala (India), the poorest families have reacted to the
continuous rise in population density in recent decades, not only by
intensifying production, but also by reducing their fertility. It
appears that this development has been facilitated by agrarian reform,
which has considerably attenuated the concentration of land ownership
and provided families with a chance to obtain access to small holdings.
Demographic behaviour was then adapted to the need to keep the holding

The opposite case in which there is a rapid increase in the
land availability can be illustrated by the example of Sine Saloum
(Senegal). The result of the development of this area by the colonial
administration was a rapid expansion of the cultivated area, as well as
an intensification of the use of the land already cultivated. The large
increase in demand for labour resulting from these changes probably
contributed to bring about the rise in fertility which was subsequently
observed [102].

It thus appears that, as long as land is available and tenure
systems are flexible, it is a rational behaviour to have as many
children as possible (and as many wives, where possible) and to combine
as much land as possible with the available labour. In other words, as
long as the growth in agricultural production is mainly obtained by an
extension of the cultivated area, high fertility is a logical component
of the system. When land becomes more scarce and tenure systems become
more rigid, on the contrary, it is the size of the family which must be
adapted to the size of its economic base. In this phase, in effect,
progress in raising the standard of living depends more on technological
factors than on labour availability (see Chapter 5). This being so, one
is prompted to wonder about the future evolution of fertility in the
rural societies of developing countries. There are grounds for
believing that the reason why some regions lag behind others in terms of
fertility decline is partly linked to the fact that these regions are
not yet as near to the saturation of their available land as the others.
The inevitable arrival of this saturation will most likely weigh on the
future of fertility trends, for instance in Africa.


Mortality was the first demographic phenomenon to be studied on a
statistical basis (in 18th century England). The statistical tools used
in this study are therefore traditional ones and first of all, death
rates, which measure the frequency of deaths in a population.

The crude death rate (CDR) is the ratio of the annual number of
deaths in a population to the average population during the period:

CDR = Number of deaths
Average population

It is generally expressed in deaths per 1,000 inhabitants.

- 38 -

Annex I provides estimates of the CDR in 148 countries for the
period 1980-85. The lowest rates are of the order of 5 deaths per 1,000
inhabitants, and are found in the following countries: Costa Rica,
Panama, Bahrain, Qatar, Kuwait and Singapore. It may be surprising to
find that this list does not include any developed countries. The
explanation of this fact will be found hereunder. It illustrates the
defects of the CDR as an index of mortality conditions. The highest
CDRs are of the order of 23 deaths per 1,000 inhabitants. Twenty
countries have a CDR of over 20 per 1,000; 17 of them are African
countries and 3 are found in Asia.

When the age of the deceased persons is recorded and the age
distribution of the population is known, it is possible to calculate,
with the above formula, age-specific mortality rates (generally by age
group). These rates are always calculated separately for males and

Mortality rates vary markedly with age. Graph 2 shows a typical
"profile" of mortality rates for developing countries. The rates are
fairly high immediately after birth, then they fall rapidly (to less
than 2 percent) between 10 and 40; the minimum is found at about ages 10
to 15. Thereafter, the rates increase with age, first of all gradually
and then more rapidly: hence the typical U-shaped curve. Age-specific
mortality rates are traditionally presented as arranged in a life table,
which describes the effects of mortality throughout the life cycle of a
group of persons. Table 9 provides an example of such a table.

The marked variability of the age-specific mortality rates
explains why the crude death rate is an unsatisfactory index. Of two
populations with the same levels of age-specific mortality in a given
year, one may have a lower crude death rate than the other if it has a
higher proportion of youth. In effect, the CDR is a weighted average of
the age-specific rates, and will therefore vary depending not only on
the levels of these rates, but also on the "weights", i.e., the
proportions of population in the various sex and age groups. In fact, a
number of developing countries have lower CDRs than most developed
countries, although the levels of age-specific mortality are indeed
lower in developed countries. This is because the older generations -
those with high mortality rates have a much greater "weight" in the
age distribution of developed countries (see section 1.4), a fact which
causes the "older" populations to have more deaths in proportion to
their size.

For this reason, the best index of mortality is one which is
independent of the age structure of the population. This index is the
expectation of life at birth (ELB), which represents the mean length of
life of individuals who would be subjected from birth to the mortality
conditions of the life table. Let us clarify the concept of life
expectancy by examining a life table: see Table 9.

The levels of age-specific mortality are expressed in the table
by "death probabilities" (column 2), i.e., the probabilities that an
individual of exact age x will die before exact age x+n (in our case,
before age x+5). These probabilities are derived from the age-specific

- 39 -

Graph 2 Profiles of mortality by sex and age (Ecuador 1978)

(p. 1000)

-0 25 3 A 45 50 5 Age
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85

- 40 -

mortality rates (see above) by means of an algebraic transformation
which may be omitted here.

The death probabilities are used to describe the action of
mortality on a hypothetical generation 1/ (column 3). In our example,
an initial number of 10,000 is assumed. The probability of a newborn
child dying before reaching its first birthday is 0.0835 (or 83.5 per
1,000). There will then be 835 deaths among the 10,000 persons, which
leaves 9,165 survivors at age 1. As the probability of dying between the
ages of 1 and 5 is 0.0423, there will be 9,165 x 0.0423 = 388 deaths
during this period, which leaves 8,777 survivors at age 5, and so on.

Table 9 Life table:

Tunisia, 1973-77, males

1 2 3 4
Age Mortality ISurvivors Deaths between|
x Quotients I at age x x and (x + n)

0 1 0.0835 10 000 835
1 I 0.0423 9 165 388
5 1 0.0089 8 777 I 78
10 1 0.0067 8 699 I 59
15 I 0.0095 8 640 1 82
20 0.0121 8 558 I 103
25 0.0137 8 455 116
30 1 0.0164 8 339 137
35 I 0.0188 8 202 I 154
40 1 0.0248 8 048 I 200
45 1 0.0371 1 7 848 293
50 1 0.0541 7 555 409
55 1 0.0802 7 146 I 573
60 1 0.1281 1 6 573 842
S 65 1 0.2179 5 731 1 249
S 70 I 0.2892 4 482 1 296
S 75 I 0.4891 1 3 186 1 558
80 1 1 628 1 558

Source: [166]

The principle underlying the calculation of life expectancy at
birth is as,follows. It is assumed that:

the 835 persons dying between ages 0 and 1 have lived on an
average 0.5 years;

I/ I.e., a group of persons born within a calendar year.

- 41 -

the 388 persons dying between ages 1 and 5 have lived on an
average 3 years;

the 78 persons dying between ages 5 and 10 have lived on an
average 7.5 years;

the 1,628 survivors at 80 die on an average at 85.

The total (T) of the years lived by the members of our
hypothetical generation is therefore:

T = (835 x 0.5) + (388 x 3) + (78 x 7.5)
+ (1,558 x 77.5) + (1,628 x 85) = 586,929

and ELB = T = 58.7 years.

Annex I gives estimates of ELBs by country for the period 1980-85.
It is easy to confirm that the classification of countries by level of
mortality, as supplied by ELBs, differs sharply from that provided by
crude death rates. Thus, all industrial countries have an ELB of over
70, but only half of them have a CDR of less than 10 per 1,000. As
against this, only a quarter of the 43 developing countries with a CDR
of under 10 per 1,000 have an ELB of over 70, that is, a really low
mortality level. It will therefore be seen that the differences in age
structure between "young" and "aging" countries cause considerable
distortions in the comparison of crude death rates.

In developing countries, the general level of mortality is
strongly influenced by the level of infant mortality, that is, by the
mortality of infants under age one. Infant mortality rates are always
high in relation to the other quotients, and the population at risk,
i.e., the children born in the year, is generally speaking, the largest
group of the population single-year. Infant deaths may therefore
account for as many as a third of all deaths.

Apart from accidental factors, mortality is obviously influenced
by general health conditions, the incidence of illness, the ability of
the population to stand up to illness (which is influenced by
nutritional status) and the extent to which they have access to medical
care. Mortality is generally higher in rural areas than in towns
because of the less favourable conditions as regards the above factors.


Migration is a move in geographic space which involves a change in
usual place of residence and implies movement across an administrative

- 42 -

The propensity to migrate is strongly influenced by individual
demographic characteristics:

Sex: Apart from family migration (which affects both sexes
equally), there are two main cases. Where the main cause of
individual migration is the search for work, migration
predominantly affects males, although there are notable
exceptions to this tendency (e.g. in Latin America). Where
patrilocal marriage is the main cause of migration, migrations
then involve mostly females.

Age: Children and elderly persons migrate much less frequently
than adults. The propensity to migrate broadly follows the
curve of the level of participation in economic activity.

Matrimonial status: Unmarried persons migrate much less than
those who are married. Marriage itself obliges at least one of
the spouses to change residence.

All observations on migration between rural and urban areas,
whether in developed or developing countries, show that there is a net
migration from the rural areas and a net immigration into towns.

A knowledge of migratory flows is obviously of interest in
agricultural and rural planning 1/. It is perhaps in the preparation of
regional programmes and projects that it is most useful to evaluate the
emigration and immigration flows in the area concerned by the plans
being worked out. The measurement of migration is a complex field in
which the choice of method is of great importance. It cannot therefore
be too strongly recommended that professionals be careful in formulating
the questions to which they are looking for the answer. Are they
interested in knowing about seasonal migration ? or long-term migration ?
the intensity of these phenomena ? their causes ? or possible means of
influencing them ? The replies will have to be sought in collaboration
with demographers and sociologists.

The main indices of migration are the migration rates and the
proportions of migrants. Usually, annual migration rates would be
calculated; they are ratios of the average annual number of movements
during a certain period, to the average population of the period. In
particular, the rate of net migration (RNM) will be the ratio of net
migration (immigration movements minus emigration movements) to the
population. Obviously, this will be equal to the rate of immigration
minus the rate of emigration:

RNM = Immigration Emigration = Immigration Emigration
Average population Aver. pop. Aver. pop.

1/ See section 6.5.

- 43 -

Proportions of migrants can be obtained by relating the number of
migrants during a period to the population to which or from which they
are migrating. When the proportion of out-migrants is obtained by
dividing the number who reported moving out of the area by the
population residing in the area at the beginning of the period and alive
at the end, this index measures the probability of moving for the
population, and among other uses, it can be used in the preparation of
population projections where migration is accounted for separately.
Similarly, the proportion of in-migrants is sometimes obtained by
dividing the number of in-migrants in an area during a period, by the
population of the area at the end of the period (but the denominator
could also be the population at the beginning of the period, or the
average population).


Natural increase is the net result of births and deaths in the
population. The rate of natural increase (RNI) is the ratio of the
excess of births over deaths, to the average population 1/. It is
therefore also the difference between the crude birth rate and the crude
death rate:

RNI= Number of births Number of deaths = CBR CDR
Average population

The rate of natural increase may be negative.

The growth of an open population takes into account external
migrations. At the national level, the growth rate (GR) is therefore
the sum of the rate of natural increase and of the rate of net migration

GR= Births Deaths + Immigration Emigration = RNI + RNM
Average population

It is of course possible to define and caculate similar rates for
regions within countries and for the subpopulations therein. Internal
migration is then also taken into account. In this way, it is possible
to evaluate the differences in the rate of increase between regions and
the resulting changes in the spatial distribution of the population.


We shall describe here population growth and structural
transformations as observable at the present time in the world, and in
particular in developing countries 2/.

1/ The population at the beginning of the period is sometimes used as
the denominator.

2/ See [108]. This short booklet provides a succinct but complete
description of the world population situation.

- 44 -

3.5.1 Fertility

For the period 1980-1985, the birth rate worldwide is estimated
at about 27 per 1,000. For the last 20 years, birth rates have declined
considerably in all the main regions except perhaps Africa.

The fall in these rates has been particularly marked in eastern
Asia, where they declined by about a third. Given the size of the
populations concerned, this fall has had a marked impact on world
population trends. Rates have also declined significantly in Latin
America and Southern Asia. In Africa, which has the highest birth rate
among the world's regions (46 per 1,000), the lack of reliable
statistics makes it difficult to assess the trends. There may have been
a slight decline in the regional birth rate in that continent since

For the developing regions as a whole, the birth rate appears to
have fallen from 40 per 1,000 in 1960-65 to 31 per 1,000 in 1980-1985,
which is still twice the average rate in developed regions.

