A FARMING SYSTEMS CONTRIBUTION TO IMPROVED
RELEVANCY IN AGRICULTURAL RESEARCH
CONCEPTS AND PROCEDURES AND THEIR
PROMOTION BY CIMMYT IN EASTERN AFRICA
By: MICHAEL COLLINSON
CIMMYT, BOX 25171,
I Some history of CIMMYT economists adoption of FSR procedures to
improve relevance in agricultural research.
II The FSR procedures developed by CIMMYT for use in adaptive
III Issues in CIMMYT concepts for the use of FSR in adaptive
IV Progress in the promotion and implementation of CIMMYT FSR
procedures in Eastern Africa 1976-80.
V The evolution of strategies for introducing FSR procedures
into research establishments from experience in Eastern
SOME HISTORY OF CIMMYT ECONOMISTS ADOPTION OF FARM SYSTEMS
RESEARCH PROCEDURES TO IMPROVE RELEVANCE IN AGRICULTURAL RESEARCH.
The.idea of small farmers irrationality has been dispelled. Evidence has
accumulated from studies throughout the developing world, that like many
of us, the small farmer does what he does in his own best interest. The
degree of his success varies with the management ability of the individual.
We can readily example the way small farmers have taken to changes which
they perceive to be in their interest. Cotton in Sukumaland, Tanzania is
a good-case in point. Government policy for expanding cotton production
in Tanzania from 1950-1966 was very consistent. Plant early, tie ridge
and use fertilizer and production expanded dramatically --
Table 1. The development of cotton production in Sukumaland, Tanzania.
Year 1950 1966
Production (bales) 39,000 405,000
Approximate total cotton acreage 120,000 1,000,000
Acreage using fertilizer n.a. 28,000
Acreage of cotton per farm .68 3.36
% area in cotton 9 44
Source: IBRD: Studies in Employment and Rural Development, 13, 1974.
but official extension policy, aimed at high yields, played a minor role
in the expansion. After 15 years exposure to the idea of fertilizer use,
less than 3% of the area planted was receiving it. Yet the total crop
expanded tenfold and cotton, as an enterprise, had grown from 9% to 44%
of the area cultivated per farm. A well organised marketing channel was
provided which removed the price uncertainties of selling cotton, and the
Sukuma farmer dovetailed an expanding cotton acreage into his farming
system in a way which complemented his food crop activities. Tea, coffee,
milk in Kenya, Pyrethrum in South Tanzania, Hybrid maize in Kenya, Zambia
The views expressed in this paper do not necessarily represent those of CLMMYT
I thank Steve Franzel and Duncan Bougton for their comments onthe paper.
and Malawi, these are all success stories not because of high yields
from following recommended management practices, but due to high
profitability, complementary with farmers' other production activities
and effective marketing. These programmes succeeded, some of them in
spite of inappropriate research recommendations. However, many programmes
fail due to faulty organisation, others to a poor policy framework, but
perhaps more than is often appreciated due to inappropriate technology
at the heart of the planning. It is these, given the urgency of the
agricultural development problem and the scarcity of resources,
especially qualified manpower, that bring out a need for increased
relevance in agricultural research.
Inappropriate technology generates two types of adoption failure:
1) Non or partial adoption of new management packages offered
as improvements as a result of their ability to raise yield
levels in research programmes.
2) Adoption by particular sections of the farm population
with results which distort policy objectives.
There is widespread experience of non-adoption and partial adoption.
As an East African example Gerhart (1) has shown varying levels of
adoption for the complementary management practices extended with hybrid
maize in Western Kenya, designed to exploit its yield potential. In some
zones with high adoption of hybrid seed, low adoption of timely planting,
pure stands and fertilizer application are found.
There is a widening documentation of the second type of failure where the
introduction of new practices has had a social and economic impact
contrary to government policy objectives. The introduction of the HYV
short strawed wheats increased the income gap between poor and rich farmers
in some parts of the Indian sub-continent;'where only the richer
farmers had access to the irrigation water and cash for fertilizer which
was needed to exploit the potential of the new varieties. Kiray and
Hinderink (2) have described the effects of replacing Yeri, the traditional
cotton varieties in a part of Turkey, with Akala and Deltapine through
a government controlled seed distribution scheme. The new varieties
matured within a three week period this concentrated the harvest labour
requirement and required the employment of casual labour. Only the
wealthy farmers had the cash to hire labour and cotton was lost as a
cash crop to the small farmers. They turned to casual cotton picking
as a source of cash income and found themselves worse off as a result.
