UNITED NATIONS
ECONOMIC AND SOCIAL COUNCIL
Distr.
LIMITED
E/CIM4/SH/36 15 May 19G1
Original: ENGLISH
ECOMOKIC COMMISSION FOR AFRICA
Working Group on Organization, Content and Methodology of Household Surveys Addis Ababa, 29 June - 3 July 1981
FOOD PROCUCTION/AGPICULTURAL SURVEYS
Prepared by
the Food and Agriculture Organization of the United Nations
E/CM.14/SK/36
Introduction
1. The National Household Survey Capability Programme (i;KSGF) sponsored by the United Nations, fl^Panc the Fcrld Dank has as ,its basic aim "to make optimum use of all available resources to tui-ltj. up
an effective and enduring statistical infrastructure ir the area of household statistics, with a resulting improvement of related national,
statistical services". At the sane tine 'the IIHSCP is also a programme for data collection.../ (1)
2. The majority of the households in rur?.l areas of developing;
countries derive their livelihood fro™, agricultural activities. Hence, household surveys can be efficiently used to collect food and agri
cultural statistics relating to the traditional sector of agriculture especially within the framework of a national household survey
programme
3. The first section ' explains -the need for collecting food and agri cultural statistics ar.d their pain uses, with particular reference made to the reconx.endctions of irternrtional bodies. The second
section deals witl the lasic concepts of agricultural holding and household. A list of items relating to food and agriculture, which can be collected through household surveys- is proposed in section 3.
The definition of the main'proposed itei;.s are given in section 4.
Different sources of data on agriculture are descriled in section 5.
A brief description of country practices ir. section C illustrates the different ways of using households as units for collecting statistics on food and agriculture. Sample Questionnaires are attached as
Appendix 1.
SECTION 1
Meeds and uses of data on agriculture
4. During the past decade there has leen an ever increasing demand for tifnely and reliable statistics on food and- agriculture, especially in developing countries where the majority of people derive their
livelihood froir agriculture.
5. Gn the national level data are needed for planning of social and economic jttevelopment. Meaningful plans can be developed only on the basis o.r relevant, timely and accurate data covering a wide range of socio-economic and demographic characteristics of the population.
6. Data are also needed for different .regions in the country, enabling a planned and balanced development of the regions.
IIOTE: The numbers in parenthesis refer to the list of references>
Appendix Z.
.J4/SK/26 Page 2
7. There is also need for statistics on special population groups, such as landless agricultural labourers, women, children, etc. These statistics are of particular value in the design and control of
policies and programmes dealing with issues such as poverty, unemploy
ment , malnutrition, etc.
£. In monitoring and evaluating specific projects, data are needed not only for the project area itself, but also for the broader region where the project is situated, in order to be able to compare the results in the project area with those outside the area.
Recommendations of international bodies
9. The need for statistics on food and agriculture was recognized
also by the FAO Governing Todies.
'''Governments of the Member Nations of the Food and Agriculture Organization of the United Nations have recognised the great
importance of food and agricultural statistics in carrying out national and international programmes aimed at fulfilling the main objectives of the Organization - namely, to raise nutrition
levels and living standards to improve the efficiency of agri cultural production and the distribution of food and agricultural products, and to Vetter the overall conditions of rural popula tions. Periodic agricultural censuses are particularly important,
as they are the main source of basic quantitative information on structure and other characteristics of agriculture that is needed in development planning, socio-economic policy formulation ana establishment of national priorities." (3)
10. The World Conference on Agrarian Reform and nural Development (V.'CARRD), held in Roire in 1979, recommended that :;FAO and other organizations of the UN system, in cooperation with Member Nations, develop indicators of agrarian reform and rural development and help collect and analyse pertinent data in order to monitor progress toward respective national targets of rurp.l development as laid down in this
Programme of Action. "■■ (4)
SECTION 2
11. The concepts and definitions of holding and households are of
primary importance to the use of household surveys for collectingagricultural statistics. A detailed discussion on them may be founa
in references (2) and (3). . .,.!..
Holding
12. 'A holding, for agricultural census purposes, is » techno-econoiic unit of agricultural production comprising all livestock kept and all
land used wholly or partly for agricultural purposes and operated under
the management of one person or more, without regarc to title, legal form, size or location. The holding as a techno-econoicic unit unr single management generally has the same means of production, suchlabour, farm structure, machinery or draught animals.
a as
14/SU/Z6 Page 3
"The above definition covers practically all holdings engaged in agricultural production and includes livestock holdings with practically no significant agricultural land or for which
land is not an indispensable element of production.:: (3)
13. As follows froFi the above, a holding is a techno-economic unit.
In order to be able to collect information about the holding it is necessary to identify a.id contact the person who operates it, i.e.s the holder. There are different ways of identifying the person(s) who operate(s) the holding.
14. One way is in first identifying the land for which information has to be collected. This can be done on the basis of geographic
maps or aerial photographs. The second step is then to find out who operates this land. This approach is used in some countries for area samples (e.p.s United States3 Tunisia, Sierra Leone). It should be noted, however, that sone agricultural information en be obtained even without contacting tho holder. For example, forecasting of crop yields or mer.suring of crop areas can be carried out by using a crop field as a unit, and if no other information is needed, there is no need to contact the holder. However, for crop cutting to estimate the crop yield the holder's permission is usually sought.
15. A much nore widespread method of identifying the holding in the
"traditional1- sector of agriculture is by approaching the households.
Lists oi households are prepared or are already available from a previous household survey or a population census. Adding a few more simple screening questions for each household, such as whether any member of the household operates some land or keeps some livestock,, a list of households with agricultural holdings can be obtained. This approach is used in the majority of developing countries.