Researchers have made great efforts to identify the factors
affecting fertility and its evolution in developing countries. Although
work in this field has been rendered difficult by the lack of reliable
data, observations suggest "that declines in fertility in less developed
countries between 1960 and 1975 were determined mainly by factors other
than progress in social and economic development and the reduction of
mortality rates during this period". However, the findings confirm the
view that "the attainment of a relatively high level of development and
low level of mortality is a condition favourable to the decline in
fertility, although it may not be an absolute prerequisite" [108].

The available data on this question are derived from country
surveys, and have therefore no general validity. Nevertheless, it was
repeatedly found during the 1970s that there were differences in
fertility linked to the level of women's education, to a tendency
towards later marriage, and to the practice of family planning 1/.

3.5.2 Mortality

There has been a remarkable reduction in the death rates in the
developing regions as a result of the progress in improving health
conditions. Between 1950-55 and 1980-85, expectation of life at birth
in these regions rose from 41 to 57 years. A smaller, but substantial
improvement was also registered in the developed regions, where average
ELB rose during the same period from 65 to 73 years.

It has already been noted that infant mortality weighs heavily on
general mortality. It will therefore be understood that the decline of
mortality in developing countries was largely due in the past (and will
depend to a great extent in the future) to the reduction of the risk of
death to new born babies and to infants. In developed countries, a

1/ The work carried out by the World Fertility Survey deserves
particular attention in this connection, especially the Colombia,
Fiji, Malaysia, Republic of Korea, Dominican Republic and Thailand

- 45 -

substantial improvement in life expectancy could only come from a fall
in mortality among the elderly, which will be difficult to achieve.

There are not, of course, two clearly differentiated types of
mortality only. Developing countries are at different stages of the
transition from the high mortality generally prevailing before great
progress was made in medicine and hygiene, toward the low mortality of
developed countries who benefit from this progress and from favourable
nutritional and environmental conditions.

Table 10 provides estimates of the ELB (for
females) in 1950-55 and in 1980-85 by region.

Table 10 Estimates of expectation of life at birth

for some regions, 1950-55 and 1980-85

.both males and

worldwide and

I Males 1 Females I Total

S1950- 11980- 1950- 1980- 1950- 1980-
S1955 1985 I 1955 1985 1955 1985

SWorld total I 44.7 I 57.5 I 47.1 1 60.3 I 45.8 58.9

IMore developed I 63.0 I 69.4 I 68.7 I 76.9 I 65.1 1 73.0

ILess developed I 40.3 1 55.5 1 41.8 I 57.7 I 41.0 I 56.6
Africa 36.2 I 48.2 I 38.8 I 51.3 I 37.5 I 49.7 I

America 55.5 I 63.9 59.8 69.4 I 57.6 I 66.6

Asia 40.6 57.2 41.8 58.7 41.2 I 57.9

Europe 63.2 70.0 67.6 76.6 65.3 1 73.2

SOceania 59.2 65.5 62.9 I 69.9 61.0 1 67.6

The estimates for the developing regions cannot be taken as
certain, either as regards the level of life expectancy or its trend 1/.
Nevertheless, there is no doubt that the lengthening of the average
length of life has been substantial in these regions.

1/ The only reliable data are those for Latin America.

- 46 -

Latin America had the lowest mortality among the developing
regions in the early 50s, and there has been a further improvement in
that respect since that time. The most striking progress has been
registered in Central America and the Caribbean. Most of the 24
countries in the region are now in the low mortality category (with a
life expectancy of over 60). Only two countries (Haiti and Guatemala)
still have a life expectancy of less than 60 years.

In Asia, there is still uncertainty about the levels of mortality
for China and India, where the huge size of the population influences
the average demographic levels of the region. Nevertheless, it seems
from available data that life expectancy has progressed considerably in
China in the last few decades. The picture appears to be less
favourable in India, where progress seems distinctly slower. Most of
the countries in Southern Asia, including the four largest ones
(India, Indonesia, Bangladesh and Pakistan), are in the high mortality
group. In Eastern Asia, on the contrary, most of the countries should
be placed in the low mortality group (very low for Hong Kong and Japan).

In Africa, the situation as regards mortality is much better in
the northern (Mediterranean) part of the continent than in the
subsaharan region where mortality is higher than in any other region.
Certain countries in the south and east of Africa avoid inclusion in the
high mortality group because of the presence of large allochtonous
minorities with a low mortality. The lack of reliable data makes it
impossible to make firmly based estimates, but it is thought that there
has probably not been any rapid progress in the reduction of death rates
in Africa during the last decade. Food shortages and malnourishment are
the main obstacles to a reduction of this level.

3.5.3 Population growth

The recent growth of world population is unprecedented in the
history of the species. At the end of the eighteenth century after
several million years the world population amounted to one billion
people. It took little more than a century (roughly 1800-1925) to add
the next billion, half a century (1925-76) for the third and the fourth
billions, and the less than twelve years for the fifth billion. The
last quarter of the present century will no doubt witness a further
increase of two billion, to which still others will be added in the
course of the next century.

The maximum speed of this increase was reached around 1960, when
the annual rate was 2.0 percent (i.e., four times the average rate in
the 19th century). There has been a slight slowing down since then.
The present rate is about 1.7 percent. The average rate for the period
1975-2000 might be 1.6 percent,.which would still be very high in a
long-term perspective.

A considerable proportion of the present population increase is
taking place in the developing regions. Between 1950 and 1975, the

- 47 -

population of these regions 1/ increased twice as much as that of the
developed regions (75 percent against 36 percent). Moreover, the size
of the population at the beginning of that period was about double that
of the developing regions (1.6 billion against 0.8 billion), the result
is that the latter regions accounted for 83 percent of the world-wide
increase (1.26 billion out of 1.52).

Table 11 gives estimates and projections of the population for
eight regions. The projections are those of the medium variant of the
United Nations.

Table 11 Estimates and projections of the population of the world
and major regions, 1950-2000

S Population (millions) ITotal growth (percent)I
1950 I 1975 1 1985 I 2000 1 1950- 1975-
SI I I 1975 I 2000

IWorld total 2504 I 4076 I 4842 I 6127 62.8 I 50.3

IMore Developedl 832 1095 1 1173 I 1276 I 31.6 I 16.5

ILess Developedl 1672 2981 I 3669 I 4851 1 78.3 I 62.7

Africa 222 I 410 I 553 1 877 I 84.7 1 113.9

America 331 1 560 1 670 848 I 69.2 51.4

Asia 1366 1 2357 I 2824 I 3544 72.6 50.4

Europe 392 474 492 513 20.9 8.2

SOceania 13 21 25 30 61.5 42.9

Source: [121]

1/ Africa, Latin America, Asia (minus Japan and the USSR), and Oceania
(minus Australia and New Zealand).

- 48 -

The region with the fastest population growth in the third
quarter of the present century was Latin America, where the population
has practically doubled despite slow growth in Argentina, Uruguay and
countries in the Caribbean. Despite the recent fall in the rate of
population growth in that region, the United Nations "medium" projection
still gives an increase of about 90 percent for the period 1975-2000.

In Africa, a higher rate of growth is forecast for the last
quarter of the century than for 1950-1975. The reason is that there will
doubtless be progress in reducing the high rates of mortality which are
still largely prevalent in this region, whereas no fall in the average
birth rate for the whole of Africa is expected before the 90s.

In terms of numbers, Asia (Southern and Eastern) has completely
outdistanced the other regional groups since 1950. Southern Asia (from
Turkey to Indonesia) will long remain the most populous region in the
world. India accounts for almost half the population of this region.
In Estern Asia, China has 84 percent of the population and Japan 10
percent. The "population giants" dominate this part of the world which,
even without counting the USSR, contains six of the nine most populous
countries of the world. Between 1950 and 1975, Eastern and Southern
Asia accounted for over 60 percent of population growth world-wide. In
the "medium" projection, the population of Southern Asia will rise to
2.2 billion at the end of the century. No great slowing down in growth
is expected in South Asia, while in Eastern Asia the growth rate will
doubtless further decline, being strongly influenced by demographic
trends in China, where the rate of growth might fall to close 1 percent
a year in the period 1975-2000.

In Africa, it is possible that, during the past twenty years, the
overall birth rate has fallen slightly, but the death rate has declined
more sharply, so that the (extremely high) rate of natural increase
seems to have accelerated since the 60s. The other regions with a high
growth rate are Latin America and Southern Asia. However, the birth
rate has fallen in these regions, but only just enough to balance the
reduction in mortality. The whole of Eastern Asia has achieved a level
of natural growth not much higher than that of the developed regions as
a result of a sharp decline in the birth rate. At the present time,
Eastern Asia has much the same characteristics as Oceania, despite the
predominance, in terms of numbers, of two developed countries (Australia
and New Zealand) in the latter region.

3.5.4 Age structure

We have seen in a previous chapter that a decline in fertility,
given constant mortality, results in an immediate drop in the proportion
of children in the population. .In the longer run, a prolonged fall in
fertility leads to an upward shift of age distribution, that is to say,
it causes an aging of the population. The effects of a fall in morality
make themselves felt, but to a smaller extent, in the opposite
direction. A lower mortality among children is translated, at unchanged
fertility, into an increase in the proportion of children. An
improvement in the survival of adults and the elderly, on the contrary,
makes for the aging of the population.

- 49 -

The net effects of these factors on the age structure of
developed countries were described in Chapter 1: a reduction in the
proportion of children, a considerable increase in the proportion of
adults and a more modest increase in the proportion of the elderly.
These effects are brought out in the contrasting age structures of the
developed and the developing regions.

In developing regions, the percentage of children in the
population increased slightly from 1950 to 1970. On the one hand, there
was a fall in fertility in a number of countries, but this was more than
offset by the decline in infant mortality. Since then, there have been
only slight changes in the stucture in these regions as a whole,
although there has been a marked reduction in the proportion of children
in certain countries those in which the decline in fertility has been
rapid. Africa is the region with the youngest structure at present,
with about 45 percent of the population under 15. This proportion is
slightly less in Latin America and Southern Asia, and much less in
Eastern Asia.

Between now and the end of the century, it is to be expected that
the percentages of children will fall in the whole of the developing
regions from about 41 percent in 1975 to about 34 percent in 2000 for
those under 15. The consequences of this evolution (and of its
longer-term impact on the active population) undoubtedly deserve to be
studied at the country level.

3.5.5 The growth of urban and rural populations

Urbanization 1/ is one of the distinctive features of modern
population dynamics. The growth of towns, particularly in developing
regions, has been spectacular in the course of the last quarter of a
century: the urban population world-wide has more than doubled, and, in
the developing countries, it has almost tripled. The proportion of the
urban population has gone up from 29 to 41 percent for the world as a
whole and from 17 to 31 percent in the developing regions.

The rate of increase in the urban population is particularly high
in Africa, where the number of urban residents will probably have
decupled between 1950 and 2000. But in fact, all the developing regions
have registered a marked increase in urban population since 1950. Table
12 gives estimates and projections of the proportion of urban
populations in the major regions for the end of the present century.

I/ That is, the increase in the proportion of the population living in
urban areas.

- 50 -

Table 12 Percentage of urban population world-wide and for
various regional groups, 1980 and projection 2000

Whole world
Developing regions
Developed regions
Latin America
North America
Eastern Asia
Southern Asia





Source: [108]

Although the proportion of the rural population is declining,
there continues to be an increase in absolute numbers in the developing
regions. Between 1950 and 1975, this increase averaged 50 percent in
these regions, and indeed was even more in Africa and in Southern Asia.
For the future, it is possible that the rural population growth in
Eastern Asia will stop towards the end of the century, and slow down in
Latin America and in Southern Asia, but no significant reduction is
expected in Africa for that period.

It is worth emphasizing the reason for the more rapid growth of
urban population in developing countries. This is not due to natural
increase, which in fact is as high or even higher in rural areas, but to
internal migration which (together with the change in the category of
certain localities) accounts for 40 percent of the increase in the urban
population of these countries 1/.

