On-farm research and household economics


Material Information

On-farm research and household economics
Physical Description:
30 leaves : ; 28 cm.
Low, Allan
Publication Date:


Subjects / Keywords:
Agriculture -- Research -- On-farm   ( lcsh )
Agriculture -- Economic aspects   ( lcsh )
Farms, Small -- Economic aspects   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Includes bibliographical references (leaves 29-30).
Statement of Responsibility:
Allan Low.
General Note:
General Note:
Caption title.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 695401588
lcc - S540.O53 L69 1985
System ID:

This item is only available as the following downloads:

Full Text

On-Farm Riesearch and Household Economics

Allan Low'

1. Introduction

The topic I wish to address in this paper concerns the disparity
I see in the theoretical and practical relationships between y
on-farm research and household economics. The discussion will be
based on my experience of on-farm research in Eastern and
Southern Africa and in particular on the application of CIMMYT's
methodologies and concepts in this area.

The argument that I will put forward is that although there is in
theory a close relationship between the two new philosophies, of
"farming systems research" and householdd economics", these
relationships have not been sufficiently recognized or adequately
developed in practice. I will concentrate on the need for
on-farm research methodologies to be more orientated towards
household economics concepts than is currently the case.

My task: is complicated by the fact that there seems to be
considerable disagreement and misunderstanding about both the
terms "farming systems" and "household economicts". It will be
necessary therefore to start by defining what I mean by the5e
terms. First I am going to refer to on-farm research as being
that particular aspect of farming systems research in which
CIMMYT is involved. CIMMYT describes its farming systems
PERSPECTIVE (Byerlee, Harrington and Winkelmann, 1982). This
involves farm level research to:

(a) understand farmers' circumstances,

(b) generate hypotheses about how farm productivity can best
be improved in the near term given these circumstances
and current technical knowledge and,

(c) design and test technologies based on these hypotheses on
farmers' fields or,

(d) direct station research towards more relevant programmes
if the technical information base is lack:ino.

By household economics I ref er to the concept of hou !s e h ol
product ion behavi our that has its basis in the new theory of
consumer choice developed by Eeclker (19?65), Lancaster (196t) and
Muth (1966). This new theory of consumption sees ho~useholas eE
productio~n/consumption units in wh ich utility is not derived
directly from the conrsumrption of market goods and services bu-t of
inter-mediate non-mrark~et "I goods", which ar-e produced within the
ho us ehold. Mark::et coods and household resources (principally
time) ar-e combined in a household technorlogy2 to produce these 2
goods. which are then consumed in combination that generate

PaEge 1


maximum utility or satisfaction or welfare for the household.
The strengths of this theory are that it brings the production of
non-market goods into the analysis of consumer choice and it
provides a framework for analysing intra-household labour
allocation where the opportunity costs of time vary between
members for specific tasks.

The significance of recognizing that men and women in particular
have different opportunity costs or exchange values is brought
out in a number of the conference papers (Peters, Guyer, Jones).
Other papers ( V ier i ch, Hyden) emphasis se the i mpor tan ce of
non-marki;et t ran sactions GE labour anyestmentens necessary y t o g ai n
ac~c~ss tto ~l~annd an lab our resources inAfrica where,
traditionally, these resources do not enter the mark~et.

Improved opportunities for cash cropping and access to wage
employment and consumer goods (including staple supplies) means
that African households are increasingly faced with mak:ing
comparisons between the consumption of traditional non-mark-et and
modern mar ket goods and services. The household economic s
frameeworkP allows a direct analysis of these comp ar ison s on the
basis of the opportuni ty cost of household members' time in
alternative mark~ef-f~5F non-ma k:et goods and services.

Few of the growing number of studies on intra-houlsehold processes
in rural settings (e.g. see Binswanger et al, 198i0) have been
explicitly based on the new household economics theory of
consumer choice (a not abl e ex:ception is Evenson et al. 1980).
However most have highlighted the Importance of non-farm
n on -mTarki:et household p rod u ct ion act iv it ies and show how these
often influence the ability to conduct farming
activities. Household economics has been shown to provide a useful
addition to the neoclassical tool kit orunders~tanding and
analysing farmer response to economic and technical change in
Southern Africa (Low 1982(a), 1982 (b), 1982(c)). where resource
access and p rodvc t i on op portuni t ies are related to household
ccE omp O 0CIfeES~ where the c onsu5Ltmpt i on of market_ goodsc_ and
services takes place alongside the attacation .o~f considerable
amounts of t ime i nt o ec on om ic ally superiorc~l~ or socily nencessa y_~_
non-markI:et activi t ies.

This brings us to the theoretical relationship between on-farm
research and horuseholdd economics. On-farm research is aimed at
see):ing ways of increasing farm production, be It for the miarke:t
or for home consumption. On small farms crop or livestock~
production is organised within the context of the farm-househo~ld
which is not just a production but is also a c on sump t ion unit.
Non-market non-farm production of Z goods within the
f arm-h ouseh old (hou~seholld p rod uc t ion) f orms part of this
pr oduc t ion / consump t ion mil1i eu. Of ten on small African faris a
high proportion of available household resources are used in
household pr odulcti on act ivi ties. These activities may include
household maintenance or child care. but ma~EY also include the
investment of time into socially necessary activities, such as

P'age 2


ex change l_ 1abour,.. .to ensure access to essential non-mark~~et
resources of land and labour. Increased farm production through
technologies that use more household resources implies using
fewer of the available resources on household production. This
in turn implies the consumption of more farm goods or more market
goods purchased with the proceeds of the increased farm
production. The appropriateness of such a technology clearly
depends not only on the extent to which it increases the
productivity of household resources used in farming (the
product on aspect) but al so on a c omp ar ison of current
p roduccct i on -- ~w it h potential f ut ur e p rod u ct ion (the
investment/security aspect), as well as on a comparison of the
subjective value of the non-market household production goods
forgone with the utility and/or price of the substitute farm
derived goods consumed (the consumption aspect).

It would seem logical -therefore that household economics theory
and the study of intra-household processes would form an
important part of on-farm research methodology. This however is
not the case. On-farm research methodology tends to concentrate
on the interactions between different farming activities. Some
attempt is made to account for opportunity costs of time and
funds used in non-farm market production, but little attention is
gi veni to opportuni ty costs of resources used in non-farm
nn l-r-marketQroduct rion, inv~iies~fFItm~~ _~enSti an nsmpt~ion .

Furthermore the relationship between agricultural productivity
and household welfare is thought of as A one-way process and
assumed to be (or hoped will be) positive. As Norman et a1
(1982) put it, "The primary aim (of the farming systems approach
to research) is to increase the overall productivity of the
farming system -therefore, hopefully, the welfare of the
individual farming families".