16. It follow s from the ctefini tion that, holdings comprise all live stock kept and all land used wholly or partly for agricultural purposes .! wevor, some countries, for practical or operational reasonss set a minimun r.vraber cf livestock kept or a minimum size in terms of area operated, or value or volume of production as a lower limit. Holdings belou these l:.uits are not covered in the censuses and surveys.
Although the volu::ic or value of production of such holdings may be of marginal importance to the total agricultural output of the country^
by this restriction a sometimes significant proportion of small holders is missing in the census or survey results.
17. The FAO Progra;..;>ie for the 19C0 V'orld Census of Agriculture recommends (2, p.,17) that the minimum size limit be set as low as possible and, where.the holdings lolow the specified limits are
important, to investigate then through special inquiries. Information on small holders becomes very important in monitoring and evaluating rural development, as was stressed ly the Vforld Conference on
Agrarian P.e^orir. and Rural Development (tfCARRD) . (4)
]■■»,-■
.-,, .- - > ..
Page 4
Holder
13 According to the generally accepted definition proposed by FAO,
■■■the holder is a person who exercises management control over the operations of the agricultural holding and takes the major decisions regarding the utilization of the available resources.' (u)
19 Fhen two or more members of'the same household jointly share the operation of a holding, usually the head of the household is considered
as a holder. On the other hand, when two or more members of the house hold operate independently some pieces of land, each of them should be considered as holder and their land counted as a holding.2C In many developing countries there are situations where it is difficult to decide on the basis of these definitions who is the holder and which land is a separate holding.
21 A common case is found in some African countries where the wife of the head of household cultivates a piece of land on her own. This land may be a small vegetable garden and sometimes a plot with some cash-crop. It aay be difficult to decide whether the wife should be considered as an independent holder and therefore the piece of land as a separate holding, because the degree of the wife's independence
in operating the plot varies fror: case to case.
Household
22. The following definition of household is given in the FAC Pro gramme for the 19S0 rorld Census; of Agriculture (VXA) (3, p. 22):
"The concept of 'household' is based on the arrangements made cy persons, individually or in groups, for providing themselves with food or other essentials for living. A household may be either.
(a) a one^person household, that is, a person who makes provision for his own food or other essentials for living without combining with any other person to form part of a multi-person household, or
ft) a multi-person household, that is, a group of two or more persons who make common provision for food or other essentials for living.
The persons in the group may pool their incomes ana have a common
budget to a greater cr lesser extent;; they may be related or unrelated
persons, or a combination of *:oth.
23. -Households usually occupy the whole, part of, or more than one
housing unit, but they may also be found living in camps in boarding
houses or hotels, or as administrative personnel in institutions or they nay be homeless. Households consisting of extended familieswhich I cLcn provision for food, or of potentially separate house
holds with-a common head, result ing .fron polygamous unions, -.ay occupy more than one housing unit.
E/CN.14/SM/36 Page 5
24. !sThe persons involved in the above arrangements are counted as members of the household. Members of the holder's household also
include the holder himself. Care should be exercised not to miss members of the holder's household who are temporarily away, while temporary visits should be excluded.5'
25. "For the agricultural census a large household with more than one holder may have to be split into separate sub-households, each treated as a single household.5'
26. For the purpose of collecting data on agriculture an extended definition of agricultural household is recommended (2S p.31); ''A household is considered to be an agricultural household when at least one member of the household is operating a holding or_ when the house hold head, reference person or main income earner is economically active mainly in agriculture.- This definition covers two types of households: those whose nep>er (s) operate(s) a holding (farming households) and those whose head (or main income earner) is economi cally active in agriculture although he does not operate a holding
(non-farming agricultural households).
27. The notions of main income earner are introduced because some countries may prefer not to take into account the head of the house hold when he or she is not the real supporter of the dependents.
However, most countries still attach the dependents to the head of the household.
2S. The recommended definition of the agricultural household covers also households of landless agricultural workers as well as members of the households of holdings operated by a juridical person. Only a small proportion of households connected with agriculture will not be covered, namely those households which do not operate a holding or where the head (reference person or main income earner) is not economically active mainly in agriculture. Such cases are very rare in rural areas in developing countries.
Relationship between agricultural holding and farming household
29. When household surveys are used to collect data on agriculture, some problems may arise in establishing the relationship between an agricultural holding and a farming household. Different relations may exist:
a) One-tc-one correspondence between a holding and a farming household, that iss one holding only is operated by members of one agricultural household.
b) Two or mo^e households are jointly operating one holding.
c) Two or more members of the same household operate two or more holdings .
d) Combinations of a), b) and c).
E/CN.14/SM/36 Page 6
30. Case a), i.e., the one-to-one correspondence between households and holdings is typical in the majority of developing countries. The other cases are rather infrequent.
31. One example of case b) is when two persons, usually relatives, each having his own household, jointly operate one holding.
32. If in a sample survey households are selectee1 with equal probability then the joint holdings (case b) will have a higher probability of
selection, and this will result in biased estimates of holdings5 chara cteristics. A simple solution is to attach proper weights to such
holdings in the estimation formulae. In case of n^ households operating jointly one holding/ each selectee' household will get one n-th of the value of the hull' -.; ' s characteristics.
33. Tore frequent are cases vrhere two or more members of the same
household (case cs para.24) operate their own holdings. In some countries in Africa it is common that the wife operates her own piece of land.
Strictly speaking, these cases should be treated as separate holdingss if all conditions of the definition of holdings are fulfilled.