I/ A study of the last intercensal period in a sample of 29 developing
countries gives average rates of natural increase of 2.7 percent in
rural areas and 2.5 percent in urban ones [103].



In this chapter, we shall describe the general framework for the
use of demographic variables in agricultural planning. For this
purpose, we shall illustrate in general terms the relevance of the
main variables and indicate how they can be dealt with in the
perspective approach which is inherent to planning. The three following
chapters, dealing with employment, migrations and food, will delve into
the substance of the main issues involving socio-demographic components
which are constantly coming to the attention of agricultural planners.


Here we shall consider the size and the sex and age structure of
the total population, and indicate first of all in what respects these
variables are useful, and subsequently the methods by which projections
of these variables can be obtained.

4.1.1 The use of the variables

The total population size at the terminal date (or horizon) of
the planning exercise and, if need be, at intermediate dates, is an
essential datum for this exercise. It is the basis of all calculations
of needs, and is the denominator of all economic indices expressed as
"per caput" (in particular per caput income), some of which often
constitute the targets of the plan being studied. For example, in
agricultural planning, the (projected) per caput demand for a particular
commodity will be multiplied by the size of the total population in
order to derive total demand. In this context, the total population
size is obviously the reference datum which ensures consistency between
the different variables handled in the course of the analysis [7].

The sex and age structure of the population is also extremely
important. As was noted in an earlier chapter, it is relevant for the
study of employment questions. That structure also affects the analysis
of requirements, in particular those for food (see Chapter 7). Lastly,
it is indispensable for the analysis of population trends influencing
the rate of growth and the future level of the total population and of
the various sub-populations relevant to planning 1/.

1/ See the following subsection.

- 52 -

Population dynamics, i.e., the evolution of size and structures,
is in fact a fundamental aspect of the processes which will affect the
economy and will in turn be affected by it during the planning period.
For example:

the rate of population growth is closely linked to the
dependency ratio, which may have an influence on the related
rates of consumption and savings;

this rate also influences the requirements for productive
investment ("broadening of capital");

it may, through its action on age structures, also influence
the composition of consumption needs and expenditure;

if related to available agricultural resources, the population
growth rate must be considered in choosing the strategy to be
followed for agricultural development (self-sufficiency versus
dependency on trade, etc.) [7].

It should be added that even more specific uses of these
variables can be identified when planning addresses the problems of
small populations, particularly at the level of investment projects. We
shall revert to this point later.

4.1.2 Population projections

Projections are intended to provide forecasts, that is, estimates
of the future state of the population. At the simplest level, which is
the projection of size alone, it will be enough to apply a hypothetical
rate of growth to the estimated population size at the beginning of the
planning period. In this way, an estimate of future size will be
obtained directly. This method may be sufficient for a short-term
forecast, for it is then unlikely that there will be a sharp divergence
between the rate of growth and the recent trend (which is taken as a
basis for determining the hypothetical rate), and even if there is such
a divergence, the error will be relatively small in view of the short
period involved. But in practice the planner needs to know not only the
size, but also the sex and age structure of the population. He is
therefore obliged to resort to more detailed methods, such as the
"components method" which is universally used nowadays.

If the mechanism of population projections is well understood,
the professional can himself carry out such a projection in case of
need, and, at any rate, more easily communicate with a statistical
office or with demographers, from whom the planner may obtain the
projection of a particular population (that of the country over a
specified period, or that of a project area). The said mechanism is
also interesting for the layman in that it reflects and illustrates the
internal dynamics of a population.

- 53 -

Let us take as an example the population of Indonesia in 1985,
and try to estimate the growth and changes in structures of this
population between 1985 and 1990. We shall show here how the projection
of the female population is made; the method would be exactly the same
for the male population. Calculation of survivors

The problem here is, given the size of each five-year age group
at the beginning of the projection period, to calculate the number of
survivors in these groups at the end of the period. This number is
obtained by applying to the initial size a survival ratio, which ought
to be selected ex hypothesis. At this stage, it is therefore necessary
to formulate assumptions regarding the future levels of mortality by age
group. In substance, this hypothesis will depend on the available
information about the present level of and the trends in mortality.
The assumption will be formalized as a series of 5-year survival ratios
(see column 3 in Table 13). When no detailed information is available
on mortality at the country level (which is often the case in developing
countries) recourse is made to model life tables 1/. In the present
case, female life expectancy at birth for the period 1980-1985 was
estimated at 53.9 years. We have therefore selected from a collection
of model life tables [124] a female table (Far Eastern type) in which
life expectancy is 56.5 years, and applied the ratios indicated by that

The application of 5-year survival ratios to the population at
the beginning of the period provides an estimate of part of the
population at the end of the period:

persons of between 0 and 4 (in complete years ) in 1975 will
be aged 5 to 9 in 1980;

persons of between 5 and 9 in 1975 will be aged 10 to 14 in
1980; etc.

1/ See Annex III.

- 54 -

Table 13. Projection of the female population of Indonesia, 1985 to

1990: calculation of


1 2 3 4 5
Age Numbers 1985 Survival Numbers 1990 Age
group (thousands) ratio (thousands) group
1985 1990

S 0-4 11 497.4 .9752 11 212.3 5-9
S 5-9 10 451.8 .9912 10 359.8 10-14
S10-14 9 429.2 .9890 9 325.5 15-19
S15-19 8 400.9 .9810 8 241.3 20-24
20-24 7 472.0 .9753 7 287.4 25-29
S25-29 6 680.6 .9706 6 484.2 I 30-34
S30-34 5 371.1 .9654 5 185.3 I 35-39
35-39 4 300.1 .9581 4 119.9 40-44
40-44 4 066.4 .9470 3 850.9 45-49
45-49 3 659.0 .9286 3 397.7 50-54
50-54 2 999.6 .9009 2 702.3 55-59
55-59 2 310.5 .8624 1 992.6 60-64
60-64 1 660.2 .8102 1 345.1 65-69
I65-69 I 1 241.9 .7401 1 919.1 70-74
S70-74 1 839.7 I .6502 1 546.0 75-79
S75-79 483.4 I .5377 I 259.9 80-84
80 + 322.3 1 .3610 116.4 85 +

All 81 186.1 (77 345.7) (5 & +)
I I I I I I Calculation of

births and of surviving children

The calculation of the births to be expected during the five
following years obviously involves formulating a hypothesis regarding the
future level of fertility. This hypothesis may be formulated with
a variable degree of detail, depending on the nature of the information

If possible, a hypothesis will be formulated regarding the
level of the age-specific fertility rates, and each of these
rates will be multiplied by the average size of the female
population of the corresponding age group, and then by 5,
before adding up the births from all groups in order to
obtain the total of these births.

- 55 -

In the absence of better data, one will formulate a
hypothesis regarding the level of the overall fertility rate
(as an average for the projection period), and this rate will
be multiplied by the average number of women of childbearing
age, and then by 5, in order to obtain the number of births
in the five-year period.

In the present case, we have an estimate of the overall fertility
rate (125 per 1,000) in 1980-1985, and we know that the trend of
fertility is a declining one. Let us assume that the overall fertility
rate in 1985-1990 will be 110 per 1,000. We must apply this rate to the
average number, during this period, of women of childbearing age (15 to

In 1985, the number of women of between 15 and 49 (see Table 13,
column 2) was 39,950,000. In 1990 (column 4), it will be 44,494,500.
The average size in the period is half the sum of these figures, i.e.,
42,222,300. The number of births for the five-year period will
therefore be:

5 x 0.110 x 42,222,300 = 23,222,300

If we assume that there will be 51 percent of male births and 49
percent of female ones, the respective figures for these births will be:
11 843 400 male births and 11 378 900 female births. We then apply to
the latter figure the appropriate survival ratio in order to obtain the
number of surviving children at the end of the period:

11,378,900 x 0.9164 = 10,427,600

The series of numbers by age group in 1990 is thus complete.

It will be noted that the procedure described above takes no
account of possible migratory movements. This is the way in which most
projections of population at the national level are prepared and

However, in quite a few cases, it is known that international
migration is by no means negligible, and it is therefore desirable to
take account of it. Data are often lacking for the formulation of sound
hypotheses regarding migration, but it is better to have an imperfect
hypothesis than a projection leaving out migration altogether, which in
fact would mean a projection using a zero migration hypothesis often a
particularly unrealistic one. This is true not only for national
projections, but also, and especially, for population projections at the
regional level.

The classical way of handling migrations in a projection
procedure is to set the net expected numbers of migrants (arrivals less
departures), by sex and age group, and to add these numbers to the

- 56 -

figures provided by the calculation of the survivors 1/. The rest of
the procedure remains unchanged (unless an amount of net migration is
added for children between age 0 and age 5).

We have presented in the preceding paragraphs, an example of a
population projection with a single hpothesis for each of the phenomena
affecting the future evolution of the population (mortality, fertility
and migration). It may of course be desirable to make several
hypotheses regarding one or other of these variables. However, this
could lead to a proliferation of projections which would reduce their
usefulness. For instance, with two hypotheses for mortality, three for
fertility and two for migration, we would obtain 12 projections, which
is too cumbersome an apparatus to be useful.

In practice, one should seek to cover the range of possible
evolution, but stick to the most probable ones, and hence limit the
number of variants. In most cases, since the projections for planning
purposes are short-term ones 2/, one should limit them to two or three.
It should be added that all economic calculations are not to be repeated
for each of these projections. One or two projections can be used as a
reference for the preparation of the plan framework, i.e., that of the
major economic aggregates, while other projections may be drawn up to
meet the needs of certain analyses: for example, differing alternatives
for the level and structure of migrations which will be used for the
analysis of employment policies.

Certain projections are carried out with the sole aim of
illustrating the medium or long-term consequences of specified
demographic trends, or of the hypothetical evolution of certain
variables. These projections, since they are not designed for
forecasting, cannot be assigned any a priori coefficient of likelihood
and can generally not be used for planning. What is needed in planning
is a type of population projection adapted to its requirements, and
these the planner will seek to obtain from demographers. He should
therefore understand the nature of the work carried out by the latter.

Lastly, we cannot insist too strongly on the fact that the
quality of a projection does not depend on the mathematical procedure
employed. What is important is the quality of the basic data and
especially the relevance of the hypotheses. These, in their turn, call
for a sound analysis of the situation and trends.

1/ Net migration may of course be negative, and therefore result in a
diminution of the totals.

2/ In a demographic perspective, five years are regarded as short term.

- 57 -


4.2.1 The use of the variables

Projections of the agricultural population indicate the probable
future level of the population which depends on agriculture for its
livelihood. As such, they are used in many different ways for
agricultural planning. For example, a desired level of per caput income
will be applied to the estimated agricultural population, and this will
make it possible to estimate the target growth for agricultural
production. Conversely, if the targets for agricultural growth are
given, the projections of agricultural population will enable planners
to check whether these targets are acceptable in terms of per caput
income. Moreover, significant conclusions can be drawn from comparisons
between the present and potential use of land on the one hand and
projections of the agricultural population on the other [125].

The active agricultural population, or agricultural labour force,
indicates the extent of the sector's employment requirements. This
consideration is of the greatest importance for agricultural planning
when it is related to manpower requirements. The labour force variable
is then incorporated into two main types of analysis, which themselves
are subject to multiple variants [7].

The first type of analysis is that of the formalized models of
sector analysis. In this perspective, forecasts of agricultural
manpower are exogenous. It is within the model that the levels of
utilization of that manpower are determined, area by area and with
a distinction between family workers and wage labour. The supply
functions of labour used in these models link the amount of labour
available to the wage level and distinguish different types of behaviour
for family and wage labour. The activities included in the model
correspond to alternative techniques for the production of a given
foodstuff in a particular area. The labour coefficients for each
activity are estimated on the basis of actual data. The solution of the
model makes a selection of the levels of the different activities and
indicates the corresponding levels of employment and wages. When the
levels of labour utilization indicated by this solution are regarded as
unsatisfactory (as being too low, for example), alternative solutions
may be obtained within certain limits by taking adequate levels as
constraints of the model.