Non adoption or distorted adoption may result in a failure to improve
the living standards of the intended clients and implies an ineffective
development effort in four main ways:
1) Scarce manpower and funds have been absorbed in research
on issues which are irrelevant to farmers.
2) Relationships between farmers and extension workers, and
government servants in general, are soured. Morale and
effort decline and create trends in people/government
relationships which are undesirable and difficult to reverse.
3) Where new techniques are incorporated as the content of area
based development programmes, these require investment in
enabling infrastructure and institutions: Roads, vehicles,
supplies and credit are common examples. The anticipated
increase in production requires investment in servicing
institutions for transport, marketing, storage and processing.
With non-adoption such investments are drastically underutilised.
4) Where programme content is irrelevant to the target population
such benefits as are realized go elsewhere, frustrating
government policy objectives.
Thus inefficiencies in research planning and focus result in nonadoption or
distorted adoption and have wide repercussions on agricultural development
In the early 1970's, following adverse commentaries on some aspects
of the Green Revolution, CIMMYT's Economics Programme sponsored studies
in seven countries (3) which sought to identify the factors most influential
in shaping the adoption of new maize and wheat technology. The studies
showed the critical importance of their natural and economic circumstances
to farmers in selecting between alternative technologies for use on their
The dominating influence of economic circusmtances on farmers' decisions
pointed to a role for economists in planning and interpreting agricultural
experiments seeking new techniques for farmers.
Through the 1970's some of the characteristics of experimental methods
and research organisation which give rise to irrelevant results have
crystallised. At the same time an awareness has emerged that small
farmers are different from large farmers in ways which require special
consideration in planning and interpreting research for their situations.
Three features of research organisation and method are at the heart
(1) Biological potential, a perspective too narrow for planning
experiments and evaluating new techniques for farmers.
The exploitation of biological potential persists as a
perspective for the planning and evaluation of experiments. It
gives rise to output per unit of land yield as the dominant
decision criterion in the selection of recommendations for farmers.
Yet farmers never seek biological potential for its own sake.
Sometimes yield is used as an intermediate criterion, but it
is never a sufficeint criterion alone for small farmers to decide
what crop and animal products to produce and what methods to use
in producing them.
(2) The prescriptive tradition in agricultural research.
Historically research has been oriented to large scale farmers
who, because of their influence have often been able to determine
research priorities. Their access and education have allowed
them to select out those experimental results they perceive as
relevant to their own priorities and conditions. Small farmers
do not have this influence, government acts on their behalf, it
decides priority lines of research and which findings will be
promoted. Researchers themselves select out 'the best' results
of experiments and prescribe these as'improved management
practices' for the extension services to teach to farmers. 'I
know what is best for you' very much dominates the process.
(3) The isolation of researchers from the farmers, their clients.
Located on experimental stations researchers are isolated
from the farmers, their clients. Reward systems are such that
researchers have no incentive to understand the farmers they work for.
This mental isolation compounds the prescriptive nature of the.
recommendation process. The physical isolation of station enclave
also promotes irrelevance. Natural and economic conditions there
are often very different from those in which the farmers' have to
operate. For example soil fertility and labour supply are likely
to be very different on an experiment station and on local farms.
Dominance of the inappropriate perspective of biological potential and
the associated decision criterion of yield per unit area is perhaps the
key to irrelevant research. Relevancy to farmers needs and circumstances
demands farmers' perspectives in planning experiments and farmers'
decision criteria in evaluating the results.
A further complicating factor is that farmers' perspectives and decision
criteria are by no means uniform, they change with farmers' circumstances.