34. However, it may sometimes be difficult to decide whether the wives should be taken as holders, since the degree of independence in operating their piece of land may vary from case tc case. In order to avoid complicated field instructions some countries consider all hold ings operated by wives as part of the husband's (or the head of the household ?s) holding.
35. In studyi - J'r.n role of women in agriculture, information on v:r;en operating * h'-ir own pieces of land, regardless whether fully independently or not, may be of importance. It is therefore suggested as a minimum to record this fact on the holder's questionnaire,, when
listing all members of the holder's household.
36. As was mentioned in para. 25 large households tiay be pllt into two or more sub-households, each having only one holding. Tnis pro cedure requires precise field instructions and well trained enumerators and therefore it may not be always the most practical solution. For this reason, in many surveys all holdings operated by members of the same household are treated as one single holding. In that case the number of individual holdings is under-estimated, but estimates of other characteristics, such as area, production, inputs - etc., will still be valid.
37. It should te n:ted that very often in the census or survey reports the holding is defined in accordance with the recommended definition, but from the field instructions it can be seen that actually the
definition of a farming household was used, i.e. a household was
considered to be a holding if one or more members operated some piece of land or owned livestock.
E/CM.14/SM/36 Page 7
38. Only few developing countries gave indication o£ the number of joint holdings. I;. India it was found that less than 7 per cent of the total number of households have joint holdings with other house holds. In Kenya it is escin^t.eU that.-Lhere <*re less than 5 per cent of such cases.: - - .
39. In order to .shed more light on this problem, it is recommended that a few additional questions be . added to the house,ho.ld lists, indicating for each farming household whether any raeroLers operate their own land and whether any other household is jointly operating
the holding.
40. If such cases are relatively few, then the countries may decide to neglect them, and use the concept of farming household instead of
the concept of holding.
41. However, in some surveys it may be necessary to use the concept of holding. A typical case are the farn management surveys, where the notion of a technc-economic unit is essential. Also when the main objective of the survey is to study the land tenure system, the concept of holding is more appropriate.
42. At the beginning of an integrated system of household surveys for most developing countries it will be perhaps more convenient to use
the concept of farming household and disregard the cases where this does not fully correspond to the concept of holding. .This will enable the countries to benefit immediately, from, the. household survey system as a means of collecting data on households agricultural activities.
SECTION 3 .-■■.■.
Main items to be collected General considerations
43. Some of the proposed ileus ^e normally included or can easily be included in multi-purpose household surveys. These are, among-
others, some items on demographic and social aspects, such as sex, age, employment, etc.' of the agricultural population. Other items may in some cases require specialized surveys as, for example, area
of the agricultural holding, yield and production of crops, etc.
Which topics will be included in a survey, how and when will depend on the general programme of household surveys in a particular country, and this, in turn, will depend-on the-priorities established by the
country. ..•:.-..
44. The topics to te covered could be divided into three groups:
a) topics on characteristics of individual persons
b) topics on characteristics of households c) topics on characteristics of,holdings.
E/CN.14/SH/36 p&«e 8
The FAC document or, agricultural population and employment (2) contains a detailed list cf topics in the first and second £rour_ and a list of topics in the third proup related to the stuay of agricultural popula tion and employment. The FAC Propraip.no for the 1980 V.'orld Census of Agriculture (3) contains a list of topics recommended for collection in the decennial world agricultural censuses, but most of them can also be collected in household surveys.
Topics on characteristics of individual persons Topic
^esilence characteristics Place of usual residence Place of previous residence
Appurtenance to a holder's household Personal characteristics
Sex
Karital status
Age (at last birthday]
Ape (summary classes) ; Level of education
Relationship to a reference person (head; supporter or main income
earner)
Identification of main supporter of a dependent person
Characteristics of economic activity Activity stetus (current)
Activity status (usual) Category of dependent1 Occupation
- main
(i) in which post (ii) secondary (iii) . subsidiary .
Industry (branch of economic activity)
- of nain occupation
(i) in which most of the time is spent (ii) which brings major incoir.e
- of secondary occupation
- of subsidiary occupation .
Status in employment one vear
- for main occupation
- for secondary or subsidiary occupation (if (-. .C, 6.1, 6.2 of ISCC)
Economic activity on the holding
(agricultural work) one year
Time reference period
one year
fixed point
point point point point uoint
date in the past of
of of of of of
time
time time time time time
point of time point of time
short period, one week
one year
point of time
one year
of the time is spent
one. year
E/CK.14/SM/36 Page P
Topic on characteristics cf individual persons (cont'd)
Topic Time reference period
Characteristics of economic activity - (cont'd)
Economic activity off the holding one year - non-agricultural work off the holding
- agricultural work on other holdings - non-agricultural work outside the .
household
Time worked (hours/week) one week
- on the holding
- non-agricultural work in the household enterprise
- paid agricultural work on other holdings - paid non-agricultural work outside the
household '!
Time worked (days/year) one year
- on the holding
- non-agricultural work in the house hold enterprise
All these iterr.s car. be collected using the farming household as a un i t.