The second type of analysis, which is the most common one,
consists in determining coefficients for the use of labour for each
agricultural activity, that is to say, by crop and production technique,
distinguishing between the operations to be carried out, from soil
preparation to harvesting, and taking account of the seasonal
distribution of the activities. This analysis can be effected a
posteriori, that is to say, after the mix of crops and techniques has
been determined in the light of other criteria such as the demand
structure or the natural resource potential. The incidence of this
analysis on employment can be taken as a point of departure for a
process of iteration designed to harmonize the technical aims and the
plans for the absorption of labour.

- 58 -

The analysis of agricultural employment may show that there is no
solution for the problem of finding an acceptable level of absorption
for agricultural labour. What is needed then is an intersectoral
approach at the rural level. The planner will be obliged to modify
certain parameters previously regarded as exogenous, such as the
distribution of the population between urban and rural areas or the rate
of growth of non-agricultural sectors [7].

4.2.2 Projections

Agricultural population and labour, as opposed to the total
population, cannot be considered as exogenous variables, that is to say,
exclusively determined outside the plan, by the agricultural planner.
Their size and relative weight are influenced by the evolution of other
variables in the system, since the most visible component in their
evolution is intersectoral transfers (with or without migration) of
population and labour. These transfers are of course affected by the
evolution of the economic system and by the policies included in the
plans and programmes themselves. For the planner, these variables ought
therefore to be determined jointly with the others, and not taken as

The demographic framework of projections of agricultural
population and labour may be described as follows [125]. The factors
affecting the evolution of the agricultural population are (as in the
case of any other sub-population) fertility, mortality, immigration and
emigration. In the case of migration, what we are dealing with is
obviously transfers to the non-agricultural sectors ("emigration") or
from these sectors ("immigration"), which do not necessarily imply a
change of residence. The agricultural population (Pag) after a given
period (t) may then be expressed by the formula:

Pag (t) = Pag (0) + n(0,t) Pag (0) m(0,t) Pag (0)

+ i(0,t) Pag (0) e (0,t) Pag (0)

where n(0,t), m(0,t), i(0.t) and e (0,t) are respectively the birth
rate, the death rate and the assumed rates of immigration and emigration
for the period extending from the present moment (0) to date (t) 1/.
The corresponding active population, or agricultural labour force (Mag),
is obtained by applying to the agricultural population the assumed
activity rate for the end of the period, at time (t):

Mag (t) = a(t). Pag (t)

1/ The reader may have noted that these "rates" differ from the
classical rates (see Chapter 3), since they refer to the population
at the beginning of the period, and not to the average population for
the period.

- 59 -

Projections of the agricultural population may theoretically be
obtained by the component method. However, in most countries, the lack
of data on the components of the evolution of agricultural population
makes it impossible to use this method. On the other hand, population
censuses always, or almost always, provide data on the active population
by sector, and hence on the agricultural labour force. This is why the
procedure used by FAO [87], and suggested as being suitable for
countries to apply, is to prepare, in the first place, projections of
agricultural labour, and then to derive from them projections of the
agricultural population, using a hypothesis on the activity rate within
that population. This is to say that the framework set out above is
reversed and that the final calculation is as follows:

Fag (t) = Mag (t)

The projection of the agricultural labour force is linked to the
projections of total labour force carried out with conventional methods
the principle underlying which is the application to the total
population (projected by the component method) of hypothetical activity
rates by sex and age. Once these projections are given, the procedure
amounts to making a hypothesis about the division of total manpower into
two sectors (agricultural and non-agricultural) and to applying the
respective proportions to the total manpower as already calculated. We
shall illustrate these principles by reverting to the example of
Indonesia. Projection of the total labour force

As noted above, the future size of the labour force (taking all
economic sectors together) is estimated by applying to the total
projected population, by sex and age group, the activity rates by sex
and age which are assumed to represent the levels of activity at the
date of reference of the projection.

In our example, we have to estimate the female activity rates in
1990. We could look at activity rates as assessed by a recent census or
survey, and formulate assumptions regarding the evolution of these rates
under the effect of the progress of schooling, and of various economic
factors. If we could avail ourselves of time series of activity rates,
we also could extrapolate their trends. 1/ For our present purpose,
however, we shall use the projection of activity rates prepared by ILO
[167]: the calculation of the size of the active female population by
age is shown in Table 14.

1/ On labour force projection methods, see [126].

- 60 -

Table 14. Projection

of the female labour force of Indonesia to 1990

Age Female Activity I Active
groups population I rates population
(thousands) (percent) (thousands)

0-9 21 639.9 -
10-14 10 359.8 5.95 I 616.4
15-19 9 325.5 29.40 1 2 741.7
20-24 8 241.3 36.00 1 2 966.9
25-29 7 287.4 37.00 I 2 696.3
30-34 6 484.2 40.00 1 2 593.7
35-39 5 185.3 43.00 1 2 229.7
40-44 4 119.9 46.50 1 915.8
45-49 3 850.9 47.00 1 809.9
50-54 3 397.7 44.00 1 495.0
55-59 2 702.3 36.25 I 979.6
60-64 1 992.6 28.75 1 572.9
65 & + 1 3 186.5 16.50 525.8

All 87 773.3 24.1 21 143.7
I All I 87 773.3 24.1 I 21 143.7 I
I i I I

L9993 Fati-in~tin~

the proportion of agricultural labour force

At the present time, the proportion of agricultural labour force
is declining almost everywhere world-wide (see the following chapter).
For forecasting purposes, it is generally assumed that this decline will
continue. The whole problem is to make a realistic hypothesis about the
rate of this decline. Various methods can be used for this purpose.

One method (used by FAO in 1971) consists in establishing, on the
basis of observations provided by censuses or surveys, equations linking
the rate of decline in the proportion of agricultural labour to
different economic indicators (originally, these were the rate of growth
of agricultural GDP as a ratio of the rate of growth of total GDP; the
share of agriculture in total GDP; and per caput GDP). By making
hypotheses regarding the future level of the independent variables, one
obtains, by means of equations, an estimate of the rate of decline in
question. The drawback of this method is that there is a lack of data,
and that a large number of measurement problems arise in the attempt to
arrive at a correct estimate of the functions to be used for the



----- -~

- 61 -

In practice, it is therefore necessary in most cases to fall back
on simple methods which do not explicitly rely on relations between
economic variables. These methods are based on direct hypotheses
regarding sectoral movements of manpower overtime, i.e.:

a) a constant difference between the rates of growth of the
non-agricultural and the total labour force; or

b) a constant difference between the rates of growth of the
agricultural and the total labour force; or

c) a constant difference between the rates of growth of the
non-agricultural and the agricultural labour force.

The latter method is preferable for long-term projections (that
is, those for more than ten years), for it assumes a logistic form for
the growth in the proportion of the non-agricultural force, which is
indeed the actual pattern over the long run. The formula giving the
proportion of non-agricultural labour force, p, at a future date (t) is

p (t)=
p(0) + [l-p(0)] exp [(n-a)t]

where (n) and (a) respectively are the rates of growth of the
non-agricultural and agricultural labour force. All that is needed is
data on agricultural and non-agricultural labour force at two reference
dates in order to estimate the parameters.

It should be noted that the hypothesis regarding the respective
proportions of agricultural and non-agricultural labour force is not
independent of the contents of the plan or project for which the
estimate and projection have been carried out. For example, the trend
is often towards the diminution of the proportion of agricultural labour
force, and this trend can be extrapolated. But, if one of the
objectives of the plan is to retain labour in agriculture and if steps
are taken to that end, we must logically make a hypothesis on the final
proportion of agricultural labour force which reflects the desired
change in trend.

Having estimated the size of the agricultural labour force at a
future date, we then divide this size by the activity rate of the
agricultural population in order to obtain the size of that population.
This activity rate is usually unknown, and must be estimated. On this
point, FAO suggests to make the hypothesis that this rate is equal to
the rate of activity of the whole of the population [87]. This is
obviously an arbitrary hypothesis, but it is not an unreasonable one.
The activity rates of the agricultural population by sex and age are
doubtless different from those of the non-agricultural population
because of a greater participation by women and children, but, in most
developing countries, the proportion of the agricultural population is
very high, so that the overall activity rate mainly reflects the level
of participation in the agricultural sector.

- 62 -

This method has the additional advantage of ensuring the
consistence of the sectoral projections with the overall ones. In this
framework, the problem of preparing projections for the agricultural
population and labour force can therefore be reduced to estimating
future variations in the quantitative ratio of agricultural labour force
to total labour force. Now, ex hypothesi, the activity rates for the
agricultural and the non-agricultural population are not very different.
Moreover, the rates of natural increase of these two populations are
also not very different if we can go by the slight differences between
the rates measured for the rural and urban populations. The result is
that changes in the share of agricultural labour in total labour are
mainly due to the balance of intersectoral labour movements.


The distribution of the population between rural and urban areas
is of interest to agricultural planners from various points of view.
One of the habitual uses of the concept of rural population is, it must
be admitted, informal and unsatisfactory [66]. This is the practice of
using it as a substitute for agricultural population when information on
that population is lacking. The weakness in this utilization is
obviously the disregard for two segments of population which, if we look
at the matter closely, are of special interest to planners, i.e.:

The non-agricultural rural population, which depends for its
subsistence either on activities linked to agriculture (the
supply of inputs or the processing of commodities) or on
service activities furnished to the agricultural population
(handicrafts, trade, etc.) and which forms the potential
market for agricultural commodities in rural areas, the
diversification of which is a precondition and an indicator of
rural development;

The agricultural urban population, arising from the existence
of agricultural holdings within the territory of certain
agglomerations administratively classed as urban, or by the
development, in certain regions (such as the Near East or
Latin America), of the extremely interesting phenomenon of
"part-time agriculture 1/.

Within the urban areas, it is essential to have data on the
population of large and small towns, particularly for regional planning
in which the influence of small economic centres and local markets has
to be analyzed. In the rural areas, data are needed on the size of the
population classified by area of production or by ecological area,
since these areas differ as regards their productive potential, the
crops grown in them, the surpluses obtainable from them, their
consumption structures, and so on [8].

I/ See, for example [55].

- 63 -


The increase in population is obviously an essential factor in
the evaluation of the needs and resources involved in planning. This
increase ought to be looked at with an eye for a differential increase
in various relevant categories: the urban and rural populations; the
agricultural and non-agricultural populations; agricultural and
non-agricultural labour force; the population living off subsistence and
market-oriented agriculture, etc. As the growth differentials are
largely conditioned by the transfers of population from one category to
another, there is an added reason for studying migrations.

Migration modifies the size and characteristics of
sub-populations which are of interest to planners. Changes in food
demand, for instance, are strongly influenced by the movements of
population leaving the agricultural sector (or the subsistence
sub-sector), as well as by urbanization (see chapter 7). On another
level, migration may be regarded by intersectoral planning as a transfer
of human capital. In this context, one could imagine quantifying it as
the value (discounted at the date of migration) of the contribution
which would have been made by the migrant during the rest of his active
life, to the production of the sector of origin. Lastly, another
phenomenon accompanying migration is of clear importance for planning,
namely intersectoral transfers of income. These transfers occur in both
directions. The sector of origin provides migrants with means of
subsistence, and later receives remittances from them. The extent and
destination of these flows, especially the flow of funds into the
agricultural sector (or at least toward the agricultural population) are
of interest to the agricultural planner. The overall balance sheet of
these transfers is another interesting subject, for it is very doubtful
whether the balance is always in favour of the areas and sectors of
origin (see chapter 6).



The agricultural labour force differs from that of the other
sectors because, to a certain extent, it can be regarded as possessing a
specific dynamics. The population of working age in agricultural
households is usually regarded as forming the basis of the agricultural
labour force. Of course, nobody would think of dealing with the
industrial labour force in the same way. But, rightly or wrongly,
agricultural households are implicitly considered as particular
entities, in which the exercise of one and the same economic activity by
successive generations is fairly normal. For that reason, it is
considered that population dynamics is the first determinant of the
active agricultural population, though of course intersectoral transfers
may then have to be taken into account.