Caldwell (1983) suggests that we need to move beyond the
hopefulky=S link: between farm income and household welfare. He
notes that welfare is a function of the total mix: of monetary and
non-monetar-y tangible and intangible goods, that perceptions of
welfare affect the goals of farm-household members, and therefore
thei~-r all1oc at ion of resources and management. Thus welfare is
not only a fun~~c~tio bt lo a determinant of the management of
a gr icul tuIr a l- proCduc t i vit y. Whe re houts ehoTD-;F~f'-~~r- TV 71f n1 I~ec ed
by non-farm factors (e.g. the wage employment market, access to
consumer goods or household composition)) and this determines the
objective and commitment to farming, it is clear that on-farm '
r ese arcoh ai med at generating appropriate technology must tak:e
account of these broader household factors.

In the remainder of this paper I will first consider why on-farm
research methodology and concepts have been limited to the
farming aspects of farm-household activity. I will then suggest
that on -f ar mr research results are indicating the need to thin)
more in household terms and will go on to Euggest areas where I
think that the new household economics perspective and an

F'agf =


appreciation of intra-household processes can contribute to the
effectiveness of on-farm research methodology and help to move us
beyond the notion of a one-way linkage between farm income and
household welfare.

2. The Development of On-Farm Research Methodologies;

2.1 The Rationale _f~or On-Farm Research

The need for on-farm research developed as a result of the
observation that much of the technology devYel oped from st ation
research programmes was not adopted by small farmers in
developing countries. One of the reasons for this lack of
adopti on was that the technol ogi es bei ng produced were not
consistent with the circumstances of many small farmers. i
Technol ogi es were seen to be i nappr opri ate ora number of:

(a) the natural circumstances (soils, topography, climate)
facing small f armer s in local specific situations vary
from location to location and are generally different
from those on research stations:

(b) the institutional support services needed to supply
inputs were either nonexistent or unreliable;

(c) the costs and risks: of using the new techno og ies were
too high for small farmers relative to the benefits;

(d) small farmers had multiple objectives stemming from their
need to consume much of what they produced, to mi ni mise
riskP and to max-imise returns to heterogpenous resources
su-ch as family l abour whose opp or tun i ty cost var i e
through the season as well as for different individuals.

By contrast the on-station researcher's environment i
characterized by particular (often favourable) natural
circumstances, availability of inputs, little concern with cost
or risk: and generally a single objective: to increase output per
unit of land.

On-farmr research methodology is aimed at sensitising agr-icultural
resea c h er-g P _toT_hE_circumstanrceso theit~11~~~~llll~~~rlC fame cien t s. Ap ~ar t
from getting researchers to conduct experiments on farmers'
fields. it was reconri sed that far-merS' circumstances are
determs ned by bothl physical and social factors. 'Thas led to th~e
need For i nt er -d isc ipli n ir y i nteracti on between so~ci al and
technical scientists. It was also rercognised that family farming
systems tended to be co~mpl.e?. W~thin the farm cont e:-t this
complen- ty was especi ally evident In tropical areas with long
growing seasons where intercropping and multiple e cropping was
practised'>. The recognition that fam21y farmingJ systems are

P'age 4

On-Farm Research Methodology

complex led to the need to adopt a systems perspective in which
i nteracti ons between acti vi ties could be accounted for. Again
this contrasts with the commodity and disciplinary orientation of
station based research.

The characteristics of on-farm research then are that it Is based
on first hand interaction between farmers and researchers, it i
interdisciplinary and it encompasses a systems perspective.

2.2 The Aims and Methodoloqv of On-Farm Research

The aim of on-farm research is to increase farm production
through the generation of appropriate technology in the near
term. The strategy for achieving this aim is based on the
adaptation of existing technology to better fit the needs and
circumstances of small farmers In developing countries.

The starting point is to gain an understanding of the farming
system and identify k:ey research opportunities which are lkl
to give a substantial pay-off in the near term. On-farm
researchers are increasingly relying on unstructured explortaory
surveys o sondeos) rapidly to gain an understanding of the
farming system (s) in an area. The techni que used is
characterized by a high degree of researcher participation in
farmer interviews and field observations (Collinson 1981,
Hil1derbrand 1981). A multidisciplinary team of researchers
interviews farmers in an informal and interactive manner. The
aim of the interviews is quickly to focus on areas where there is
research potential for adapting current: technology so that it is
more consistent with farmer circumstances and in ways that are
likely to improve farm productivity.

The output of efforts to understand farmer circumstances in the
e:-:pl1or ator y survey is the g ener at ion of hypotheses abouLt how
current technology can be adapted in ways that are consistent
with these circumstances. These hypotheses may be tested in more
formal, verification surveys in which emphasis is placed on the
collection of quantitative data to test the hypotheses formulated
at the exploratory stage.

On the basis of the formal surrvey results (or sometimes i n the
absence of for-mal veri fi cati on) research opp ortuni ti ec are
identified, priorritised in terms of farmers' needs and likely
pay/-off and then screened for systems compatibjility The two or
three most promising opportunities are designed as trials to be
set ouLt on farmers' fields.

The guiding principle of this methodology is that data collection
is designed as a sequential process, with information becoming
more detailed and focused at each stage. The objective of the
research is limited and clear: to generate appropriate technolocy
in the near term based on the use and adaptation of existingg
information and a k:nowledae of farmers' circum~stances

P'age 5

On-Farm Research Methodology

2.3 The Consideration of Intra-Household Processes

With the objective of on-farm research being to increase farm
production in the short term and the sharp focus of the
methodology on the farm, a minimal amount of consideration has
been gi ven to non-farm household activities and decision making
processes. On-farm research concepts and analysis techniques
have tended to concentrate on tak:ing account of how farmer
adoption of new technologies is influenced by (a) natural
circumstances, (b) institutional support or (c) cash costs and
riss.The final item listed in section 2.1, farmers' multiple
objectives (d), has been less thoroughly treated, due in part to
the lack of a theoretical basis for analysis of multiple market
and non-market objectives of farm-households' '. B~ut it is also
because items (a), (bj) and (c) can be handled within the contex:t
of the farming system. The need to adjust input rates
(f ertili sers, plant popul ati ons) to better fit l ocal si
characteristics or seedbed quality, for example, or the
suitability of a new crop in an e:;isting cropping system can be
established without reference to non-farm activities and
intra-household decision making processes.

Taking account of item (d), however, implies extending the ar-ea
of analysis from the farm to the farm-household and from a
concentration on production to more emphasis on consumption.
This reducaE the focus and coppolicates the analysis. However-
experience with on--farm research work: in Eastern and Southern
Africa is beginning to point towards the need to consider-
household/farm linkages more explicitly in technology generat ion
and suggests~ that there may be a case for extending the concepts
of on-farm research beyond the boundaries of the farm to
encompass the larger farm-household unit, despite added
complexity and some possible loss of focus.