Topics on characteristics of farming households
Topic Time reference period
Structural characteristics
Number of persons point of time
-economically active
- economically active in agriculture
- economically active in other industries - not economically active
- independent .-,
-dependent
Characteristics of economic activity
Total time worked (man-hours per week) one week - work on the holding
- non-agricultural work in the household enterprise
- paid agricultural work on other holdings
- paid non-agricultural work outside the household
- unpaid work off the holding (exchange labour)
Page 10
one year
one year
Topic on characteristics of farming households (cont'd)
Topic Time reference period
Characteristics of economic activity
(contfd)
Total time worked (man-days per year) - work on the holding
- non-agricultural work in the household enterprise
- paid agricultural work on other holdings - paid non-agricultural work outside the
household
- unpaid work off the holding (exchange labour)
Other economic characteristics
Dependency of holder's household on the holding
- Holder's household income - Total income
i) from the holding
- monetary
- non-monetary (valued at retail prices)
ii) from other sources - monetary
■■.: ;■ ■•:■- non-monetary (valued at retail prices)
Topics on characteristics of holdings Topic
Structural characteristics
Size of holding in terms of:
- total area
- agricultural area - cropland area - arable land
Size of holding in terms of number of
heads of livestock Sector of agriculture
Type of holding (destination of output) Type of holding (crop, livestock)
Type of holding (decree of specializa tion of cultivation
Type of holding (level of improved techniques)
Tenure of lane
Level of mechanization
Use of progressive techniques
Time reference period
point of time
point of time point of tine Agricultural year point of time point of time point of time point of time agricultural year agricultural year
Page 11
Topics on characteristics of holdings (cbi»t !d)
Characteristics of- econor.ic- activity Agricultural work (operation's)
- by menbers of holder's household - by hired permanent workers
- by hired occasions 1 or seasonal workers Number of permanent workers
- ;ae!r:ters of the holder's household -hired workers
Total time worked on the holding . .(raaiv-hours per week)
- .by members, of, -l-|cl-d.orf; s household - by hired permanent workers' "'
- by hired occasional or seasonal workers Total tine worker1 on the holding
(mar.-days per year)
- by members of hold.er's household - by hired permanent workers
- Vy hired occasional or seasonal workers Total time worked on each of the agri-
cultural operations (man-days per year
- preparation of the land
- sowing or planting. . . .. ■
- crop husbandry (care of growing crop)
■ - harvesting - " : - ' /.■ . - ■ ■■■ ■
- animal husbandry . ,.
-marketing the'produce .' ■ .!
- administration -and1 supervision - other
Other economic charrcteristics
Volume of production for each main crdp Volume of sales,for each nain cash crop Value of total,production (valued at
prices . of first point of sale) Value of total sales
Time reference period
agricultural year
agricultural year
one week
.f-ricultural year
agricultural year
agricultural year agricultural year
agricultural year
4£. It should be noted that all topics on characteristics of holdings can be collected also from fariring households. In this case classifica tion by size of farming household will not be the same as the classi fication by size of holding in cases where joint holdings exist.
SECTION 4
Review of sone definitions
46. This section deals mainly with definitions of those concepts which are norrally not used in household surveys. A particular definition _, based on 9 certain criterion, may be suitable under the conditions prevailing in a country or to the aims for collecting the
Pp.pe 12 .
information but not so under ether conditions or if the aims for
collecting the data are different. Moreover, if and when more than = one objective is to be satisfied and thus-more than one aspect °-
the concept is tc he considered, collection of the primary data should
"e ■'one in such a- way as to permit the derivation of the oata cor- res^rrding to the required aspects, of, the. concept. Relevant concepts, international recommendations on definitions, classifications ant,
tiwe-re^erence periods are piven in references (?) and (3). The following discussion can be applied also to farming household.
Size and type of holding
47. The size and type of holdinp are major characteristics investi gated in agricultural censuses and in agro-economic surveys not only for their intrinsic value as desiraMe information but also as
criteria for the cross-classification of the different characteristics of the holding anc, in perticular, the characteristics of the, related
population and latour force. r., . . , • .
Size of the holding , ,..
4£. The concept of size of holdine can be defined in a number of. ways according to the tasis of measurement. It can be measured in terms of area, output, livestock, trees, latour, etc. Moreover, each of these can have different facets which satisfy certin objectives lut
not others.
49. The concept of area to be considered could be;
- ToteI area: which may or may not be satisfactory according tc
the size of unproductive land
- Agricultural area: which includes meadows and pastures, may
"be more relevant when the holding includes both crops and
livestock
- CrcD area or aratle land area: one or the other of which
could be particularly relevant i.f the size is to le correlated
with the laboitr inputs. ...
50. The agricultural output can be measured in terms of volune or
of value. Koreover the volume could be the volume of total production or only the 'volume'' of the marketed products (sales) and similarly the value could be the total value of the or only the incone of the farmer from the. sales. However,. since the volume of production of .different crops'ds.not additive; it should ,Ve shown separately by crop and only
for the la a in... crop or crops,.
51. Since the number of heads of different species of livestock are not additive, the use of the tot&l number of heads of livestock as a measure of the size has no ireanin? unless only one and the sane species is considered throughout or if some kind of conversion table is so prepared as to give the equivalent of the different species to a standard unit. It would be still more useful, if this livestock standard uniV Were rade equivalent , to some land r.rca sothat a n.o e comprehensive measure of the size of the holO.ir.jr combining botn the area of the holdiiic and its livestock is defined.
E/CN.14/SM/36 Pa?e 13
52. In holdings consisting only of tree plantations, the size of the holding (other than the area) could be the total number of trees or only the number of trees in producing age. However, if the planta tion consists of different species of trees of different importance, some conversion table should be established in order to make the number of trees of different species additive.
53. The size of labour requirements may also be a measure of size of the holding crtcept that it varies appreciably with the degree of mechanization. Moreover, in the traditional sector of agriculture, the work on the farm is mostly done by members of the holder's house hold in an irregular and perhaps not very efficient way and this may - decrease the value of such a measure of size.
54. If area of the holding is used as a measure of size, then many problems may arise. In the first place the respondent may not know;
at all the area which he operates, or he may know it expressed in some local units, which in turn may vary from one region to another.
Or the operator may know the time needed for. ploughing or the amount of seed. This is typical of the situation in most developing African countries and many developing countries in other regions.