Again there are particularly significant differences between the circumstances
of Large Scale Commercial farmers and small holders, but also between
small holders from different areas. These differences reinforce' the
need for new procedures in planning and interpreting experiments for
the traditional agricultural sectors in LDC's. For example smallholders
operating close to the subsistence level, with low resource endowments,
have priorities for a secure food supply. Averseness to risks of food
supply failure influences their decision making giving rise to resource
allocations and management strategies different from those of large
farmers with higher resource endowments. Different technologies are
relevant to these different situations. Again, amongst smallholders
in different situations decision criteria will vary. Take maize. and
beans as a common crop mixture, farmers' decision criteria on the
appropriate balance between maize and beans in the mixture will vary
with their circumstances
Criteria will depend on:
(1) The relative potential of-the two crops in the natural circumstances
of the area.
(2) How the two crops are combined in the staple dishes of the area.
(3) Whether there are profitable market opportunities locally for
Ca) Maize (b) beans (c) both crops Cd) other crops grown
in the system.
It will clearly be impossible to produce relevant improved management
practices for the maize and bean enterprise without understanding how
local circumstances dictate local farmers' decisions in balancing the
The gradual clarification of these issues on the relevancy of research
to farmers is supported by the evidence from the CIMMYT Adoption
Studies. A stronger link between farmers and researchers is needed to
incorporate farmers' decision criteria and circumstances into the
planning and interpretation of experiments to produce relevant recomme-
ndations. Farm Systems Research (FSR) offers a linking mechanism capable
of reflecting local farmers' priorities and circumstances. A set of
procedures for the use of FSR in this context were developed (4) and
have since been promoted to national agricultural research services
by CIMMYT's Regional Economics Programmes. The first of these,
established at the beginning of 1976, was the Eastern African Economics
Programme, based in Nairobi, Kenya.
The implementation of a philosophy of agricultural research which has
relevancy to .farmers problems as its central tenet, and
uses Farm Systems Research as a farmer link to operationalise its
philosophy, requires two major developments in research organisation and
(1) Recognition of a distinction between Applied and Adaptive
(2) Recognition of a coordinating role for a Farm System
Economist in the Adaptive Research cycle.
Figure .1. models the ideal interactions between Applied .and Adaptive
research and Target Groups of farmers. Ideally the organisation of
research parallels these interactions. Applied research, seen as
station based, is seeking new materials and practices which are
technically feasible in the natural conditions of the country or
region. These are additions to an accumulating body of knowledge on
potential management improvements. Adaptive research selects out and
tests components of this body of knowledge identified as potentially
relevant to the needs and circumstances of target groups of farmers.
FIGURE 1. INTERACTIONS BETWEEN APPLIED AND ADAPTIVE RESEARCH.
Diagnosis of priorities
Identification of new
techniques and materials
available and apparently
3. relevant solutions
Body of Knowledge
of new materials
Station based commodity and
solving priority technical
The coordinating role for the Farm Systems Economist is justified
by his ability to understand and reflect the farmer's perspective.
The farmer's role as a decision maker is to allocate his scarce
resources of land, labour and cash between crop, livestock and off-
farm production opportunities in a manner which best satisfies his,
and his family's priorities. This is essentially the economic problem
and the FSE's professional task is to represent the farmer's perspective
as a decision maker. His function is to modify the traditional
perspective of biological potential to understand farmers' decision
criteria and identify how and why, in managing his farm, the farmer
compromises on the optimal technical management of any one enterprise
in order to raise the productivity of the whole system.
Operationally, locally based Adaptive Research teams, normally made
up of an Economist and an Agronomist, but with call on whichever
specialists are appropriate, carry out stages 1 through 3 of the
Adaptive Research cycle and stage 4, the link stage to the Applied
(1) The description and interpretation of the farmers' situations,
the identification of system problems and possible development
(2) The identification of materials and techniques which show
potential for improving farmers' incomes by overcoming problems
or exploiting opportunities in their situation.
(3) Testing those potential improvements selected as relevant and
feasible for target group farmers, undeb the conditions they
will face in production.
(4) Identifying unsolved technical problems and passing these
back as priorities for Applied Research.
These technical problems, identified on farms in the course of adaptive
research as important to farmers' development, form a logical focus
for disciplinary and commodity oriented station research (Stage 5).
The output from this Applied Research is the Body of Knowledge (Stage 6).
Such a model has guided the strategies of CLMMYT's Eastern African
Programme while helping develop the capacity to use cost effective FSR
procedures in improving relevancy in national agricultural research services.