3. Some On-Farm Research Findings

In this section I will outline somte of the findings coming out of
on-farm research in Eastern and Southern Africa, which
demonstrate the linIkages that exi:st between on-farm research And
household econToTi1cs.

3.1 The Importance of the Time Constraint

According to household economics theory, the time of itc. member~ts
is the basic resource of households. The oppoartuni ty cost olf
thi s resourLIce vari es over t ime and ait any one pornt In time
between household members of different gender, age and tls
A~n imrp location of the theory i s that ti me and cash are
slbst Itutable. Time can be "so03ld to generate cash or noln-marke:et
goocds and it can also be "purchased" by Epenoing cash on

P'age 6

On-Farm R~esearch Findings

time-saving inputs.

Diagnostic work in on-farm research is indicating that farmers
very often compromise on crop and livestock management, not
because of lack: of knowledge or for lack ofcash to purchase
inputs or because inputs are not available, but because of time

Often seemingly appropriate production increasing innovations are
not adopted because of their implications in terms of time. For
example, commenting on the results of: ex per mental workC: on
livest oc k feeding in the Kienya Dryl and Farming Research and
Development project, Tessema (1983) concluded that the rate of
adoption of innovations was disappointingly poor. He observed

K:enyan farmers valued their leisure more than the gains they
could get from clearing bush to encourage good forage

-most farmers are grazing their crop residues in situ and
realise that they are wasting about 40% of production in so
doing. Since they still go ahead with this practice, it
seems that, in terms of labour use, farmers choose the least
burdensome way of doing a job, even if they are aware that
an increased input will give A higher return;

-the growing of fodder crops creates greater demand for
labour- and o::en time, which the farmer cannot cope with if
he has to carry out operations of ploughing, planting and
weeding for food crop production. ThuIS only a handful( of
farmers were able to be persuaded to Include fodder crops in
their cropping system.

If we lookl: at a summary of the research thrusts that are coming
ourt of on-farm research work: in Southern A~frica (Table 1), the
importance of time in evaluating potential technical innovations
is again quiite evident.

P'ace 7

On-Farm Research Findings


Location/Problem Trials Evaluation Criteria

Zimbabwe (Mangwende)
Late seize planting
Late fertiliser application

Zimbabwe (Chibi South)
Staggered stize plantings
Shortage of oxen

Malawi (Local Maize)
Yitch weed
Lack of fertilisation

Malawi (hybrid maize)
Cob rot and late planting

Bot swana
Variable germination and poDF Need
control due to broadcasting

ninisue tillage/herbicide
Fertiliser sanagement

Variety x tires f planting

forage intercropping

Yield/ha, tise costs
YiEld/ha, tire EDst, fisC

sinious fields, risk
Yiedldha, tire costs

Reduiced reedinq reas~ireseots

planting tires


tise of planting/stalk
bending .

row planting methods

As for 2iababue and botswana

YeLRlidL, disease incidence, other
Operations at stalk bending time

returns to oren and labour time
and returns to cash

Land not a constraint and returns
per unit of land is low priority


Swazi land
Poor esergence low plant populations

Late and inadequate weeding

Poor seedbeds, late planting

Poor animal nutrition

Yield/ha, emergence, fertilizer
rates,farler Interest/assesscent
Yield/hla, time costs,
farmer interest!assesseent

YiEldlha, DFeT, work rates
gersination percentage
Weight gains, aren orkr rates

modified ox planter


fall and EarTy Ninter
stall feeding of grown
forage or bought hominy chop

'Further details on the diagnosed problems and implications for
research thrutsts are given in the appendi;:.

P'age B3

On-Farm Research Findings

3.2 Household Di ff erent iat ion

Household economics theory relates differences in behaviour
between households to differences in their characteristics and
comlposi ti on and, in particular, to the way these affect the
relative time values of members within a household. On-farm
research methodology recognizes that differences in the economic
and natural circumstances facing households will affect their
ability to and interest in adopting particular farm
technologies. The identification of different recommendation
domains in on-farm research has tended to be based on external
factors such as; agroclimatic conditions and access to markets or
inputs. However as research proceeds the importance of internal
household factors in determining appropriate technology is
beginning to emerge.

Table 2 Characteristics of Two Recommendation Domains
in Manqwende. Zimbabwe

Cattle Ownership
Owners Non-Owners


Family size (personc) 8.4 6.4
Farm workers 3.4 2.8
Size of holding (ha) 3.9 2.9
Aree cultivated (ha) 3.6 2. 1
X Farms with head working away 7 13
% Farms with head (55 years 17 42
%/ Farms with woman head 12 30

Crop Yields (t/ha)
Maize .21
Groundnuts .7 *5
Sunflower .2 .04
Income Sources (Z8/annum)
Maize sales 347 166
Vegetable sales 140: 84
Groundnut sales 40, 26
Off-farm income 159 149
Total income 752 449

In Table 2 we see that higher crop yields are achieved by cattle
owners than non-owners. These yield differences are related to
management factors. Cattle owners plant and weed earlier and
greater proportion of them winter plough and apply manure.
These management differences are In turn related to Internal
household factors. As Shumba (1983) states:

"Whil1e non-own ers and owners obtained the same abs~olute

P'age 9

On-Farm Research Findings

income from off-farm sources, this represents a much higher
proportion of total income for non-owners, who have lower
productive capacities in farming because of their smaller
labour forces, lack of oxen and greater tendency for the
household head to be away. The greater tendency for
household heads to be absent in non-owning households is
related to the younger age of these households. Job
prospects for younger household heads are better than for
their older counterparts and wages provide a relatively low
risk means for young households to generate the necessary
funds to hire cattle and purchase ferti liser. The i ncenti ve
for members of non-owning households to seek wage employment
is therefore quite high and, given their already smaller work:
forces, this further reduces time availIable for farm
activities and contributes to the lower levels of crop
management, lower yields and lower farm incomes of non-owners
compared with owners."

From a household economics perspective theinlec of h
domestic devel opment cycle on the productive capacity o
Tarr~households is clear.' D~en ownership is a critical factor
enabling better crop husbandry and the distribution of cattle in
this society is associated with household development and
maturrity. whi ch are related to the other factors mentioned by
Shumba and combine to result in poorer crop management by the
less mature non-owning households ***

Given the relationshiip between cattle ownership and crop
productivity and the observation that cattle numbers have
dec lined in the area due to a breakdown in health control and
drought, on-farm researchers have l ook::ed towards research
opportunities, such as improved feeding, that would increase the
size and capacity ofthe draught cattle pool. However,
recogni tion o the devel opment cycle li nkage poses two

(a) Would these extra cattle be any better distributed
between households?