55. Objective measurements have been recommended by FAO (5) whenever.
needed and feasible. It is recognized, however, that objective measurements are a rather costly and time consuming operation. In many cases, there "ore, objective measurements are taken only on a
small sub-spmple of fields.
Minimum size
56. In the censuses of agriculture, especially those carried out on a complete enumeration basis, practical considerations make it necessary to limit the enumeration to holdings above a certain limit of size.
Further, most of the agro-economic surveys carried out subsequent to a census use .is a frame the results of the census of agriculture and thus do not cover holdings b-low that lower limit. This results in a situation in which the most vulnerable section of the agricultural population i" not investigated at all.
57. It is recommended that such holdings be included in the list and
subsequently sampled with a small sampling fraction.
Type of holding
53. From the above review of size of holding it can be seen that the measures of ^ize depend on the type of holding. In the Programme for the 1970 World Census of Agriculture, a particular item on type of holding was introduced. This item covered two of the facets of the concept ftype of holding' , namely the concept based on the destination of the agricultural output (mainly for home consumption or mainly for sale) and the concept based on vegetal or animal production^ The latter was limited to those holdings which produced mainly for sale.
and the classification based on the most important activity was; crop holdings, livestock and poultry holdings, mixed holdings and others.
E/CN.-14/SM/36 Page 14
59. Because of the discouraging experience of many developing countries in 1970 World Census of Agriculture in collecting data on type of hold ing and the lack of agreement on criteria of general applicability; the items on type of holding were dropped in the.Programme for.the 19S0 World Census of Agriculture. The 1900 Programme left it open to countries to collect relevant data on type of holdings according tc their own requirements and countries were encouraged to undertake the
necessary research for this purpose.
Sectors of agriculture
Private, public and collective
60. Holdings operated by a civil person cr persons (one or more
individuals or one or wore households) have widely different character istics from those operated by a juridical person (i.e. corporation, cooperative, collective, etc.). Moreover, it is also found that the characteristics of holdings operated ty juridical persons may still differ widely according to the type of juridical persons. Thus, it ^ was found useful to sub-divide the agriculture into a number of distinct
(with no overlap) sectors and to present the data separately for each of these sectors. This could be done a priori and, moreover, the
questionnaires and the methods used for collecting the data may differ
from one sector to the ether.
61. A reasonably simple classification would be to sub-divide the
holdings into the following three sectors:
1. The Private sector which includes all holdings operated by individuals or households;.
2. The Public sector which covers those holdings operated by a central or local government either directly or through a special Vody:
3. The Collective sector which covers the holdings operated ty a group of persons who voluntarily or by mandate of the governing authority join together to exercise land rights in common. However, if a member of a collective receives a plot for his own personal use, this plot will be part of
the private sector.
Modern, progressive and traditional
62. Another classification of agricultural holdings is based on the importance of the holding and the degree of utilization of advanced agricultural techniques. In this classification, the holdings are sub-divided into: p.odern, progressive and traditional.
63. Such a classification would be very useful especially in stratified sample surveys of holdings with different sampling fractions for the three sectors. However, no simple definition which woulc show the exact limits of the different sectors can be given. The following
attempts have been made to define these sectors'
Page 15
64. Holdings in the uiodern sector (sometimes called estates, agri cultural establishments or simply large farms) have to conform to certain recognized criteria. These criteria are based on:
- Size of the holding: atove a certain fixed, limit:
-.Destination o.f the products: for sale
- Labour inputs: use of paic! permanent workers;
- Mechanization: use of important machinery and equipment"
- Organization: 1ook-keeping of records of activities, inputs
and output.
65. Holdings in the progressive sector conform to some of the above criteria tut not all. They are. of a moderately large size, they produce mainly for sale but some of the produce is for home consumption and
some mechanization and modern agricultural techniques have been intro
duced in the operation of the holding.
66. Holdings in the traditional sector are generally small in size, produce mainly for home consumption, do not employ, or only occasionally employ paid labour and use simple agricultural implements.
67. From the above it can be seen that the definitions are rather
vague and thus are differently interpreted by the countries which quite often use only the dichotomy modern/traditional farming or commercial/
subsistence farming.
6G. For precision purposes, in the organization of sample surveys it would be very useful that the distinction between the different sectors be made a priori and that they re considered as different domains of study with perhaps different questionnaires and different methods of collecting the data. It is recommended that at least the so-called modern sector, whatever be the definition used by the Country be
identified, that a separate list of the holdings belonging to it be established and that a larger sampling fraction (or the totality) of
these holdings \e investigatec.
69. It is obvious from the atove defined classifications that household surveys will be a suitable ireans for collecting information from the private, traditional and progressive sectors of agriculture.
70. Data from holdings in the public and collective sectors will be more conveniently collected separately by interview or ty mail inquiry.
The same,applies for raodern holdings in the private sector.
Agricultural population
71. The concept of agricultural population covers different groups of individuals according to the criterion used to identify the members of the grour- In the past, the different agencies collecting data on agricultural population used different types of agricultural population.
Thus, the statistical data on agricultural population were not compar able or col^erent. In order to remedy the situation, these different groups of population were given different names which, however,, were not always well adapted. In what follows, the two concepts of rural population" aiul..farm population are reviewed and the two other concepts, namely agricultural population activity-wise and agricultural popula tion incoipe-wise, are proposed.
Page 16
Rural population . :
72. There has Veen no internationally recognized definition of rural population or rural aieas nor :of its complementary part, urbar popula tion or urban areas. Each country decides for itself what it considers to l.e an urban area; the remaining area is then defined as rural.
This is unfortunate but cannot be fully remedied.