THE FSR PROCEDURES DEVELOPED BY CIMMYT FOR USE IN
ADAPTIVE AGRICULTURAL RESEARCH
The procedures have been designed as low cost with a rapid'turnaround
to suit the scarce manpower situation in most research establishments
in LDC's. The procedures used first identify target groups of
farmers for whom the same research and development effort will be
relevant. An investigative sequence is implemented within target
groups selected as priorities for intensified research and development
efforts. This sequence is concluded with design of a set of experiments
tailored to the identified problems and development opportunities
of the farming system of the target group. The experiments are
implemented under the operating conditions of target group farmers
to ensure recommendations emanating from the work will perform as
tested when in the hands of local farmers.
(1) IDENTIFYING TARGET GROUPS OF FARMERS
Inevitably there has to be a compromise between research for the
particular situation of the individual farmer which is far too
expensive, and for the heterogenous conditions of the country as a
whole, which is far too generalised. A pre-requisite for the use of
FSR procedures in diagnosing farmers' situations is the setting up of
a framework of target groups as a basis for planning research priorities
and identifying adaptive research focii.
Farmers are grouped into relatively homogenous populations on the basis
of their existing farming systems. As a basis for grouping the farming
system has three justifications
(a) The farmer's existing system is a manifestation of a weighted
interaction between his exogenous natural, economic and
cultural circumstances and his own priorities and resource
capabilities; It best reflects the balance of factors
important in identifying homogenous groups of farmers.
(b) Farmers operating the same system have the same researchable
problems and exploitable development opportunities. The same
new technologies will be relevant to the group.
(c) The existing farming system is the starting point for development,
the base onto which productivity improvements have to be grafted.
Grouping farmers on the basis of present activities has two dimensions;
activities alter geographically with changes in the natural and economic
environment and hierachically with changing resource endowments. Both
dimensions are important to present and to potential crop management.
A short questionnaire is developed to collect descriptive information
about farmers in each local Administrative Unit. An example questionnaire
is shown as Figure 2. It seeks to tap the experience of agricultural
staff locally involved in day to day agricultural administration in the
areas to be covered. Foreknowledge of their likely biases is used to
ensure balanced information.
The aim in data interpretation is to separate different farming
systems. The key step in interpretation is to identify sources of
variation which play a significant role in resource allocation in the
farming systems covered. This reduces the collected information to
The cost of defining homogenous target groups, using this methodology,
are low. In terms of professional time three stages can be distinguished.
(a) Preparation : 6 8 mandays.
(i) Developing and testing the questionnaire
(ii) Arranging the programme of visits.
(iii) Preparing background material and maps.
(b) Data collection 6 10 mandays: (Administering a questionnaire
for some 100 enumeration units)
(c) Target Group Identification 8-12 mandays
(i) Tabulation of the collected data (done by clerks)
(ii) Interpretation of the data
(iii) Deriving,describing and a preliminary mapping.
It is important to emphasise that these are preliminary groupings.
As the diagnostic sequence is implemented among these identified target
groups, boundaries between groups will be defined more precisely, and
group characteristics detailed more fully.
FARM SYSTEM ZONING QUESTIONNAIRE. CENTRAL PROVINCE, ZAMBIA
A. ANIMAL KEPT BY 1. THREE MAIN TYPES 1
MOST FARMERS OF ANIMALS KEPT 2
2. IF CATTLE; MAIN 1
PURPOSES FOR 2
B. FOODS GROWN (G) 1. STARCH STAPLES 1
OR BOUGHT (B) BY 2
2. RELISH CROPS TO 1
3. ANIMAL PRODUCTS 1
FOR FOOD Overall
1. NEW CASH CROPS 1 ranking
AND % GROWING
C. MAIN CASH 2. CROP SALES AS 1
SOURCES FOR CASH SOURCE 2
(RANK OVERALL) 3. LIVESTOCK AS A 1
(1st, 2nd,3rd) CASH SOURCE
4. OFF FARM 1
CASH SOURCE 2
1. YEARS CULTIVATED 1
D. LAND USE 2. TYPICAL AREA (HA) 1
METHODS AND 3. MAIN METHODS OF 1
TIME OF LAND PREPARATION
MOST FARMERS _
4. MAIN METHODS OF 1
1. TYPES OF HIRED 1
E. HIRE AND PURCHASE LABOUR & PAYMENT
OF RESOURCES BY 2. WORK DONE BY 1
3. MAIN INPUTS 1
AND % FARMERS
Within selected target groups the adaptive research cycle is established
as described in Figure .1, there are three stages. The first stage,
the main application of FSR method, is the understanding of farmers'
priorities and their circumstances to perceive the rationale of their
management strategies and isolate the problems of system expansion.