(b) Would having cattle enable less mature households with
smaller work:forces to practise better crop management and
would the relative incentives to seek~ wage employment be
sufficiently reduced to encourage them to do so"

An answer to (a) may be deduced from looking at the situation in
neighboring Botswana, where cattle numbers have increased at
4.7% per annum over the last decade and average herd size has
increased from 30, to 45 head. Despite this sustained increase in
the siZe of the draught cattle pool, the proportion of households
owning cattle has remained unchanged and more than 50% of farmers
still do not own their own draught animals.

Pace 10,

On-Farm Research Findings

3.3. Women Farmers

As on-farm researchers conduct surveys and establish trials in
Eastern and Southern Africa, they increasingly find themselves
dealing with women farmers. At farmers' group meetings women
invariably outnumber men and it is said that 50-70% of farmers in
Africa are women.

Given that women the world over are responsible for household
production activities (household maintenance, child care etc.),
it follows that much of the agricultural work in Africa competes
with household production activities for the allocation of
women' time.

On-farm researchers; and farm management economists are accustomed
to assessing potential technical innovations in terms of labour
demands for competing farm acti vi ti es. Or altern ati ve mark::et
wage activities are accounted for by imputing an opportunity cost
of time. Seldom are the demands household production
considered, either directly or indirectly through an imputed
opportunity cost.

Rural household studies are beginning to highlight the large
amounts of t ime all1oc ated to non-farm non-mark~et household
activities, especially by women. Often the costs of not
per for m inrg some of these essential or socially neress~~_y_'ary _tasks~
(e.g.- fetching water- or ~worki~-ng in another's field) wi 1ll be
quite high and will s~ignificantly reduce the real benefits of
technologies that compete for the time of household members
responsible for such household production activities.

Factors affecting who does what within farm-households and how
many hands are available to do farming clearly have significant
implications for the appropriateness of new farm technology.
Tessemra's observations (page 9) that farmers; value leisure above
gains from bush clearing or choose the least burdensome way of
feeding crop residues, event though they k:now that an extra input
of time would give higher feed production, are made from a
farming systems perspective which lacks a household perspective
of farm-household decision making.

4. Towards a Household Economics Perspective in
On-Farm Research

In this final section I argue that the application of a household
economics persperspective can contribute to the effectiveness of
on farm research in three particular areas:

understanding farmers' objectives and strategies
defining recommendation domains and,
-evaluation of new technologies

P'age 11

Towards a Household Economics P'erspective

4.1 Understanding Farmers' Objectives and Strategies

On-farm research is designed to look at technology development
from the farmers' point of view. Understanding farmers'
objectives and values is crucial to this:

"The goals and motivations of farmers, which will affect
the degree and type of effort they will be willing to
devote to improving the productivity of their farming
systems, are essential inputs to the process of identifying
or designing potentially appropriate improved technologies."
(Norman et al., 1982:25).

While on-farm research recognizes that farmers have multiple
objectives, these objectives are generally looked at in terms of
the farming system. Multiple and intercropping strategies are
mani festati ons of farmers' mul ti ple object ves for cash,
preferred staple foods, food security and maximization of returns
to farm resources. Non-farm and non-market objectives of farmers
have been given less, if any, attention. As Bhenke and Kerven
state, this concentration on the farming system may have two
undesirable results.

"First it may encourage researchers to think of those who
f arm as pri mar ily or solely farmers, and thereby
underest mate the role of non-agricu~ltutral activities in
the l arger household economy. Secondly, an ex cl3usi ve
concentration on farming may ill equip FSR to address one
of the major issues in agricultural development in Africa:
the withdrawal of l abour from agricult ure due to
rur al -urban mi great ion. (Bhenk:e and K~erven, 1 983 :9)

In Eastern and Southern Africa farming is seldom the only source
ofincome and in many cases it is not the major one. Wage
employment, beer brewing, handicrafts, trading and teaching are
common additional sources of income for rural households. While
on-farm researchers are concerned with measuring and increasing
farm income, farmers are concerned with stabilizing and
increasing their entire welfare, much of which may comE frOM
non-farm production. Thus in order to understand "farmers"'
goals and objectives, on-farm researchers need to adopt a
household economics perspective and attempt to see how diverse
production activities are combined to maximise household utility.
To quote Bhenk:e and Kerven again:

"the acceptability of a farming innovation cannot be
adequately judged solely by its technical and economic c
impact on farming. It must also be assessed in terms of
its positive or negative contri buti on tothe household
Economy as a whole. This will especially be the case when
technical innovations require additional labour or capital
that could be invested elsewhere, for ex-:ample, in the
search for urban jobs or in the educati on of child en. "
(ibjd, p.10)

Pasge 12

Towards a Household Economics P~erspective

Application of a household economics perspective will help
on-farm researchers to understand farmer strategies in a
household context and thus to search for farm technologies that
are appropriate to the overall, farm and non-farm, circumstances
facing farmers. Given a household perspective it is possible to
see, for example, that one important risk-reducing strategy
adopted by many farmers is the search for wage employment by one
or more members of the rural household. Over the last two
drought years in Southern Africa those households that have had a
wage earning member have suffered much less than those that have
not had a reliable non-farm source of income. Clearly where the
chances of obtaining off-farm employment are quite good, any farm
based risk: avoidance strategy, such as planting an extra area of
cassava, or tied ridging or mulching or insect control must be
compared with the returns and reliability of obtaining income
from wage employment.

Norman (1983:7) notes that in the case of B~otswana it may be
necessary to accept that farmers will be reluctant to invest very
much (money or time) in crop production because such investment
is risk::y compared to putting it in to other areas such as
livestock~ and off-farm activities. Such a real isation clearly
has important implic nations for the generation of relevant
improved agricultural technology for small farmers in Botswana.

4.2 Defining Recommendation Domains

Given the recognition that different farmers face different
circumstances, but that resources do not permit research to be
geared towards individual farmers, the concept of the
recommendation domain has become central to on-farm research
methodology. The definition of a recommendation domain i
homogeneous group of farmers who share the same problems and
possess similar resources for solving these problems. This group
of f ar mer s i ep ec ted to adopt (or not adopt) the same
recommendation given equal access to information about it. In
much of Southern Africa, different recommendation domains occur
not only because of differences in farmer resources, cropping
opportunities, market access and inherent land fertility but also
because, at any one time, farm households have different
opportunities for non-farm wage employment or other
i ncome-earni ng acti vi t ies. Often it is the nature and e:tent to:
which farm households exploit these non-farm opportunities that
most strongly influences farming practices and the aims and
objectives of farm production.