73. Ir. the past and until now, in many developing countries, rural area and rural population were considered to be agricultural and the so-called rural population was assumed,to be equivalent to agricultural population. However, in a number of'countries, the difference is now being taken into account in socio-economic surveys and the households are subdivided into four groups: rural.agricultural, rural non-
agricultural, urban agricultural and urban non-agricultural households.
Farr population
74. The concept of farm population has been based indifferently on one or another of two different criteria: the criterion of residence and the criterion of appurtenance to a holder(s household and also on a combination of both. In what follows, the term ''farm household1' will be restricted to the residence concept and not the definition given in the 1970 and preceding agricultural census programmes, while for the population of a holder's household the term -farming household-
will be used.
7C. In a number of censuses of housing* a question based on the residence criterion was asked, namely, 'Is the dwelling situated on a farm?' . For the 1PC0 V'orld Census of Agriculture, the census information on the population, lased or the second criterion, is restricted to the holders and members of their households.
76. On the other hand, for the previous world censuses of agriculture, the concept of farm population was based on both criteria. The farm population included all the members of the holder's household whether or not they resided on the holding and also all other persons living
on the holding.
77. The concepts of rural population, of farm population and of : population of farming households do net represent and should not, therefore, be confused with the concept of agricultural population.
In fact, the farm population in whichever way it is defined and the farming population both include persons active in other branches of economic activity and slso inactive persons who do not depend for their livelihood on the agricultural industry. Horeover, they Co not cover all persons who depend for their livelihood on agriculture.
Agricultural population (income-wise)
70. This variant of the concept of agricultural population is based on the criterion of main source of income. The definition is;
- The agricultural population consists of all the persons who derive their main incope from the agricultural industry and
their dependents.
E/CM.14/Sf.5/36
Pa^e 17
.■ere again, as tor the preceding concept of agricultural population, the aefinition can be taken in the limited sense and cover only ISIC Major Group 111 (13) or it can he extended to cover the totality of ISIC Major Division 1.
Recommended- definition . o-f ■agricultural population
79. For purposes of regional and international cowparalility, it might he better to select the definition v\.ich is most suitable and recommend it for universal use. The following definition of Agric- cultural Population, based on the usual type of economic activity and the industry is recommended:
All economically active persons engaged mainly in the agricultural industry (ISIC Major Group 111) during the reference year, irrespective of where they live or work,
and. their dependents.
30. .Retailed discussion of. these definitions with recommended classifications may te found in FAC publications', references (2)
(3). .. . ; . anc
:ECTICN 5 Sources of data
31. In this,.sect ion the principal sources of data on agriculture and possibilities of supplementing tit em by household surveys are
1^ £\ If 1 t^t w. r f ,*' ■
revised
Census of agriculture
02 . The irain advantage of a complete enumeration census is that it provides data at the lowest administrative level, and also auxiliary
data for subsequent samples. Because of the high cost of an agri cultural census few countries can afford to carry cut a complete enumeration every live or ten years.
IS. Tahing i,nto account the rather long period "let we en two censuses (5 or 10 years) ?.nd considering also the time needed for processing and tabulation of data (the final census results are often published more than two years after ..the actual census takinr) it is obvious that
the ever increasing demands for timely data cannot be net by censuses alone. Because there are also increasing demandsfor data on small administrative units, ancj these cannot always be met by relatively sir-all size sample surveys, again a combination of data from different sources with household surveys r.ay help.
C4. Special techniques have been developed for estimates of small areas , using a combination cf census data, large sample data, data from administrative registers, etc. For more details see ((:) and
references listed there.
iminrr ••
E/CF. 14 Pare IS
£.5. Another possibility of obtaining data for-small areas is to use a rotating sample design'. For characteristics that do not change very rapidly- 1/50 or 1/iOO of primary units could Le used in yearly rotation, which will give in 5 years averages for 20 or 10 per cent of the units.
This is sometimes more than a developing country could .afford in a quinquennial or decennial sample census. For more detailed discussion see reference (7).
36. Host developing countries take a sample o£ holdings in the ....
traditional sector and enumerate completely holdings in the modern sector (for definitions of traditional and modern see paras. 68-74) .
Ap r i cu 11 ur a 1- s amp le su r v- ey s ; '^ , ■
87. Tliese sample surveys may serve different purposes. In many developing countries they replace the complete enumeration census, in others' they .'supplement , on.,a yearly or even shorter basis, the data collected .through the decennial or quinquennial censuses. Other uses of sample surveys connected with censuses are pilot studies which procede the census, coverage checks; quality control, preliminary
tabulations from census records, etc.
G£. Household sawple surveys can supplement an agricultural census
by providing nore diverse and detailed data. For example, an agri - :. .>;'l : cultural census; will r.ornaliy net provide data on landless households engaged in some''agricultural activity, and this information can be obtained through household sample surveys. Also data on employment and time use in agricultural activities are not covered in much detail in the censuses and supplementary information can be obtained through
household surveys.
Cther sources of da/La ,
CO. In some developed countries, e.g.* Pemaark, Sweden, Australia, farm registers are an important source of information on basic
characteristics of agricultural holdings. Since the registers are established mainly for non-statistical purposes, financing comes fror.
their principal functions and statistical, uses are only secondary'.
Data from registers provide information at the lowest administrative levels, Vut usually they do not contain all, needed data. They are frequently used as a sampling frarje for sample surveys carried out to
obtain the needed detailed data.
°C. More recently information from farm registers is stored on
magnetic tapes and discs and therefore it is more readily accessible.