It identifies development opportunities for the target group and specifies
relevant adaptive research content.
(2) FARMING SYSTEM DIAGNOSIS
Diagnosis of the Farm System has three steps. The sequence acts like
a sieve, progressively sifting out unimportant facets of the farming
system and irrelevant research issues. Each step funnels the diagnosis
towards key circumstances of target group farmers, key problems of their
farming system, and the identification of appropriate development
opportunities. At the same time steps are increasingly expensive in
terms of professional manpower. Cheaper diagnostic procedures are used
to focus subsequent, more expensive procedures onto key issues.
Step .1. The Use of background information to evaluate Target Group
This first step reviews available secondary information on the natural
and economic circumstances of the target population of farmers. The
objective of the review is to identify management problems posed by
the local circumstances within which target farmers must operate.
The focus of enquiryis consistently that of how these local circumstances
will influence farmers' resource allocation decisions. Important facets,
and some of the implications for farmer management, are listed here.
(1) Natural circumstances
(a) Rainfall amount and reliability
(i) the length and timing of the growing seasons
(ii) uncertainties caused by periods of excess rain or drought.
(b) Seasonal temperatures
(i) As an indicator of growth rates
(ii) Frost incidence as a source of uncertainty to growth.
(c) Soil characteristics and topography
(i) As a guide to possible erosion and soil fertility
(ii) To indicate possible locational advantages and
disadvantages; flooding and the use of residual
moisture are examples.
(d) Pest and disease incidence as a source of uncertainty of crop
(2) Institutional circumstances.
(a) Types and usage of marketing and supply channels.
(b) Types, usage and reliability of food distribution channels.
(c) Existing extension and credit programmes; the types of
programmes, the numbers participating and the types of
(d) Land tenure arrangements.
(e) Farmer groupings; cooperatives whether voluntary, organised,
official or unofficial and their planned and actual functions.
This information describes the available physical infrastructure. On
marketing and government intervention programmes it allows some
assessment of the content and degree of penetration of the local
institutions. Information on land tenure and farmer groupings
describes part of the institutional environment which will influence
farmer decision making. Details on extension and credit programmes are
particularly useful in identifying extension biases. For example, if
the investigators know in advance that 5% of the areas farmers' have
credit for the purchase of grade dairy animals, it gives a basis for
judging the comments of extension staff on the incidence of grade animals
in the area.
(3) Economic circumstances.
(a) Population numbers and density, and the pattern of settlement.
This information gives a useful indication of whether the extensive
margin is likely to compete with the intensive margin for system
expansion. Different settlement patterns have different implications
for access to water and grazing, as well as to advantageous locations
for special crops. Patterns of land use will be closely related to
whether farmers live in villages or on their holdings.
(h) Available acreage and production figures give a prior idea
of what should be found in the area and the relative importance of the
various enterprises making up the Farming System.
(c) Marketed Products.
(i) Volume, trends over years and over the calendar, in
outputs sold and inputs purchased via the market and
(ii) Foods purchased, relative volume, trends are the years
and between seasons.
(iii) Prices, trends over years and over the calendar, and
marketing margins if available.
Information on the workings of the market improves the understanding
of what farmers face in both producing for sale and buying for
consumption and for use on the farm. Trend information helps to
highlight how things are changing in local farming what is coming
in and what is going out, trends are much easier to reinforce than
reverse. Each facet reflects circumstances of local farmers' production
environment and is a potential influence on their decisions on what to
produce and how to produce it. Reviewing secondary sources of information
on farmers' circumstances is a two or three day task. Where secondary
information is limited it can be supplemented by discussion among the
target group with local officials and business men.