Thus it is commonly observed that within homogenous agro-climatic
locations with similar market opportunities, neighboring farmers
with similar income or resource levels will farm in very
different ways. Households that are in a position to exploit
non-farm income opportunities, by dint of better qualifications.
experience or enterprise, will tend not to have the ti me or

Page 13

Towards a Household Economics Perspective

inclination to manage their farming operations in as thorough a
manner as their less wage employment orientated neighbours. With
these farm--households, the aims and objectives of farming tend to
be less production-orientated and more social- and
security-orientated. Their neighbours who are less able to
exploit non-farm opportunities will look on farming more in terms
of production and income and will tend to give more time and
attention to farming activities. The cultivation practices of
these two types of farmers will differ as will relevant
interventions and recommendations.

For example a recommendation domain exercise was carried out in
Swaz iland wi th the exipectati on that different farming systems
would be observed in the very different ecological conditions of
the highveld, middleveld and l owveld areas in the country
(Watson, 1983). However, it was found that the variations within
the regi ons were much greater than the variations in cropping
systems between the regions. Th e wit~hin reg _~_)__ia~~on .aitns temmed
from differences in internal household ci rcumstances rather than
from external circumstances. Table 3 gives a breakdown of these
farm-household types according to household characteristics and
relates these differences to the typical cropping practices
emp loyed by each group and the potential relevant interventions
implied by farm-household circumstances and current practices.

P'age 14


Relationship between Houlsehold Characteristic Determined
Recommendation Domains and on-Farm Trials

Cropping Practices
Fars Household Type Distinguishing Features fixed lion-Experisental Potential Interventions

Towards a Household Economics Perspective

(a) Minter or 2 x plough,
early plant, 2+ usedings
(b) High level of input use,
e.g. fertilisers (top
dressing), hybrid seed,

Top dressing IEVEl5
Tied ridging
Winter ploughingitractor)
Early planting (hybrids)
Double cropping

1. Cash/resource
rich and
labour rich

(a) 4 + adult equivalent in
fasily fars orkrfore
(b) Access to significant
non-fare income

2. C~ash/resource (a) (4 adult equivalent in (a) Only one ploughing, late (a) Top dressing
rich but labour family fare workforce planting, I useding, use (b) Botsrana plough/planter
poor (b) Access to significant of planter (c) Botswans improved planter
non-fars income (b) High levels of input use (d) winter ploughing(tractor)
(c) May or say not own e.g. fertilisers and top (e) Short season varietiES
oxen dressing, hybrid seed, (f) Herbicides

3. Cash/resource (a) 4+ adult equivalent in (a) 2 x plough, early plant (a) Minter ploughing
poor, but labour family fars workforce 2 x usedings (b) 2 x ploughing
rich (b) Poor access to (b) Lower levels of input (c) Better weeding
non-fare income use, E.g. no top Id) Double cropping
(c) Have some own oxen dressing, less hybrid (e) Intensive 5Neet potato
saize, no tractors production
(f) Cutwors banding
and scouting
(g) Early planting
(h) Fodder conservation
(i) Tied ridging

4. Cash/resource (a) (4 adult equivalent in (a) lxw plough, late (a) Minious tillage
poor and family fars workforce planting, l x useding, (b) Tyne plough
labour poor (b) Poor access to hand planting in furrow (e.g. Zirbabwe)
non-fare income (b) Low levels of input use, (c) Short season varieties
(C) Few if any cattle local or open pollinated (open pollinated)
varieties, no tractors

Page 15

Towards a Household Economics Perspective

In Table 3 the farm household types have been broken down on the
basis of off-farm income/resource endowments and labour committed
to farming. These factors are not independent. Previous surveys
(e.g. de Wletter, 1981) have found that there is a strong
positive relationship between off-farm income and farm-household
resource endowments. There is al so an i nver se rel at ion shi p
between off-farm income and the time and attention directed to
farming. However, some households are able to exploit off-farm
income opportunities and to commit time and attention to farming
as well. These generally are households where the head is not
engaged in off-farm employment but other household members are
earning cash away from the farm. Such farm households would fall
into category 1 in Table 3. Other farm households may find that
they are able to exploit off-farm earning opportunities, but in
order to do so they compromise on time devoted to farming. These
fall into category 2 in Table 3.

Category 3 represents households with relatively little potential
for ex:ploi ti ng of f-farm income opp ort uni ti es, but wi th a
reasonable labour and resource endowment for farming. The last
category may be represented by old households or women-headed
households, where opportunities for off-farm income exploitation
are poor and labour and resource endowments are al so relati vely

We can identify some of the distinguishing features of these
household types as is done in Column 2 of Table 3. On the basis
of survey data we can also list the different types of observed
cropping practice that apply to each category, see column 3 of
Table 3. These lead us to suggest different sets of potential
relevant interventions for each farm-household type. For
example, type 1 households can contemplate cash expensive inputs
and have the resources and commitment to manage these inputs
reasonably well. Group 2 households may have the cash and
incentive to purchase inputs but will tend not to manage them so
intensively. Thus fertilizer top dressing may be a relevant
intervention for both group 1 and 2 farm households. But the
conditions under which the resource of top dressing is tested
should differ significantly for the two groups households.
Trials related to group 1 households would be conducted with good
land preparations, early planting and adequate weed control. The
response oftop dressing under conditions relevant to group 2
farm households (poor seedbed preparation, late planting and
little weeding) is likely to be very different. Not only that,
the relevant evals 1uation of the trial results may' also be
different. If group 1 farm households are surplus producers the
relevant value for any increased production is the mark::et price
of maize. For group 2, deficit producers, the relevant value to
place on yield increases wi ll rather be the equivalent food
purchase price.

Page 16

Towards a Household Economics perspective

Another example is the introduction of an early maturing short
season maize variety for group 1 farmers. This may open the door
ordouble cropping and if so the benefit of an early maturing
variety should take the value of the second crop into account.
For households in groups 2 or 4, however, where circumstances
dictate late planting, the potential benefit of a short season
variety will be that it can better exploit the limited growing
period. It should therefore be evaluated in terms of its
production compared with current varieties, when planted late.

The point of this example is to demonstrate how the specific!
composition of the household affects the relevance of improved
technology. On-farm research methodology takes the farm as thel
unit of analysis and the farmer as the single decision maker.
Rather little thought has gone into the question of how the
family farm unit is defined, whether it is managed within a
nuclear family structure, or through an extended family structure
for ex:ampl.e. Given that household composi ti on can aff ect the
appropriateness of technology, there would seem to be a case for
on farm researchers to pay more attention to how farming
households are organised and how farming units are defined.

On-farm researchers in Swaziland attempted to do this through a
series of questions in the formal survey designed to establ ish
the roranisational structure of each homestead and how this
affected farming. Fi rst, qiesti ons were asked to establ ish
wfitter-therth homestead was composed of a single or multiple
households on the basis of the Swazi definition of a household
(tindlu) as well as on the basis of the number of separate
kitchens (emadladla) in the homestead If the homestead
comprised more than one household on either count questions were
asked to establish whether households farmed separately or
together. A household census was completed for each household in
the homestead and subsequent questions in the survey were to
relate specifically to the household of the respondent.