The age structure of a population is an important factor in the
relative size of its labour force for, as we have seen earlier on
(Chapter 1), persons of adult age, who make up the labour force,
represent a greater or lesser proportion of the total population
depending on the general pattern of age distribution. As we have seen,
age distributions may be of very different types, especially as between
developed and developing countries. For the whole of these two groups
of countries, distribution by major age groups was estimated as follows
in 1980 (Table 15):

Table 15 Estimates of age structure for the developed and the
developing regions in 1980 (percentages)

Age groups More developed ILess developed
countries countries

0-14 23.0 40.0
15-64 65.6 56.2
65 & + 11.4 3.8

Total 100.0 100.0
II 1

Source: [108]

- 66 -

From this distribution one can derive an index specifically designed to
show the relative weights of the potentially active population (15-64
years) and of the potentially dependent population (under 15 and over
64): the dependency ratio by age (DR).

DR = Population (0-14) + Population (65 and above)
Population (15-64)

This index is usually expressed as the number of "dependents" to
100 "active" persons. With the data in Table 15, we thus have:

for the developed regions:

DR = 23.0 + 11.4 = 0.52

- for the developing regions:

DR = 40.0 + 3.8 = 0.79

i.e., 52 dependents per 100 active persons, and 79 dependents per 100
active persons, respectively.

These figures are, of course, only of an indicative nature, since
people between 15 and 64 are not all active and the others are not all
inactive. This is only a way of describing the age structure by giving
some idea of its possible economic implications. In practice,
developing societies try to arrive at more satisfactory "effective"
dependency ratios by allowing children and the elderly to work (what is
here called "effective" dependency ratio is the ratio of the number of
inactive to active persons of whatever age).

The general population dynamics influence not only the size of
the population of working age, but also its growth rate and replacement
rate, that is, the rate at which its members are replaced by new
entrants. This can be seen from Table 16, which shows how the active
male population increases in a country possessing the demographic
characteristics of developing regions. The level of mortality
corresponds to a life expectancy at birth of 50 (that is, a relatively
high level of mortality). The activity rates by age correspond to the
typical pattern of developing countries (see Chapter 2). The table
shows the rates of entry into the active population (the attainment of
age 15) and of exit from the active population (by death or retirement),
and the increase in the active population, under two hypotheses (medium
and high) regarding fertility.

- 67 -

Table 16 Growth indices of the active population in a developing
country, for two levels of fertility (percent)

S 1 I
Medium High
fertility fertility
(TFR=4) (TFR=6)

Entries 3.0 4.1
Exits 1.6 1.2

Growth 1.4 2.9

Source: [129]

The high fertility hypothesis corresponds to a total fertility of
six children per woman, and the medium fertility hypothesis to four
children. The table shows that a 50 percent higher fertility leads to
twice as rapid a growth of the active population.

It is worth noting that, when population growth begins to slow
down, the growth of the labour force does not decelerate for some 15
years, that is, the time needed for all the generations born before the
beginning of the slowdown to enter the active population. During this
period, the dependency ratio necessarily declines, and this is of course
an economic advantage. During the same period, the demand for jobs
continues to grow, while other needs (such as those for health and
schooling) have begun to decline.


The growth of the agricultural labour force is a function, on the
one hand, of the growth of the total labour force, and, on the other, of
the evolution of the share of agricultural labour in the total.

As we have already noted, the proportion of agricultural labour
is declining almost everywhere in the world. But this is not so for the
numbers of the agricultural labour force. In the developed regions,
where the numbers of the total labour force are more or less stable, the
agricultural labour force is declining. In the developing regions, on
the contrary, the total labour force is still increasing fairly rapidly.
Thus, despite the fall in the proportion of agricultural labour which we
observe in these regions too, the size of the agricultural labour force
continues to grow (see Table 17).

In almost all the developing countries, the agricultural labour
force is on the increase. An important aspect of the long-term
prospects of the agricultural sector, and of employment in general, is
the future of that increase. When can the labour force be expected to
stop growing ? What will its size be at that time ? These questions

- 68 -

Table 17 Active agricultural population: size

active population in

the world and the major

and proportion of total
regions, 1975 and 1983

Size (millions) I Proportion (percent)

1975 1983 1975 1983

World total 820.6 836.7 48.4 43.6

Africa developing 94.9 107.0 72.5 66.9

Far East developing 272.5 291.7 65.4 60.0

Latin America 37.8 39.3 37.3 31.9

Near East developing 33.3 35.1 57.1 50.4

North America developed 3.2 2.5 3.1 2.1

Western Europe 19.5 14.9 12.7 9.3

Eastern Europe and USSR 44.0 35.3 23.9 17.9

Source: [84]

may be clarified by the

calculation of the "turning point" in the

increase in the agricultural labour force; the principle underlying this
calculation is the following:

If F is the size of the total labour force,
A that of the agricultural labour force, and
N that of the non-agricultural labour force,

and if f, a and n are their respective annual rates of growth, it will
be shown 1/ that, if f and n remain constant, the number of years needed
to reach the peak of agricultural labour force numbers (the turning
point) is:

log(f/n) log(N/F)
t =-----------
log(l+n) log(l+f)

1/ See [71], pp. 133-136.

- 69 -

At the end of the t years, the size of the agricultural force
will obviously be:

A = F N = F(+f)t N(l+n)
t t t

Let us give a numerical example, based on FAO data: the case is
that of Egypt. In 1980, the size of that country's labour force was
estimated as follows (in '000):

F = 11,822
A = 5,957
N = 5,865 (i.e., 49.6 percent of F),

and the rates of increase of these three categories during the period
1970-80 were:

f = 0.024
a = 0.016
n = 0.032

The calculation gives:

log (0.024/0.032) log(0.496) 0.413
t = ------------------------------= ----- = 59.0
log (1.032) log (1.024) 0.007

On the basic hypothesis of this model (constancy of 1 and n), the
turning point would be reached in the year 2039. At that date, the size
of the agricultural labour force would be:

At = (11,822 x 1.024 59) (5865 x 1.032 59) = 10,290

which constitutes a relative increase of 73 percent over the 1980

Such a calculation does not constitute a forecast: rather, it
illustrates the consequences of the persistence of the rates of growth
of the total labour force and the non-agricultural labour force, a
hypothesis which may not be realistic in the long term. It is none the
less interesting on that account. For example, one may wonder about the
possibilities of the future agricultural labour force finding productive
employment in the sector, taking account of available cultivable land,
investment requirements, etc. If the absorption of that labour by
agriculture seems problematic, it is to be expected that intersectoral
transfers of labour will come about in any case, probably in the form of
migration from the rural to urban areas, and the consequences of these
moves should be planned for. Lastly, it is possible to study ways of
influencing the future levels of f and n, and the corresponding costs.

- 70 -

In effect, the equations show that:

the time needed to reach the turning point is all the longer,
the higher the proportion of agricultural labour at the
starting point of the projection;

given this initial situation, the turning point will be reached
all the more rapidly, the lower the rate of growth of the total
labour force and the higher the rate of growth of the
non-agricultural labour force.

In a dynamic perspective, it is therefore obvious that a fall in
the growth rate of the total labour force, or an increase in the growth
rate of the non-agricultural labour force, will result in speeding up
the attainment of the turning point. The former development can only
come from a fall in the rate of growth of total population, whether by a
decline in natural increase or by emigration. The second development
may be the result of policies promoting the growth of employment in the
secondary and tertiary sectors, the effect of which on the attainment of
the turning point are more rapid, but which call for careful
consideration costs and of consistency with the overall economic policy.

Before concluding this point, let us illustrate the vast
differences to be found even among developing countries as regards the
agricultural labour force.

Let us take a country like Brazil, where the share of the
agricultural labour force has already fallen relatively low (to 38
percent in 1980). The increase in the non-agricultural labour force is
also much faster there than in the total labour force (4.1 percent a
year against 2.8 percent between 1970 and 1980). The calculation shows
that, if the latter rates remain stable, the turning point will be
reached as early as 1988, with an increase of only 3 percent in the size
of the agricultural labour force over the 1980 level.

Or take a country like Bangladesh, where the share of the
agricultural labour force was still 84 percent in 1980. There, the
total labour force increased by 2.5 percent a year from 1970 to 1980,
against an increase of 3.8 percent for the non-agricultural labour
force. The calculation indicates that, if these rates were to remain
stable, the turning point would not be reached for over a century and
that, at that date, the size of the agricultural labour force would have
increased six times over In such an extreme case, the calculation is
obviously of a purely illustrative nature. It is clear that such an
evolution is impossible, and we are therefore left to wonder what the
future rates (f) and (n) would have to be for the evolution to be
bearable, and above all, how the adjustments will be made.

The reader is invited to calculate the turning point for the
regions in which the agricultural labour force is still on the increase,
on the basis of Table 17. It will thus be confirmed that the turning
point seems to be generally in the offing in Asia and Latin America, and
much further away in Africa, where a large number of countries will
probably not reach that point until the next century.

- 71 -


In every region of the world, transfers of labour between the
agricultural sector and the others are practically all towards the
non-agricultural sectors.

According to the most widely held development theories, these
transfers are to a large extent a precondition (and a sign) of overall
development, for development necessarily involves a shift of the active
population from agriculture which ex hypothesi is traditional and
unproductive to industry which is ex hypothesi modern and dynamic.
This process, it is contended, in turn makes it possible to modernize
agriculture by consolidating land holdings and by mechanization. In the
last few decades, a number of developing countries have based their
national strategies on this postulate. In view of the poor results
obtained, this assumption is more and more strongly contested [28].

The assumption in question corresponds to the idea of a type of
development copied from that of the now industrialized countries. But
no historical experience is automatically applicable to a fresh set of
circumstances. In the present case, several considerations raise
serious doubts regarding the possibility of the process being repeated,
especially at a faster pace as certain arguments would have it. Let us
go over the main considerations in question [1-20].

For one thing, industries in developing countries do not live up
to the role assigned to them for the absorption of the labour leaving
the land. The expansion of these industries is hampered by the shortage
of investment capital, and this handicap is increased by highly capital-
intensive technologies, which are costly to apply and which lead to
industry using much less labour per unit of output than it did in the
take-off phase in the now developed countries.

For another, the problem of labour absorption is not posed in the
same terms, for the present increase in the labour supply (which is a
function of population growth) in developing countries is from two to
four times more rapid than it was at the time when developed countries
became industrialized. Besides, the latter countries had the advantage
of an additional outlet for their supply labour in the form of
emigration to virgin territories, and practically no such areas exist

Lastly, experience shows how difficult it is to develop the
industrial sector without first of all making an effort to develop the
agricultural sector, which affords a livelihood to the majority of the
people in developing countries. The sluggish development of the market
for non-agricultural products, which holds back industry, is a
consequence of the stagnation of the agricultural and tertiary sectors
as well as of the mediocre impact of industry itself as a provider of
income. In the pattern of development triggered by industry, it is
essential to procure investment resources by skimming off the economic
surplus in the agricultural sector. But this policy (effectively
implemented, in particular through the prevailing price systems and the
internal terms of exchange unfavourable to agriculture) has resulted in

- 72 -

stagnation of the level of living, and hence in the demand for
industrial goods in the agricultural population. The development of the
tertiary sector, for its part, has been largely a result of the
urbanization caused by the flight from the land. This is a
substantially unstructured, low-income sector, whose size merely
reflects the setbacks in the other sectors. The inadequate attention
paid to intersectoral relations is one of the greatest weaknesses of
recent development policies: in practice, there has at times been a
regression to a lower degree of integration of the national economies
than was previously experienced. Dualism [127-128], opposing a modern
sector (which incidentally includes large speculative farm holdings) to
a traditional one has, if anything, been strengthened since its theory
was formulated, and what is more, dualism extends into the social sphere
[28]. In these conditions, it is certain that the role of agriculture
in the evolution of the general levels of employment is crucial and will
long remain so [11-129-145].


In general terms, the problem of the major technological options
in agriculture is closely linked to the absolute and relative
availability of the factors of production, especially land and labour

During the last decade, available cultivable area has increased
in almost all developing countries. In the major regions, the changes
in the arable area were as follows: Africa + 12 percent, America + 4
percent, Asia and Oceania + 3 percent. At the same time, the
agricultural labour increased in almost all developing countries.