The review of background information aids the understanding of the target
group by identifying the features of their production environment expected
to have a strong influence on management practices.
Step 2. The Informal Survey
The pivotal procedure in the sequence of three steps'for understanding
the farmers situation is the Informal Survey. The economist and
biologists talk to farmers of the Target Group, on their farms, over
a period of ten to fifteen days. The economist comes to understand the
farmers' perspective in production and the biological researchers
identify major shortcomings and compromises in management which appear
to under exploit the biological potential of the area. Interaction
between the social and biological scientists identifies new management
practices which would better exploit biological potential and which,
at the same time, would be consistent with the farmers' perspective;
giving better satisfaction of his priorities from within his resource
The informal survey proceeds by unstructured interviews with target
group farmers which are focused by detailed guidelines held by the researchers.
These are best conducted at a time in the season when the crops can be
observed in the field. For the first two or three days discussions
concentrate on describing the farming system as manifest in the enterprise
pattern and agricultural calendar. What are the enterprises farmed, how
are the products used, how are the inputs and outputs of production
spread over the agricultural year. What is the relative size of enter-
prise in terms of resources absorbed. This description of how the Farming
System is manifested on the ground provides the basis for understanding
the farmers priorities and decision criteria as reflected in product
end uses and in the relative importance of enterprises as resource
During the next few days discussion aim to identify constraints on the
expansion of the system and on the improved satisfaction of farmers
priorities. These discussions are concerned to describe in detail
the practices used in managing the major enterprises. Their management
practices reflect farmers' strategies to satisfy their priorities in the
fact of resource constraints and production hazards. At this point the
biologists begins to identify compromises which underexploit biological
potential, and the economist aided by the previous review of background
information, begins to understand the pressures from resource limitations
and production hazards that create these same compromises. Once farmers'
management strategies are understood the biologists and economist
interact to identify new materials and practices which would better exploit
biological potential and at the same time improve the satisfaction of
target group farmers priorities. New materials might include varieties,
fertilisers, pesticides, machinery and equipment and even new enterprises.
New practices could cover any one or any combination of the management
components involved in producing one or more enterprises; seedbed
preparation, time and method of planting, crop arrangement in the field,
weeding timing, frequency and intensity and many others relevant to crop
production. All aspects of housing, breeding, health and feeding management
are relevant to the livestock enterprises.
Better satisfaction of farmer priorities would require that the new
materials or practices improve his capacity to manage local production
hazards, to alleviate his resource limitations or, offr.er strg tough
incentives for the re-allocation of resources within existing limitations.
The third and final set of discussionswith target group farmers aim
to test their attitudes to the management changes identified by this
interaction between the biologists and economist.
The discussions are a recursive learning process. Guidelines setting
out the areas of information which may need to be covered are divided
into 'bite sized' sections. Each section forms a basis for detailed
discussion on that aspect alone. Some farmers will be sufficiently
interested to discuss several such sections, others will have had
enough after one. The researchers interview a farmer on one or more
sections making field notes. At the end of the day, after talking to
perhaps three farmers on various sections of the guidelines field
notes should be re-written and filled out as an aid to absorbing and
evaluating the material obtained. Researchers may interview the same
or different farmers on the same sections of the guidelines and
interaction between researchers at the end of the day is an essential
aid to their understanding local farming. Further farmers are interviewed
on the same sections until researchers are confident that they understand
those aspects of local farming summarised in a section. At this point
each researcher makes detailed notes against each heading of that
section of the guidelines. Gaps in the information are identified
and filled by subsequent interviews. Wherever possible interviews are
carried out in farmers fields with visible evidence of farmers
management before the researchers, the visits during the Informal
Survey represent a real opportunity for researchers to interact with
their clients. Interviews should move across the zone with an initial
set of questions to check that selected respondents fall within the
Target Group. After the second set of discussions with target group
farmers the economist writes a scenario of the farming system which
specifically covers the following facets:
(a) A description of the system covering enterprise pattern,
product end uses and the agricultural calendar.
(b) Farmers priorities and decision criteria.
(c) Factors constraining expansion of the system.
(d) Major management strategies to achieve priorities in the
face of resource limitations and production hazards.