In the case of multiple household homesteads (25% of the sample)
it proved difficult to establish clearly how independently they
farmed and to restrict the answers to subsequent questions, on
labour inputs for example, to the members of the respondent's
household alone. It became clear that some households which were
enumerated separately actually formed part of the larger
homestead, which in reality comprised the farming unit. Whi le
quite a lot was learned about the organi sat ion of households
within homesteads, the approach did lead to a considerable
extension of the questionnaire and some complications in
subsequent aneal ysi s. Clearly there is a cost to gaining
information and understanding of household structures and
organisation in terms of the time needed for surveying and
analysis, which is difficult to justify given the emphasis on
obtaining short term payoffs from on-farm research.

HowevLer some on -far m r es ear ch er s in some environments e.

Page 17

Towards a Household Economics Perspective

Norman, 198'3:12) are now suggesting that there is a dearth of
relevant experimental station data that on-farm researchers can
adapt to suit local farmers' circumstances and that in these
cases "it is difficult to visualise major pay-offs to FSR in the
next few years". Where the prospects for near term pay-off from
adaptive research are less good there may be more of a case for
ex tend ing the d iagnost ic wor k to incorporate household
characteristic analysis and gain a fuller understanding of the
household environments within which farmers operate.

4.3 Evaluation of Technologies

One of the outcomes of applying an on-farm research approach to
technology generation in a way which seeks to take the farmers'
viewpoint into account is that researchers are coming to
recognize that technologies which increase productivity per unit
of land are not the only ones that can be beneficial to small
farmers. Technologies that do not increase area yields but use
time or cash more efficiently are often equally acceptable.

It is becoming apparent that technologies which save family
l abourr ti me in parti cul ar are attractive to small family farm
units. The rapid uptakie of improved implements, herbicides and
mechanisation by small farmers around the world, as well as
farmers' own labour-saving strategies bear witness to this.

From a household economics perspective utility is maximised by
producing the desired set of goods with the least cost in terms
of the ultimate resource: the time of its members. Given the
many demands for family labour in farm and non-farm activities,
market and no~In-mark~et production, work: and lei sure, household
economics sees f am ily l abour as being at a premium, with the
major objective of houiseholdS being to employ it in alternative
uses as efficiently as possible. This Implies that households
s~eek to maximise the subtjective return to the labour of their
members and that what task):s are performed and by whom depends on
the opportunity cost of members' time.

The opportunity cost of labour time also often forms an important
cormp on ent in the eval uat ion of f ar m technologies by on-farm
researchers. However, opportunity time costs are generally
d eter mi ned in terms of alternative farm activities or of wages
that can be earned off the farm. Thus, for example, the cost of
the time of women in parts of the season when there is little
crop work~: to be done is generally assumed to be near- to zero.

Commenting on the unresponsiveness of farmers to advice on bush
clearing in Western K:enya which experimental results hao shown to
be productive, Tessema (198#) says:

"Many were unwilling to carry out the work: because they say
it is a hard and difficult task: even though it does not
conflict with other operations, as it can be done in the

Page 18

Towards a Household Economicsi Perspective

dry season when there is little other activity".

Even in times of little farm activity the demands on family
labour are many and, as the above example illustrates, it is
wrong to assume that when there is little farm work to do the
opportunity cost to family labour is negligible.

Taking a household economics perspective will help to prevent
researchers from falling into Tessema's trap, and will "provide a
basis for making some assessment of what value to place on family
labour used outside farming and wage employment. The question
researchers need to ask: is what other task~s are being perf ormed
by the relevant household members at the time. Answering this
question will probably be easier than going on to the next stage
and estimating the subjective value of a unit of the member's
time in that activity. What value do you put on an hour spent
I~looking after children or collecting Firewood or drinking beer
with friends? The important point though is that the answer is
certainly not zero just because the activity does not have to do
with farming.

The growing number of rural household studies can complement
on-farm research and contribute to its effectiveness by:

(a) highlighting the importance ofnon-farm non-wage
activities in household production and,

(b) generating some rough guidelines on the values that
might be placed on the use of labour time in household
production activities.

Even where positive opportunity time costs are assumed, the farm
based and hou1Sehold economics approaches to evaluating farm
technologies can give markedly different results. For ex ampl e
Table 4 presents A typical partial budget analysis in which
opportunity costs of labour are included and a reasonable return
on capital is obtained when extra management time and fertiliser
are applied.

Page 19

Towards a Household Economics Perspective

Table 4
Farm Based Partial Budget Analysis

Traditional New Technology

Yield kg/ha 1300 2400
Adjusted yield (-15%) 1100 2040
Gross return at cl/kg 1100 2040

cost of fertilizer 192

labour input (man days) 61 106
cost h cl0/day 610 1060

Total variable costs 610 1252

Net benefit per hectare 490 788

Return on capital (788-490)/ (1252-610) = 46%

'Data from Bruce et al (1980) table 6.2

Compare this approach with the following analysis of the same
data based on the household economics theory, that farm
households seek: to utilise the labour time of their members as
efficiently as possible. Table 5 presents the analysis on the
basis of time requirements per unit of produce.

Tabtle 5
Household Economics Time Efficiency Analysis

Tradi tional New Technology

Man days required/ton 55 52

Man days saved/ton 3
value of time saved c/ton 30

Cash cost of saving 3
man days/ton (c) 94

Net loss; 64

With the new technology each ton can be produced using three
fewer man days of time. However the technology requires an extra
cash outlay of c94 per ton produced. At the previous time value
of clO per man day the value of time saved does not justify this
e:tra cash cost.

P'age 20:

Towards a Household Economicss Perspective

More important than the different answers given by each analysis,
is the difference in the implications of each approach for
changes in the value of time of household members (or household
welfare). In the farm based approach the new technology becomes
LESS attractive as the opportunity cost of time is increased,
since the new technology uses more labour per unit of the
enterprise and net returns are reduced.

In the household economics approach, the new technology becomes
MORE attractive as the opportunity cost of time increases,
because less time is needed to produce each unit and the value of
this time saving is increased.

It seems to me logical that, where labour hiring is not prevalent
and scarce family labour time must be used in the crop activity,
increasing values of members' time (or household welfare) is
likely to encourage the use of the cash expensive, labour saving
technology, rather than discourage it as the farm based approach

Clearly an understanding of household circumstances, aims and
objectives is crucial to the evaluation and design of appropriate
technology for small farmers.