01. In this category are also cadastral (land use) registers, which are used vhere available as main source of information on land area and use. These registers are.'a very useful source of relevant data for monitoring changes in land tenure which is an important topic in analysing the effects cf agrarian reforms.
14/SK/36 Page 19
92. Administrative reports and/or records provide a variety of
information on agriculture in many developing countries. Agricultural extension workers and agricultural agents report at prescribed regular intervals on area sown, area harvested, estimated yield and production, crop diseases, agricultural inputs (fertilizers, pesticides^ etc.)>
the use of machinery, etc. Veterinary officers report on the number of animals vaccinated, from which an estimate of livestock numbers is
obtained.
93. To this category cf sources belong also records cf customs offices on imports and exports of agricultural commodities. records from Govern ment boards dealing with particular industrial crops (coffee 3 cocoa, maize, etc.) on production, marketing and prices of these crops.
94. The accuracy of these reports varies greatly, but they are a cheap source of data, being a by-product of the regular activities of these agencies.
95. To fill the gap in data on small areas more recently emphasis
is being given to the collection of information at the village (locality or community) level. This information is kept in the form of village files at the village level or stored on magnetic tapes or discs for ready access and updating in the central office.
96. Topics usually contained in village files are: number (or some times lists) of households* total population (by sex, age and main occupation), area cultivated (by crops), etc. Also information on access to roads, health facilities, availability of drinking water, access to agricultural services, etc. is collected.
Agricultural surveys as part of the National Household Survey Capability Programme (HKSCP)
97. One of the main features of the MHSCP is the integrated approach to statistical activities. Since in developing countries the majority of the population derives its livelihood from agriculture, almost all household surveys will have to cover some topics related to agriculture.
96. A typical example are the food consumption and nutrition data which are always collected throuph household surveys. 3ut other house hold surveys also deal with agriculture. A household labour force survey in rural areas deals with rural or agricultural population.
household income ant1, expenditure surveys and household economic surveys in rural areas are also related tc agriculture as the main source of income and the main activity of the majority of the rural populations.
99. Therefore it is of great importance for the developing countries that integration of data collected through different rounds of house hold surveys be secured. This can be achieved by defining common and compatible concepts and methods of data collection and analysis.
Pane 2C
100 The sur-ey programme should also be properly co-ordinated with other related-statistical -activities in order.that the household survey data are complementary to those from other sources, such as censuses and other iron-household surveys. In countries where agri cultural surveys are the responsibility of the national statistical office, such co-ordination can be easily achieved.
101 In other countries the collection of all food and agricultural statistics is the responsibility of a separate statistical office
usually in the ministry of planning. In this case proper co-ordination
should. 1-e established.
1C2. The decree of integration of household surveys may range from the harmonization of concepts, use of the common sampling frame, etc., to the full integration of all household surveys in the same central agency, using the sane office anc. field staff, data processing equip
ment , etc .
103. What arrangement is the lest solution depends on the actual
statistical organization in the country (centralized or decentralized) and on the present survey capability: availability of qualified staff,
equipment, etc.).
104. For some • Dpics it will be difficult to completely integrate agricultural surveys with household surveys, because of operational difficulties. The typical example are crop area and yield surveys.
In most developing countries in Africa and some countries in Asia, the farmers do" not know the area of land operated, cultivated ana harvested. In such cases objective measurements of areas have to be carried out on a subsample of holdings or households. This is a time consuming procedure (see also para.41) and it may not leave enough time to the enumerators to collect other data. For this reasonit would be more practical to organize a special survey for objective measurements of area using lidding or farming household as a unit, perhaps every five years, and supplement the data throufh interviews in the annual household surveys on a sub-sample basis in order to measure changes in areas. These changes are very important in many countries where shifting cultivation prevails and where operated area is, in fact, equal to cultivated area. In these cases, the size and
location of the cultivated areas change from one year to another.
105 Fcr further details about the measurement techniques and related
problems see (5). A new manual or. estimating crop areas and yielasis under preparation in FAC.
IOC Some of the household surveys are of the multi-subject_type and
cover a wide range of topics, such as demographic characteristics, health, food consumption and nutrition, housing conditions^and
facilities, education characteristics and cultural activities, employ ment, economic level of the household and others.
E/CN.14/SM/36 Page 21
107. It is recommended, when organizing a household survey, to separate from the start the households associated with agriculture from other households. Especially it is recommended to treat as a separate
domain of study those households whose members operate a holding (farming households). These can then be shown in the tables as a separate group, but still a linkage with other non-farming households
will be possible.
IOC. Since in the rural areas of many developing countries most of the households will be farminr households, a common questionnaire for all households can be used, containing questions on the characteristics of the holding. This approach was used in the 1970 Census of Agri culture and Forestry in Japan (12) and in the Integrated Rural Survey in Kenya, 1074-75 (10).
109. In urban areas where the farming households are an exception or relatively few, specific questionnaires or modules can be added to the main questionnaire for those households which are identified as farming households or agricultural households. These questionnaires will be filled at the appropriate rounds during the survey period.
SECTION 6
Actual censuses and surveys usinn household as a unit
110. A brief description is given of some agricultural censuses and surveys using the household as reporting and tabulation unit.
111. The collection and analysis of data needed for both development planning and monitoring the existing state of the economy is largely the responsibility of the Central Bureau of Statistics (CBS) of the
Ministry of Economic Planning and Development.
112. The CBS commenced the creation of a permanent survey capability in the mid-1970s. The first five-year programme of surveys, during which a continuous survey capability was achieved, was entitled the National Integrated Sample Survey Programme (HISSP). The main feature
of the NISSP was the Integrated Rural Surveys (IRS) conducted in four rounds, each covering an agricultural year, from 1974-1979. These surveys covered a wide range of demographic, agricultural and socio- economic characteristics. In addition, studies of special topics were included in the rural surveys, such as marketing of food
commodities, crop forecasting, nutritional status, literacy, housing and water supply. I'ore details can be found in the Report of the
1974/75 Rural Survey (10).