As regards total population, with extremely few exceptions, it
increased faster than the area available for agriculture. The amount of
land available per inhabitant therefore fell in almost all these
countries. This change obviously made it necessary to step up
productivity, and this increase did in fact take place. However, it may
be observed in this connection that the increase in the productivity per
unit of labour was faster than the increase in soil productivity,
because changes in technology (that is, the combination of the factors
of production) typically involved a relatively decreasing use of labour.
The question raised by these data is that of technological options in
the elaboration of agricultural strategies: for, as the share of
agricultural labour in total population falls, it is also essential to
raise labour productivity if the same degree of self-sufficiency at the
national level is to be maintained.

Now it is not possible to raise the productivity of the soil and
that of labour at the same rate, over the long term. Beyond a given
stage (determined by local conditions), a further rise in productivity
per hectare may call for a more and more intensive use of manpower, and
this will lead to a relative or absolute stagnation in productivity per
worker. Conversely, a further rise in productivity per worker may
entail a move toward an extensive type of production (with
mechanization) which leads to a stagnation of productivity per hectare.

- 73 -

The dilemma must be faced in terms of the relative priority to be
accorded to one or the other aim. In this perspective, it could be
asserted that the productivity of the "scarcest" resource should be
given a higher development priority. This is what happens implicitly in
the most clear-cut demographic situations, i.e., those in which
population density is either very high or very low. But, in intermediate
situations, things look somewhat different. In theory, decisions should
be taken on the basis of the marginal cost of the factors of production,
including land and labour, but also capital. Now, as we shall see, one
of the consequences of many development strategies is to distort the
relative costs of factors by artificially lowering those of capital. For
the rest, it seems that in practice the main technological approaches
take more account of the availability of land than that of labour.

It should be added that the logic underlying this concept
reflects the endogeneity of the labour force variable referred to in the
previous chapter. For example, if the technological choices imply a
demand for labour lower than the available quantity, economic conditions
will lead to a less rapid rise in the volume of this labour (after a
period of adaptation) as a result of intersectoral migrations and
movements prompted by the stagnation or fall in income from agricultural
work. In other words, the supply of labour is regarded as being the
most elastic factor, particularly in view of the adjustments made
possible by migrations. Of course, these adjustments operate in both
directions. A country with unused land and an insufficient labour force
can resort to immigration rather (or more intensively) than to
mechanization in order to develop it.

The drift of the above arguments is to emphasize that, on the
level of overall policies, it is not enough, and it may indeed by wrong,
to regard the labour force only as a factor of production. For
strategies differing in the intensity of their use of labour have
different results in terms of income distribution (inside the rural
areas and between urban and rural areas), and hence in terms of spatial
and intersectoral movements of the population.

Theoretical and practical discussions about the mechanization of
agriculture reflect both the analyses linked to the relative
availability of the factors of production and those connected with
employment and income. We shall examine these below, but only after
having placed them in a larger perspective which takes account of some
additional characteristics of the agricultural sector, which it is
essential to take into account if we are to deal with the problem

The agricultural sector cannot be taken as a homogeneous entity,
either as regards the role of farm holdings in the national economy or
as regards the physical characteristics of these holdings. This means
that we have to examine the problem of the sector's agrarian and
economic structures.

- 74 -

Government agricultural strategies have substantially different
effects on employment, depending on whether they strengthen or otherwise
the dualism within the sector and whether they foster an open economy or
not. In this regard, a distinction has been made [103] between:

a) "Unimodal" strategies. These aim at the gradual
modernization of agriculture from the ground up, and rely on
the general application to the whole of the sector of
technologies with a high labour coefficient. In most cases,
these strategies involve a redistribution of the land, and
also the creation of the agricultural infrastructures needed
for the spread of agricultural development to the mass of the

b) "Bimodal" (or dualistic) strategies. These concentrate on
encouraging the development of a modern commercial subsector,
with a relatively high capital coefficient, which is supposed
to provide the greater part of agricultural production, while
the rest of agriculture remains confined within a traditional
subsector which practically does not benefit from public
sector funds.

These strategies may be either "extrovert" or "introvert" as
regards foreign trade. By combining the two criteria, we obtain a
classification into four types: extrovert unimodal, introvert unimodal,
extrovert bimodal and introvert bimodal [89].

Unimodal strategies may have excellent results for general
development, employment and the standard of living, as is shown by the
example of Japan from the middle of the 19th century on 1/. Unimodal
strategies are mostly extrovert. Bimodal strategies have distinctly
less favourable effects on labour, especially when they are introvert
2/. For these reasons, it has often been recommended that countries
should use policies based on the unimodal pattern in order to increase
labour utilization in agriculture: adjustment of the domestic terms of
trade, suppression of incentives to mechanize, public-sector investment
in irrigation, agricultural research oriented towards the factors
involving intensification and greater access to improved inputs and to
credit [145]. The difficulties in applying this model should certainly
not be underestimated [160], but, for many authors, it is the only
promising one, considering the mediocre results of bimodal strategies in
most of the developing countries.

1/ See [89], pp. 57-78.

2/ The study cited puts forward, as examples of a bimodal strategy, the
Ivory Coast and Mexico (extrovert policy), Colombia, various Central
American countries, Sri Lanka, etc. (introvert policy): see [89],
pp. 13-16.

- 75 -

Introvert strategies have certain specific characteristics, the
effects of which are clearly negative, particularly the maintenance of
an over-valued rate of exchange and of various forms of premium on the
use of capital. These measures were adopted to facilitate import
substitution, and bear witness to the priority accorded to industrial
development. They encourage price distortions, and in particular
distort the comparative costs of capital and labour. In a great many
countries, they have led to the excessive mechanization of agriculture
and to an increase in agricultural underemployment and rural emigration.

Another point about these strategies is that they encourage
highly inegalitarian structures of land tenure. The modern subsector is
mainly composed of large mechanized units employing a small amount of
labour (relative to their capacity for labour absorption). In the
traditional subsector, the scope for the expansion of the holdings is
limited, not to say inexistent, and the transfers of land practically
all benefit only the modern subsector. Inegalitarian structures are
generally unfavourable to the intensification of production and
employment. On large holdings, extensive forms of production which
ensure a sufficient income to the owner (sometimes an absentee) are
preferred. In the subsistence subsector, the small size of the holdings
often limits the surplus and the scope for investment, and leads to
technical stagnation.

A large number of examples of this type of situation may be
observed in Latin America [21], and even in regions where the
concentration of land ownership is distinctly lesser [73]. A
characteristic feature of these situations is that, for both small and
large holders, agriculture becomes a secondary activity, and individual
efforts, investment and progress in productivity concentrate on other
activities. In these circumstances, it is logical that attention should
be increasingly focused on the potential role of land redistribution as
a basic step in rural development [90]. For, in view of the potential
of small holdings (above a minimum size) for labour absorption and of
their usually high productivity, this redistribution would be likely to
increase both yields and employment 1/.

If the concentration of land ownership has institutional and
political origins which can possibly be corrected, it can also be traced
back to demographic origins since the growing pressure of population on
fixed agrarian resources leads to the multiplication of small holdings
[86] which, if they are not viable, are then absorbed by larger ones.
In systems in which the holding is not split up when handed down to the
following generation, it happens more and more frequently that some of
the children have to leave the land, since the holding is not sufficient
to yield a living for a growing family nucleus 2/. In systems in which
the farm is divided every time it is transmitted to the heirs, the

1/ An excellent concrete analysis of this point is to be found in [131].

2/ See, for example, a description of this mechanism in a West African
context [132].

- 76 -

process is bound, at a certain point, to come up against its limits
because of the existence of fixed costs and of the indivisibility of
certain factors of production [133]. When costs thereby become too
high, and the decline in the size of the holdings ceases (for example,
as a result of a modification in the inheritance system), the former
process applies and a class of landless labourers appears 1/. In the
face of such an increasing divergence in land distribution (the
proliferation of tiny holdings on the one hand and the extension of
large holdings on the other), it is difficult to frame and to apply,
without encountering constraints, technological choices adapted to the
sole economic and social objectives.

The central problem here is obviously that of mechanization, and
more specifically its extent and form. The question should be tackled
in the context of the present state of land and labour resources.
Unutilized areas are becoming increasingly scarce. Where they exist,
they are more and more difficult to reach and expensive to open up. In
these circumstances, the growth of demand will lead to increasing
pressure to raise yields per unit of area. But in that case, the
requirements for labour generally increase less rapidly than the
production. It may even happen that, with certain types of
mechanization, there will be a fall in the need for labour [11]. Now
there are often grounds for seeing to it that the trend towards
techniques involving a higher capital/labour ratio does not imply a
substitution of capital for labour. In fact, certain technical changes
increase the need for labour per unit of area, either because they
entail additional activities (for example, the application of
fertilizers), or because they make it possible to practice multiple
cropping [1]. It is therefore desirable, if the aim is to preserve a
certain coefficient of labour absorption by the agricultural sector, to
be selective in mechanizing activities 2/. It is also desirable in this
case to steer a course towards intermediate technologies which make
moderate use of physical inputs, but call for a large amount of labour
per unit of area and, in spite of this, have decent (and indeed high)
yields. This choice implies foregoing techniques which would involve
the largest increases in yield, but would reduce labour requirements per
unit of output, and would therefore be relatively unfavourable in terms
of income distribution (and hence of the widening of markets and
intersectoral trade).

Ninety-nine percent of agricultural jobs are "self-created" [16]
by the peasants themselves, who combine land (provided it is available)
and labour primarily in order to satisfy their requirements for domestic
consumption. The problem of the "creation" of agricultural jobs is
therefore, strictly speaking, extremely limited; the problem of avoiding
the destruction of existing jobs, on the contrary, is definitely a real

1/ The process has been described in various national contributions to
-FAO/UNFPA seminars on as widely different countries as Rwanda, Mexico
and Bangladesh. See, for example [58].

2/ Thus, the mechanization of most of the operations immediately
downstream from planting and harvesting displaces labour and ought to
be confined to the easing of seasonal bottlenecks, except in cases
where it would make it possible to reap several crops a year [11].





Migration is certainly the most highly visible aspect of
population dynamics in the short term. Population growth, and even more
the changes in structure, are perceptible only in the long run. But
migratory movements lead to rapid modifications in growth rates and in
population structures, and have relatively immediate and visible
economic effects. For the agricultural planner, there is food for
thought in the study of the causes and characteristics of rural
migration, its repercussions on agricultural development and the
opposition between spontaneous and planned migration.


The causes of rural migration have been examined in a number of
studies, but these have been of small help to planners. We shall
rapidly review the state of knowledge on this questions, and shall
dwell, not so much on the empirical results (which are often trite), as
on the successive approaches used, as these have instructive links with
the way in which development policies evolved. I/

Attempts have often been made to analyze migration as a response
to "push" and "pull" factors which would characterize the areas of
departure and arrival [134]. This approach has been adopted primarily
in connection with rural-urban migration, which has long attracted
considerable attention. Writers have noted, on the one hand, the
underemployment and the poverty of rural areas, the lack of
infrastructure and services, and the pressure of population on the land,
etc., and, on the other, the greater employment opportunities in the
towns, the better range of educational facilities, the existence of
infrastructures and services, the greater variety in social life, etc.

The separation between push and pull factors is, at bottom,
artificial. These are simply two symmetrical ways of describing one and
the same order of phenomena, that is, the inequalities (or differences,
or development gaps) between urban and rural areas. This point is well
illustrated by the following question: do migrants move because of the
limited job prospects in rural areas or because of the better job
prospects in the urban ones ?... which shows that push and pull are one
and the same thing [92].

1/ With few exceptions, we refer in this chapter to migrations of the
active population.

2/ A variant of this classification is the one which distinguishes
between "village-based" and "urban-related" factors: the former
include in particular the scarcity of land, its relative degree of
concentration, access to it and its productivity [135].