5. Summary

In this paper I have suggested that the two new philosophies of
on-farm research and household economics have much in common
theoretically, but that household economics thinking isnot
generally incorporated into on-farm research methodology. This
is because on-farm research has a limited and clear objective to
generate appropriate farm technology in the near term based on ;
the use and adaptation of existing technical information and a/
knowledge of farmers' circumstances. This can often be achieved
by taking account of the natural and economic environment in
which the farm operates, without broadening the analysis beyond
the boundaries of the farm to encompass the larger farm-household
unit. However, on-farm research findings are beginning to;
indicate that on small farms the linkages between the farm and
household are quite strong and that household factors will often
have a significant influence on farm decisions. In the final
section I suggest that there are a number of areas of on-farm
research methodology in which the incorporation of a household
economics perspective can help to improve the effectiveness of
the approach and help researchers to more clearly perceive the
i mpl icati ons of the two way link:age that e:ists between farm
income and household welfare.

This two way relationship between household welfare and farm
income has implic nations for the macro-macro link.ages theme of
this conference which has not been addressed so far in this
paper. Hyden has emphasised three characteristics of the-
socio-economic environment within which farm-households operate

Page 21

Towards a Household Economics Perspective

in Africa:

(a) there is universal land access;

(b) there is no agricultural surpluE labour;

(c) urban migrants retain strong rural links.

Universal access to land results in specialisation taking place
within rather than between households. This means that
farm-households are often not primarily or solely farmers. Since
there is a lack of agricultural labour on the mar ket, non-farm
activities directly compete with family labour on the farm and
this often results in farm production being compromised. Even
where most income is obtained off the farm, households maintain a
rural base for social and security reasons. Overall household
welfare is maximised in this way, though it implies reduced farm
production due to labour shortages. Hyden refers to this as
premature urbanisation. It may be premature from the macro
agricultural production point of view. But in the socio-economic
envir-onment of Africa it max:imi~ses welfare at the farm-household

Page 22



Research Thrusts From OFR in Southern Africa

Al. Zimbabwe (Manqwende)

(a) Problem Diagnosis

Hybrid seed and fertilizer were widely used for maize
production, but lacki of draught power reduced the ability
to plant early. Staggered plantings occurred because of the
draught power constraint, but this was also done as a risk:
avoidance strategy. To avoid the risk of wasting fertilizer
and because of labour time constraints at planting time,
basal fertilizer was applied after germination.

(b) Research Thrusts and Evaluation Criteria

Minimum tillage trials were designed with direct planting
into winter ploughed land or tine planting with herbicides
to minimise the use of draught animals. Evaluation criteria
include yield per hectare and time costs. Earlier planting
did result in increased yields per hectare, but differences
found in time costs justified the technique even when there
were no statistically significant yield differences or
yields were reduced with minimum tillage and herbicide

Fer t~ili ser management i nvol ved pl ac ement of basal
fertiliser at planting to increase efficiency of use of
nutrients. Placement increased yields but also required
more labour time at planting. The risk: element was not
assessed directly.

(c) Implications

Diagnosis has shown that the problem in Mangwende is not
lack of use of technical inputs of improved seed and
fertilisers, burt how these are managed. Minimum tillag~e
appears to be an appr opriate intervention which could
result in increased yields due to earli er planti ng. Even
for farmers who could not plant earlier (because they don't
own oxen for example) minimum tillage and herbicide use may
be appr opriate on account of its reduced cost in terms of

Placement of basel fertilizer at plant ng does i increase
yields. but also requires more labouir. Depending on the
opportunity costs of labourr at this time, placement may or
may not be acceptable.

Page 28


A2. limbabwe (Chibi South)

(a) Problem' Diagn~osis

Maize plantings are often late and staggered on account of
the highly variable rainfall and shortage of o:en.

(b) Research Thrusts and Evaluation Criteria

Time of planting by variety trials to establish a planting
strategy that minimises the chances of achieving very low
production levels. Evaluation criteria need to include
variety performance at different planting times over a
number of years.

Intercropping maize with legumes to enhance the quality of
the stover and improve draught animal nutrition. Evaluation
criteria include the reduction in maize yield compared to
the improved quality- of fodder and how this relates to
better cattle performance and the investment of time in
planting and harvesting the legumes.

(c) Implications

These trials involve complex evaluation criteria which are
related to reducing risk: and increasing cattle performance.
Simple measures of yields per unit area of land will not
necessarily determine acceptability.

A3. Mal1awi (local maize)

(a) Problem Diagnosis

Local maize yields are limited by witch weed and lack o
fertilisation. Farmers tend to only do one late weeding on
account of labour shortages and fertiliser i applied to
hybrid-maize in preference to local maize.

(b) Research Thrusts

Establ ish the possibility~ of getting better witch weed
control through earlier planting and determine the response
of local maize to fertilisation under farmer conditions.

(c) Implications

Lack of data on effectiveness of herbicides for witch weed
control in Malawi precluded on farm herbicide trials. More
basic research is needed on this. Rotation as a means to
control witch weed is not feasible given the large
proportion of the area put to maize. Early planting is
being examined at a potential low labour i nput method of

Page 24



A4. nalawi (hybrid maize)

(a) Problem Diagnosis

Farmers indicated that they planted hybrid maize late,
because it suffered from cob rot in the field if: it matured
too early. Yield potential was being lost through late

(b) Research Thrust

Time of planting and stalk bending trial. To estimate the
effectiveness of stalk bending to reduce cob rot and the
enhanced yield from early planting.

(c) Evaluation Criteria

Evaluation criteria include cob rot incidence, yield,
labour time required for stalk: bending and other tasks:
being performed at this t ime. Eval uat ion i nvo'lves
balancing increased yield from early planting against cost
of time for stalk bending to combat cob rot.

A5. Botswana

(a) Problem Diagnosis

90%i of farmers broadcast and harrow in their maize and
sorghum seed. This leads to variable germination and poor
weeding because mechanical control is not possible.

(b) Research Thrusts

Development of row planting methods which allow .more even
germination, better plant establishment and mechanical
weeding but which do not require so murch extra labour and
o::en time that the area planted and total production is
reduced .

(c) Evaluation Criteria

Third furrow planting, performance of the Sebele plough
planter and tractor plough planting methods are being
evaluated in terms of animal draught and labour inputs
requi red f or the est abli shment of sufficient sorghumr to
provide subsistence and repay loans.

P'age 25


A6. Jambia

(a) Problem Diagnosis

Two major constraints in maize production are similar to
those found in Zimbabwe and Botswana: late planting and
weeding on account of draught and labour shortages, plus
sub-optimal fertilizer management.

(b) Riesearch Thrusts

The 1982/3 adaptive research trial programme included
trials on:

-short season varieties (5 trials)
Sfertiliser response and management (4)
-zero till and herbicides (4)
-ox planting methods and plant establishment (2)

(c) Evaluation Criteria

In most of Zambia land is not limiting. Farmers are using
technical inputs such as hybrid seed and fertilizer, but
they are not attaining maximum potential from them per unit
area because their management isbeing compromised by
shortages of draught and labour relative to land. Aswas
stated in the draught power and animal feeding network:shop
in October, the decision (of the farmer) is whether to
plant more or to go back and weed the first planting.
Priority is generally given to the establishment of crops
(rather than better management of established crops).