113 During the second half of 1979 a second phase of the programme was prepared entitled the National Sample Survey and Evaluation Programme 1980-84 (KASSEP) (in print as a CBS publication).
E/CH.14/SM/36 Page' 2 2
114 In the first phase of the IIISSP 1974-79 the control of survey operations was to a large extent decentralized to the statisticians responsible for each survey. The concentration of survey activities
in rural areas led to the de. facto emergence, of a field control unit which handled the day-to-day management of the field operations of different surveys. A rcore formal co-ordination of survey activities is envisaged in the new ilASSEP programme.
115 A sample from the IRS questionnaires is included in the Appendix.
As an illustration of the possibility of cross-tabulation of data by holdings and households a few taMes from the IRS report are also
reprinted in Appendix 2.
Philippines
116 In 1975 an integrated population census was undertaken, consisting of two phases. In the first phase data on socio-demographic character istics of the population were collected on a complete enumeration
basis. A sample of households was then selected in order to collect data on the economic activities of persons 10 years old and over in farming or g rdening, livestock and poultry raising, fishing, manu
facturing, processing, etc.
117. The following topics on agriculture were covered-;
Area of farm or garden operated, tenure of farm, area planted, production and disposal of produce, cost of operation and other expenses, farm equipment and facilities used, livestock and poultry (number, quantity consumed, quantity sold, cost of operation, etc.) fishing including off-shore or deep sea fishing -■:-■-- -" by households, fishing equipment, quantity
and value of catch.
India
112. The national Sample Survey (rSS) was started in India in 1950 to collect comprehensive, socio-economic data relating to different sectors of the economy. The thirty-second round was carried out in
1977/7S This round covered both urban and rural areas, and the main topic was employment. Other items covered were usual and weekly
activity of household members of ages 5 and above, cash expenditure and consumption of food, tobacco, fuel and light.
119. For rural labour households, both agricultural and the labour households, additional information on the type of agricultural operations and other work performed during the reference week was collected. Also information on the usual activity during the past year was collected. Data on lenc area possessed by the household,
land tenure, area of cultivated land vjas collected where appropriate, i.e. for households, with land.
120. All households were classified in 5 categories:
E/C1M4/SM/36 2 3
a) households self-employed in non-agricultural occupations . These were defined as those deriving more income from non- apricultural sources than from either rural labour or other
sources.
b) agricultural labour households are these which earned more than 50 per cent cf their total income during the past year fron wage-paid manual labour in agriculture.
c) other labour households are those which earned 50 per Cent or less froF wage-pair' labour.
d) households self-employed in agricultural occupations are those which earned pi ore than 50 per cent of their total income from self-employment in agriculture during the past
year. ■■ : ■
- -; e) other households.
121. It may be seen that this classification differs from the one recommended by FAC (2). However, from the tables it is possible to separate agricultural holdings, farming households anc other agri cultural households from other holdings, according to FAO recommenda
tions. ■-'
Japan
122. The 1970 World Census of Agriculture and Forestry in Japan is an example of combining different observation units into one survey. As can be seen from the table in Appendix.2 reproduced from the 1970 Census Report (12), survey of farn household provided information on the composition of the household working conditions of household^
members, land and its utilization, livestock, sericulture, agricultural machinery, etc. In addition a survey of holdings othei- than households
was conducted, for co-operatives, joint holdings, companies, etc.
123. A survey of rural communities was also conducted for all agri cultural rural communities, providing information on community's composition and function, production environment, life environment „ cooperative Si etc.
124. The forestry sector was covered by interviewing agriculturp.l and forestry households and other forest holders, by using the same questionnaire as for farm households.
Other country practices
125. In the 1971 agricultural census of Iraq, agricultural holding was the unit but households of hired workers were also included for some items. In Afghanistan in the 1969/71 agricultural sample survey, households were used as "enumeration units. In SyiJ.a in the 1970 Census of Agriculture, agricultural holding *as used as the enumeration unit.
A list of holdings was obtained through the household list from the population census. In Eenin the agricultural household was the unit
of enumeration in the 1976/77 agricultural survey.
E/CN.14/S"/36
Pape 24
126 In the 1970 Agricultural Census of Liberia the farming household
was'the unit of enumeration. Large farms (mostly modern sector of agriculture) v/ere enumerated exhaustively and were treated as a separate domain of study.127 A different approach was used in the 1970 Agricultural Sanple
Survey in Ethiopia. Farmers' Associations were used as primary sampling units and from the lists of members sample holdings wereselected.
128 The unit used in the 1970 Agricultural Census in Cameroon was the'holding (exploitation agricole), but the term agricultural^house hold (menage apricole) which corresponds to farming household is used as a synonym. ^The prir.ary units were the villages, ana within the sample villages dwellings (houses were listed)and -agricultural house holds" (manages agricoles) were identified anu selected.
129 The same concepts and definitions were used in the 1970 Agri cultural Census in the Central African Republic, Gabon, Chad and the
Peoples' Republic of the Congo.
130 In the 1972 survey in Swaziland blocks (groups of enumeration_
areas in the 1966 Population Census) were used as prinary units ana all homesteads in selected blocks were listed- A homestead was defined as a '-distinct single residential unit under the control oi a single head of homestead'-. Homestead heads cultivating land and/or, having livestock were considered as "holders". It seems that in this.
case the concept of holder would correspond to farming household.