- 80 -

Various authors have engaged in the construction of explanatory
models of individual migration behaviour, emphasizing mainly the
relation between the costs and benefits of the movements 1/. But the
correctness of these models has not always been confirmed by empirical
studies, and some of their hypotheses (especially the homogeneity of the
migrants as regards skills, the completeness of their information about
the town and the higher levels of income in the formal sector) have
proved to be of dubious validity. These models have also been
criticized on two more grounds:

The absence of non-economic factors among the "explanations"
of the decision to migrate: differences in the opportunities
of education, the environment, the quality of life, the risks
linked to the change in social environment, and the role of
family and social relationships;

The emphasis on urban-rural disparities and the neglect of
intra-rural ones 2/: too little attention has been paid by
theory to intra-rural migration despite the fact that it often
is greater than migration to the towns especially in Asia
[137], but also in Africa [62] and elsewhere.

It is also true that, not enough attention has been paid to the
less easily quantifiable factors (land-ownership structures, income
structures), to social factors (the distribution of power and of roles
in organizing production and social life) and to the effects of
government policies.

Moreover, it has been noticed that it is difficult to analyze
migration as only the result of individual logic which was the case
with economic theory until the appearance and development of the
"household economy". The integration of households' demographic
dynamics in migration studies is an important contribution to the
understanding of the phenomenon, and indeed of any other examined at the
microeconomic level. Whereas changes are relatively gradual in the
size, sex and age structures, and dependency ratio of a population, at
the household level these variables are, by their very nature, subject
to rapid changes.

Thus, the behaviour of a household which allocates the labour
resources at its disposal to food production on the one hand and to wage
labour (through emigration) on the other, has been analyzed in the
context of Southern Africa [78]. The analysis shows that this alloca-

1/ A classical illustration of this approach is Todaro's model, the
first one to take account of the probability of the migrant finding
work in the town [136].

2/ See [92], pp. 11-13.

- 81 -

tion depends, not only on the relative profitability of activities and
on opportunity costs, but also on the composition of the family. The
key variables are the ratio of the number of inactive persons to that of
producers, and the size of the group of producers which determines the
extent to which the diversification of activities is possible.

It is probable in the end that emigration is induced not so much
by other factors as by the need for cash [138], and, more generally
speaking, by the need for rural households to diversify their sources of
income in order to increase economic security, or at least to reduce

We shall make two further remarks to round off the discussion on
the causes of migration a subject of great practical importance, since
migrations are often the litmus test of economic disequilibria calling
for remedial measures, and any policy designed to modify migration flows
ought to examine their underlying causes.

It is necessary in the first place to examine migrations at the
household level (for a strictly economic analysis) and even at community
level (villages and ethnic groups) in order to obtain a broader
understanding of the phenomenon. Certain migration flows which are
sometimes of great economic importance can only be explained within an
analysis of the social systems, and in particular of the relations
between generations and of the distribution of socioeconomic roles and
power [17-115].

Lastly, if we are fairly clear about the reasons why people move
1/, that knowledge is by itself of limited usefulness. A real
explanation of migrations ought of necessity to be based on an analysis
of the factors governing the distribution of resources and of political
and administrative power, as well as the diffusion of information [137].
In terms of policy formulation, the two approaches yield very different
results. In particular, if one aims at the stabilization of the rural
population, one cannot avoid examining the problems of stable access to
resources and to commodities (distribution of land ownership, security
of tenure, but also food security, for example), or reassessing the
explicit or implicit technological choices made by the institutions
providing assistance to rural development.


In this section, we shall deal with the effects of migration on
the rural environment itself another subject which has been relatively
neglected because of the well-known "urban bias" which has left its mark
on studies of migrations and also on policies.

1/ Incidentally, we have not learned anything very new on this point
over the last 20 years, and the classical studies are still largely
valid [139].

- 82 -

The effects depend to a large degree on the migrants'
characteristics. As is well known, migration is a selective phenomenon:
most migrants are young; men migrate more than women (except in Latin
America, where female emigration is standard, at least in the flow from
rural to urban areas 1/). As regards the migrants' level of education,
it should be noted that it is often higher than that of the
non-migrants, especially in migration to the towns. It can easily be
shown that, from the rural family's standpoint, it is the young men with
a certain level of education who offer the highest ratio of possible
contributions to income through their emigration, to the drop in
earnings or in production caused by their departure.

Having made this point, let us look at the effects of migratory
movements on the departure areas. We should at the outset make it clear
that there can be no question of providing a constantly valid model of
these effects, and even less of suggesting an overall judgement of the
beneficial or harmful nature of emigration. We try to draw up a
reasonably complete list of possible effects on employment, production,
income, investment and other variables of importance to development. In
practice, the planner's task will be to identify the implications of
existing migratory flows and also of those which emerge in the areas of
the projects for which he is responsible, and to take account of them in
the preparation and implementation of these projects. An important
consideration, which should be borne in mind in this context, is the
great variety of migratory movements as regards their range (distance),
duration, precise economic aim and the category of persons (age and sex)
displaced by them.

Let us look in the first place at the effects on production. A
distinction must be made between the immediate effects due to the
departure of the workers and the longer term effects due to changes in
the system of production.

As regards the short-term effect of production in the departure
areas, dualistic theories assert that it will be nil, since, because of
the presence of surplus labour, tasks are divided up, and the same
production could be guaranteed by a smaller number of workers. But,
with very few exceptions, the empirical studies carried out in a large
number of countries invalidate the hypothesis of a zero marginal
productivity of labour in the emigration areas. On the contrary, it
would seem that the allocation of labour and resources is usually
efficient and that the withdrawal of labour, if it does not take place
in the off-season, will call for compensatory measures: changes in the
crop mix with a shift towards those requiring less labour; technological
changes using labour-saving equipment; a greater participation of
occasional workers, especially of women; and greater recourse to outside
wage labour [92].

1/ Other exceptions (at the national level) are the Republic of Korea
and the Philippines [137]. In addition, in a not inconsiderable
number of countries, migrations to the capital are mainly female.

- 83 -

In any case, there is little doubt that in the short term the
impact of emigration on the level of production is lesser than on the
organization of production. A number of field studies have confirmed
the increase in the participation of women in agricultural production
subsequent to emigration. Children's work (very often linked to women's
work, which it supports) also increases. Migration thus modifies the
traditional division of labour and increases women's responsibilities.
This factor should be taken account of in project preparation and in
activities affecting the emigration areas.

In the longer run, emigration can influence the level of
production through various mechanisms. For example, a fall in
production may be experienced by a population whose physical capacity
has been reduced because it must rely more on women, children and
elderly people; who has to meet smaller local subsistence needs; and who
is without ready access to the technical and financial means required
for intensification.

Other effects are on the type of production. In most cases, food
crops are paid more attention than cash crops, for example because the
latter generally require more work. However, in certain groundnut- and
cotton-growing regions in West Africa, where it was difficult to modify
the priority given to export crops, emigration seems to have affected
food production and in particular to have played a role in triggering
famines in the 70s.

The allocation of income derived from emigration is also an
important factor in the evaluation of its economic repercussions. Some
have seen emigration as a key element in a family's strategy of
agricultural intensification [68]. Admittedly, it is possible to find
examples of groups making fairly constant use of income from emigration
in order to improve local agriculture. The Soninke of West Africa are a
good example [62]. But, in many cases, this income is used:

as in Southern Africa [75-76-77] to buy cattle, which is a
form of hoarding and of accumulating prestige rather than an
investment (with, it should be mentioned, resulting ecological

or for the construction or betterment of housing;

or for ostentatious forms of consumption which are practically
imposed by the social rules and which sterilize migration from
the economic point of view 1/.

One positive effect of migration might be that of making possible
the purchase of land by families who were badly off in this respect and
who are fairly often the ones taking part in migratory movements. In
practice, a phenomenon of this kind (which incidentally

1/ For example, in the Mossi system [17].

- 84 -

on those areas and systems where land can be purchased) is met with on
relatively few occasions. On the one hand, migrants' families are short
of labour during the time covered by the migration, and this prevents
them from expanding their farm (unless they can resort to wage labour,
which, added to a hypothetical purchase of land, would call for
considerable sums of money). Moreover the disinclination for
agriculture appears to be general in emigration areas. For example, a
very small fraction of Sudanese emigrants invest in agriculture [54].
In Tunisia, less than a tenth of the income from emigration goes into
agriculture, and not necessarily into improving it [140]. In fact, the
dependence of emigration areas on income flows from emigrants usually
leads to profound social transformations (more profound than those in
the systems of production). We can witness the emergence of societies
in which agriculture becomes simply one activity among others, often
stagnant and sometimes of secondary importance [29-30-74]. In the most
urbanized countries, migration may sharply accelerate the process of
"de-ruralization" and the siphoning off of the active population into
the tertiary sector [29-32].

The effect of emigration on employment in the strict sense of the
word is anything but simple. In addition to the obvious relief procured
as regards unemployment and underemployment, we can often note the
appearance of labour shortages in peak periods (during which the labour
which has emigrated would normally have been mobilized) and even at
other times. If we disregard the case of seasonal migrations during the
dry season to better watered areas, which effect an adjustment between
complementary ecological zones [62], we shall note that the emigration
areas frequently suffer from unmet labour requirements, and this may
compromise development projects elaborated without sufficient regard for
these phenomena [92]. In this way, the initial impact of migration may
have fairly distant repercussions. This is one more reason for not
indulging in hasty judgements on the costs and benefits of migrations
and for relating any evaluation on this point to a clearly specified
reference population: who benefits or derives an advantage ? and who
bears the costs or drawbacks ? [31].


Since migrants move in search of economic advantages, whether
certain or anticipated, migration appears to stimulate a higher level of
production and employment, and in any case to encourage a smoother
functioning of the economic system. This point of view has been the
subject of various theoretical analyses, but it is really open to

On the intrarural scale, migrations may be seen as a process
tending to ensure a balance between the reproduction of labour force and
the regional labour markets, since they reduce the "overpopulation"
created, in some systems of production, on the labour markets, and
satisfy the needs of other markets, thus reducing unemployment and
underemployment. But, in practice, the characteristics of migrants are
such that they will not necessarily be absorbed by markets where labour
is in short supply. It therefore often happens that migration does not
lead to the expected equilibrium of the labour markets and does not
reduce overall rural underemployment [19].

- 85 -

On the national scale, dualistic theories regard migration as a
mechanism which eventually achieves an equilibrium between the
traditional sector (or "subsistence" sector) and the modern (or
"capitalist") sector. In this perspective, the subsistence sector is
characterized by marked underemployment. The marginal productivity of
labour in it is nil, or almost nil, and wages are equal to the
maintenance cost of the labour force. Migration towards the capitalist
sector, induced by higher wages, reduces undermployment in the
subsistence sector, and thus tends to raise wages in that sector until
wages are equal in both sectors.

The effect of this mechanism on agricultural development,
according to these theories, is threefold. In the first place, per
caput output and income in the agricultural sector increase if
emigration reduces the rate of population growth below the rate of
capital formation. Secondly, the terms of trade evolve favourably for
the agricultural sector because of the tendency for the value of its
products to rise relatively to those of the other sectors (a tendency
induced by the evolution of labour costs described above). In the third
place, the conditions are created for the introduction of advanced
techniques and a rise in productivity in agriculture. The total effect
of migrations is therefore to bring into line not only the allocation of
manpower and labour income, but also productivity.

The weaknesses of this model emerge fairly clearly. Several of
its hypotheses do not correspond to the real situation, especially as
regards the zero marginal productivity of labour and the rapid expansion
of the capitalist sector. Given the slow growth of industry in most
developing countries, migration has mainly transferred underemployment
from one sector to another. The gap between urban and rural wages has
been reduced in certain cases, but has increased in others. As to the
terms of trade, it really seems that, with very few exceptions, they
have been agriculture. As against this, it is true that migration often
has helped to bring rural subsistence economies closer to the point when
the sectors are linked by sustained exchanges, which is the first stage
on the road to the disappearance of dualism itself. Migrations have
also contributed to bringing about a certain diversification of
employment in rural areas [92]. But, in terms of technical development,
productivity and standards of living, it may be said that dualism is
still very much alive.


One factor making for migration is population pressure on rural
and agricultural resources (in effect, on land). This factor becomes
decisive in situations where the pressure is particularly intense. When
certain areas in the same country are sparsely populated, while others
are overloaded, the possibility emerges of transfers of population from
the latter to the former.