A7. Swaziland

(a) Problem D~iagnosis

Because of other work: commitments, Swazi farmers luook to
getting their land prepared and planted as quickly as
possible. This is achieved with a single or: or tractor
ploughing, which leaves a rough seedbed, and the widespread
use of the safim planter, which places the seed with the
fertillser and can cause scorching and poor germination.
The result is poor emergence and low plant populations.

Because of labour time constraints weeding is done late and
seldom more than once.

(bj) Riesearch Thruscts

Farmer reaction to a modified safim planter that places the
fertilizer to one side of the seed is being tested.

Pagoe 26


The ease of use and effectiveness of granular and liquid
herbicides are being tested under farmer conditions.

(c) Evaluation Criteria

The major evaluation criteria at this stage are farmer
interest and assessment of the technologies. Additional
criteria with the modified planter include days to
emergence, germination percentage and yields per hectare.
Lack of germination differences observed to date are
thought to be related to the rates of fertilizer used.

Given interest in one of other type of herbicide,
subsequent research cycles will test appropriate levels and
methods of application.

AB. Lesotho

(a) P'roblem Diagnosis

Lack of draught power and labour leads to poor seedbed
preparation and late planting. Labour shortages lead also
to inadequate weeding and untimely harvesting.

(b) Research Thru~sts

Fall and early winter ploughing to enable oxen to work when
they are in better condition, conserve moisture and
suppress weeds.

Stall feeding of oxen with grown and harvested fodder or
bought hominy chop.

(c) Evaluation Criteria

Yields per hectare and household levels of production
through better work~ output from ox:en leading to reductions
in fallow areas.

Time costs for harvesting and carting forage. Farmers did
not. bring sufficient forage to the pens in the ex:perimentE
so researchers provided bought hominy chop as a feed
supplement instead. Little weight gain differences were
measured, but it was observed that the stall fed animals
were worked harder.

Paege 27



1. Regional Economi st, CIMMYT Eastern and Southern Africa
Economics Program, Mbabane, Swaziland. The views expressed in
this paper are the author's alone and do not necessarily
represent those of CIMMYT.

2. Complexity may also exist in terms of the range and mix of
farm and non-farm activities carried out by small
farm-households. Such complexity may not be so readily
Apparent in a purely farm context but may still be very
significant in the broader household context, where different
members undertake a wide range of household production tasks.

.3. This theoretical gap has been filled to some extent by the new
household economics theory of consumer choice. See for
example Low, 1982(a), 1982(b), 1982(c).

4. This is not an isolated case. Other studies in Swaziland
(FMS, 1978) and Zimbabwe CIMMYT (1982) demonstrate the same
differences between cattle owners and non-owners.

P'age 28



Becker-, G.5. (1965) "A Theory of the Allocation of: Time" Economic
Journal, 75:493-517

Behnke, R. & Kerven, C. (1983) "FSR; and the Attempt to Understand
the Goals and Motivations of Farmers", Culture and
Agriculture 19:9-16

B~inswanger, H.P. et al (1980) "Rural. Household Studies in Asia",
Singapore University Press.

Bruce, K., D. Byerlee G.E. Edmeades (1980) "Maize in the
Mampong-Sekodumasi Area of Ghana", CIMMYT Working Paper,

Byerlee, D, L. Harrington & D.L. Winkelmann (1982) "Farming
Systems Research: Issues in Research Strategy and
Technology Design" Amer. J. Agr. Econ., 64:997-904

Caldwell,. J.S. (1983) "An Overview of Farming Systemt Research
and Development: Origins, Applications and Issues" in
Proceedings of Kansas State University's 1982 Farming
Systems Research Symposium farming systems in the
field, C.B. Flora, ed. Paper No. 5, Kansas State
University, Manhattan, Kansas.

CIMMYT (1982) "Demonstrations of an Interdisciplinary Approach to
Planni ng Adapti ve Agricultural Research P'rogramm~eE"
Report 45, November 1992,Ministry o Agriculture,
Zimbabwe/Department of Land Management, University of

Collinson, M.P. (1981) "A Low Cost Approach to Understanding
Small Farmers", Agr. Admin. 8:433-50

Evens~on. R.E. et al ( 1980) Nut r it ion Wor k and Demographic
Behaviour in Rural Phillipine Households: a symposium of
several Laguna household studies", in Binswanger, H.P. et
al (1980).

Farm Management Surveys (FMS) (1978) Reports #4 and 5, mimeo,
Monitoring and Evaluation Unit, Rural Development Areas
Programme, Mbabane, Swaziland

HilIderbr and P.E8 (1981) "Combi ning Disciplines in Raprd
Appraisal: the sondeo", Agr. Admin. 8:423-32

Lancaster, K::. (1i966i) "Change and Innovation in the Technology of
Consumption", Amer. Econ. Review/Supplement, May:14-23

P'age 29


Low, A.R.C. (1982a) "A Comparative Advantage Theory of the
Subsistence Farm-Household: Applications to Swazi
Farming", The South African Journal of Economics,

Low, A.R.C. (1982b) "Farm-Household Theory and Rural Development
in Swaziland", Development Study #23, Department of
Agricultural Economics and Management, University of

Low, A.R.C. (1982c) "Agricultural Development in Southern Africa:
A Household Economics Perspective", Ph .D thesis,
Department of Agricultural Economics and Management,
University of Reading

Muth, R.F. (1966) "Household Production and Consumer Demand
Functions", Econometrica, 34:699-708

Norman, D. et al


"Farming Systems in the Nigerian Savanna:
Strategies for Development", Boulder:

Norman, D3. (1983) "Helping Resource Poor Farmers: the
agricult ural technology improvement project, Botswana",
mimeograph, Department of A~gricultural Research, Sebele.

Shumba, E. (1983) "The Crop-
Farmer Adaptation to P
and Lack of Dry Seas
Zimbabwe", CIMMYT te
Swaziland, October 1983.

Li vestock~ Interrelationship and
problemss of Reduced Cattle Numbers
ion Feed in Communal areas of
,chnical net wor kshop E.- i .

Tessema, S. (1983) "Animal Feeding in Small Farm Systems", CIMMYT
technical networkshop, Ezulwini, Swaziland, October

de Wletter, F. (1981) "Report of a Sample Survey", USAID,
Mbabane, Swaziland.

Wat son V. ( 1983) "Farming Systems on Swazi Nation Land: Results
of Extension Field Officer Survey September-November
1982", mimeo, Cropping Systems Research and Extension
Training Project, Malkerns, Swaziland.

Pasge 30,