- 86 -

These disequilibria often trigger spontaneous migration,
sometimes on a mass scale, the economic, ecological, social, and indeed
political consequences of which may be profound for the destination
areas. In certain countries, these disequilibria have given rise to
policies for the fairly large-scale displacement of populations with a
view to relieving pressure on the overpopulated areas and to opening up
insufficiently farmed areas. The importance of these moves, whether
spontaneous or planned, derives both:

from matters of principle, for they raise unambiguously all
the problems of the relations between the population on the
one hand and space and resources on the other, and they force
us to realize that spatial planning, especially in the rural
environment, essentially belongs with problems of human

from practical considerations, since observation shows the
increasing extent of intra-rural migrations (including
long-term ones) and also the increasing willingness of
governments to carry out programmes of settlement, despite the
high costs involved. 1/

A wide-ranging debate has developed regarding the respective
advantages and disadvantages of leaving settlement to the peasants to
carry out and of having settlement schemes initiated and controlled by
the State 2/. We give below a summary of the relevant arguments.

Spontaneous migrants, if they achieve a production of foodstuffs
(and possibly also of cash crops once they have ensured their food
security) greater than the marginal level which they would obtain on
their overloaded original plots, may procure for the community a net
gain in foodstuffs, and indeed in foreign exchange. This positive
result is all the more visible for the State since no preliminary
investment is involved.

On the other hand, this form of settlement is not without
drawbacks, especially the accelerated depletion of the ecological
capital, the lack of future security for the settlers, and conflicts
between immigrants and the original population:

in the absence of any concern for the maintenance of the
natural capital (which the migrants are ready to abandon as
soon as it is exhausted, in order to settle elsewhere), one
often witnesses the overexploration and the depletion of the
soil and the forest; the ecological danger is felt everywhere
in countries where agricultural migration is sufficiently
intense 3/;

1/ We shall come back at a later stage to the question of the cost of
these operations, which has perhaps given rise to somewhat hasty
2/ See the case of Burkina Faso [41] which provides an excellent
illustration of most of the considerations bearing on this point.
3/ For example, several Sahelian countries, Ethiopia, Mexico, Thailand,

- 87 -

in addition to the consequences of the precarious ecological
situation, the need to adapt to the immediate market
conditions subjects the migrants to considerable pressure
from economic hazards (falls in producer prices increases in
the cost of the factors of production, etc.);

there are many possible sources of conflicts between settlers
and local communities extending from the squandering of the
agrarian patrimony by the former to rivalries over economic
and commercial influence in cases where the groups of
migrants are particularly enterprising; recent history shows
that these conflicts may assume violent forms.

Planned settlement also makes possible overall increases in
production, but at the cost of substantial investment. However, this
strategy is clearly superior as regards the protection of the soil and
ecological capital. It often provides an infrastructure which makes it
easier to settle the populations on a stable basis. On the other hand,
it often encounters considerable technical and management difficulties.
The lack of cadres often makes it more difficult to carry out the
centralized management of multifaceted assistance schemes. The land
development and farming model, usually elaborated and imposed from
outside, often encounters resistance from the population.

Accordingly, planned settlement has been subject to attacks by
economists and administrators, who stress the high cost of this
approach. These two groups, it should be added, are supported in this
case by the sociologists, who undescore the restrictive nature of the
technocratic patterns of planned settlement and oppose it to the freedom
of the spontaneous migrant. But, if we wish to avoid hasty judgements,
we must assume suitable time, spatial and social scales of reference

In the short run, spontaneous migration, which does not involve
public-sector cost, presents undoubted advantages. But, in the medium
and long-term, the populations settled in this way always call for
intervention by the States in order to provide the same infrastructure
as that enjoyed by the nation as a whole and to ensure the conservation
of the environment. Public-sector expenditure is merely put off to a
later date, and is probably rendered more onerous, since spontaneous
migrants tend to spread to a larger area the negative aspects of the
farming system which forced them to abandon their land.
In that perspective, suitably planned settlement can preserve the natural
capital which constitutes an advantage beyond price and ensure that
the populations, once installed, will be permanently settled.

More generally speaking, the cost and profitability of settlement
operations ought to be set against the cost and profitability of
alternative solutions, at the stage of project identification as well as
at that of evaluation. Given the existence of overburdened lands and
the estimated costs of transfer of part of the populations involved, one

- 88 -

should wonder:

What would be the cost of maintaining (if that were possible)
these populations on the spot, since such a step would call
for an effort to intensify production and raise it to the
appropriate level, and also no doubt entail considerable
efforts to restore and maintain soil fertility ?

what would be the deferred cost, in terms of infrastructure
and of ecological damage, of the uncontrolled settlement of
other areas by these populations ?

what would be the cost of their reception in urban areas a
solution which is open to spontaneous migrants and which
would make them greater consumers of infrastructures and
public services ?

It is doubtless true that settlement projects have, as a whole, a
lower profitability than projects for agricultural intensification 1/,
but in this context we ought at least to ask what would have been the
profitability of intensification projects carried out in the area of
origin of the migrants. This is the only valid comparison, at least
when the settlement operation aims at relieving population pressure on
specified areas. When settlers are brought in from various regions and
the main aim is the opening up of new lands, account must be taken, not
only of the long-term profitability of the investment entailed, but also
of political interest (as in the occupation of frontier areas) and of
the economic advantages of a better integration of the national space.

It is certain that considerations which are not exclusively
economic are often extremely important in the identification,
formulation and evaluation of such programmes. It is none the less
important that the relevant economic studies should be complete and
unbiased, which seems to call for serious methodological efforts.

Among the non-economic considerations involved in planning
settlement operations, we may mention several highly important
demographic aspects.

The population dynamics of a population of settlers is a peculiar
one, for the structures and characteristics of the households
transplanted are usually also of a particular nature. Whether this is
because of the principles guiding the selection of settlers (by the
agency directing the settlement operations) or because of a greater
propensity to migrate, the households concerned are usually young and in
good health, and they have relatively few children at the moment of
migrating. There are generally very few elderly persons in the
population of a settlement area. The age structure of the population is
therefore particular, concentrated as it is in adult age groups, with

1/ The World Bank has estimated at 16 percent the average internal rate
of return of 28 settlement projects it supported, against 25 percent
for a group of 16 intensification projects [141].

- 89 -

some young children, but relatively few adolescents and very few old
people 1/. With the passage of time, this structure evolves, but there
is scant likelihood of its coming back spontaneously to balance. The
result is the emergence of structures and trends in needs, which are
also rather special. The needs in health and education induced by the
children, for example, may be on a relatively reduced scale at the
beginning of the operation, but they will rise to fairly high levels
after a few years. Later on, it may well be that these needs show a
rather rapid decline (as the children reach adulthood).

Moreover, the fertility and mortality levels of a population of
settlers are usually very different from those of a demographically
balanced population. In the former, mortality is low because of the
predominance of adult age groups with a low death rate. Fertility is
high because of the predominance of relatively young couples. Overall,
the rate of natural increase is very high.

Now the dynamics of the populations in question is all too rarely
taken into account at the planning stage of these projects. The size of
the farms is generally fixed by taking account (in addition to the
technical parameters linked to the quality of soil) of an average size
of immigrant family. But the future of family farms, in the context of
the growth of the households, is rarely envisaged explicitly. It may be
that a settlement operation sets as its aim the "fixing" of a population
of a given dimension, but never, it would appear, that of fixing a given
number of families throughtime. The problem is essentially that of land
allocation. There is a danger of planning for too small a farm size, as
we have seen. But there also is a danger of planning for too large a
size since the families do not, in the beginning, have enough manpower
to ensure the optimum exploitation of the land assigned to them.

It is clear that the difficulty is to obtain, in a planned
settlement scheme, the flexible adjustment mechanisms typical of
spontaneous settlement. But it will be observed that this type of
settlement, even more than planned migration, is subject to the risk of
oversizing as the tendency to grab the maximum surface may get out of

In so far as the aim is to achieve a durable transfer of
population, and hence to provide a basis for the long-term subsistence
of the families, one should underline the importance of avoiding too
dense a population, of adopting a policy and rules for the transmission
of the farms, and to adapt their area accordingly. It is also necessary

1/ There is obviously an even greater demographic disequilibrium in the
population of settlers composed mostly of single males. In practice,
promoting the mistake of such operations is now rarely committed.
For, even on a temporary basis, this type of human settlement gives
poor results in terms of the stability of the populations.

- 90 -

this context to give timely thought to the legal conditions for the
outright purchase by the settlers of the lands and to protect them
against the risks of being ousted from them [24].

Another imperative for planning in this field is to ensure the
socio-economic balance of the populations transplanted. We refer here
in particular to ethnic and socio-professional characteristics. In
these two respects also, experience shows that populations suffering
from disequilibria cannot maintain and stabilize themselves as social
bodies functioning harmoniously in their environment. The Indonesian
experience, for example, has shown that establishments composed
exclusively of farmers suffered far more setbacks than the settlements
in which a whole village agreed to move.


Generally speaking, the idea ought to be accepted that sound
agricultural planning demands a good knowledge of migratory movements in
all their diversity, and a correct awareness of the extent to which
these movements can affect development plans and projects through
manpower structures and dynamics.

It is obvious, for example, that labour demand and supply balance
sheets ought not only to be seasonal in order to take account of the
variability of factors in time, but also regionalized in order to take
acccount of the variability in space and in particular of labour force
transfers [10].

In addition, if we refer, not to the general planning process but
to the study of specified types of agricultural polices and projects, or
of particular problems of rural development, we shall see that the
importance of manpower structures and trends, and of manpower use and
mobility, is clearly a vital factor in a fair number of cases, such as
the study of:

the advantages and feasibility of a policy of expansion of the
cultivated areas;

the advantages and feasibility of a policy of intensification
of agricultural production;

the advantages and feasibility of programmes providing for
labour-intensive rural infrastructure schemes;

policies of diversification of economic activities in rural

the evolution of the agrarian base (splitting up of farms,
changes in property and customary rights, land ownership
concentration and agrarian reforms);

the protection of natural resources;

the orientation or limitation of migratory flows (particularly
from the rural to the urban areas); and

- 91 -

technological choices, both in the context of project studies
and of agricultural planning in general [47].

In the case of the preparation of agricultural investment
projects, a sound study of the patterns of participation in economic
activity and of the socio-demographic phenomena influencing them is
needed at the formulation stage, but also very often in the follow-up.
In the case of small populations, such as those of project areas, the
patterns of activity and migration may be fairly diversified. It is not
possible to apply standard hypotheses to them a priori: direct
knowledge of the facts is essential. Moreover, ignoring characteristics
such as the migratory habits of the populations involved (and the
collective strategies which they relect) may easily result in the
estimates and hypotheses on employment effected during project
preparation becoming out of date and inappropriate [32].

In this context, the relevant variables appear to be:

the sex and age structures of the population;

the sex and age-specific activity rates, including those for
children and the elderly;

the patterns of time use for the main categories of manpower,
with, however, particular emphasis on the economic
contributions of women and children;

the division of labour between these categories for the main

the patterns of labour substitution between categories;

the patterns of participation, whether full-time or otherwise,
in non-agricultural activity; and

the patterns of migration by category of labour involved, the
spatial considerations (destination of the emigrants, origin
of the immigrants, and distance) as well as the time
dimensions (seasonal patterns and duration) [47].

These are fundamental data for any forecast, whether formalized
or not, of the available manpower and of its characteristics. For
several of these variables, quantitative data are rarely available on
the desired scale (although it is always possible to collect them by
well-tried methods), but it is indispensable to obtain at least
qualitative data (which indicate the direction of the phenomena, if not
their magnitude). Failure to take account of a variable because it
cannot be quantified exactly amounts to giving it a zero value. Such a
solution is very rarely the most satisfactory one.

92 -

If these considerations are carefully taken into account in
project preparation, the corollary will be that they should be kept
under watch during the life of these projects. Although they are still
insufficiently incorporated into project monitoring systems,
socio-demographic variables should on the contrary be full components of
those systems.

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

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