This establishes hypotheses on points (b), (c) and (d) and is used as
a base paper for the interaction between the biologists and the
economist in devising further hypotheses on possible improvements in
materials and methods for the development of the system. The final
four days of discussions aim to verify that these proposed changes
are attractive to local farmers, or to modify proposed changes until
farmers react positively to the ideas put forward. The changes form
the proposed content for a programme of adaptive research, which together
wi.th.the hypotheses on the characteristics of the farming system, is the
output from the informal survey.
Step .3. The Verification Survey
The Informal Survey uses an almost anthropological approach to understand
the farming system of the target group. It is followed up by a formal
sample survey which verifies that the understanding obtained by informal
discussions with target group farmers is indeed valid for the target
population as a whole. The formal sample survey may also seek deeper
understanding and occasionally quantification of key parameters to
improve the quality of experimental planning and interpretation.
(1) Verification The major objective of the Verification Survey
is to ensure that the understanding of the Target Group Farming
System has not been distorted by the informal selection of farmers
with whom its characteristics.have been discussed. The initial facet
for verification is the homogeneity of the Target Group itself. Survey
questions cover the variables by which the Target group has been specified
and demarcated. The data from these questions is tabulated to show the
distribution of sample farmers across these variables ,reasons are
sought for sub-groupings which emerge. Re-definition of the Target
Group may be necessary. The description and understanding of the farming
system are verified by the formal survey.
Particular emphasis is placed on farmers priorities, their decision
criteria for each enterprise, the resource limitations operating,
the relative importance of production hazards and the way these facets
relate to resource allocations, management strategies and production
Finally conclusions reached on the management possibilities for
the improved exploitation of biological potential are tested. The
process of interaction between the biologists and economist has
rejected some management changes because of the demand for resources
they imply or because of the strong clash with presently expressed
farmer priorities by their resource reallocation implications. Where
such changes would offer dramatic opportunities for target group farmers
it is important to test attitudes and verify that the resource reallocations
implied are either infeasible or unacceptable for their present circumstances.
Other management changes have been accepted both by the biologist and
by the economist as content for an adaptive experimental programme. The
implications of these proposed changes are verified as acceptable to
(2) Quantification A major consideration in the development of the
sequence of procedures being promoted has been to keep costs down and
allow a rapid turnaround of information. It is important however to
maintain a flexibility in design to accommodate very specific facets
which might dramatically enhance the effective planning or interpretation
of a research programme. In developing the procedures the quantification
of parameters has been deliberately avoided, the emphasis has been on
understanding the farming system. Normally, if precise quantification
is required to tell us whether changes will not benefit farmers, it
will be a doubtful development opportunity. Those opportunities
significant enough to stimulate a shift in the system will normally
be very clear. If quantative support is needed for the understanding
on which proposals are based, 'back of the envelope' calculations,
using interpolated figures, will give an adequate indication of the
likely benefits arising. Exceptionally the quantification of a limited
number of parameters may add considerably to the design of experiments
or to their interpretation.
Where methods of collecting such quantative data are available and
are consistent with the major methodological goals of low cost and
rapid turnaround ,survey design is flexible enough to incorporate
these. For example, labour peaks are often identified as constraints
on system expansion or on better husbandry methods.: Qantification of
the enterprise-operation combinations contributing to the peaks may
add a further dimension to the planning and interpretation of experiments
on management components designed to alleviate the peak and allow
expansion of the system. Adequate labour data on the identified peak
period can be collected in a single visit to the farm.
(3) Innovators. Many development opportunities will have been
identified and the better farmers in any community will have worked
out their own ways to improve their situation. These should be
identified in the formal survey. Farmers who appear to have found
and developed management strategies which relieve system constraints
should be followed up after the survey with a view to detailing their
solutions which, after experimentation, could be spread across the
(4) Methods. Normal Farm Survey methods are used to mount the
Verification Survey. To allow rapid turnaround it will usually consist
of a single visit of between 1l-2 hours to fifty or sixty target group
farmers. The Target Group has been identified as a homogenous sub-
population and many sources of variation have been externalised in
establishing the Target Group framework. Some thirty respondents
would normally be adequate to reflect the local situation. To allow
for the emergence of sub-groups and the possible need to re-define
the Target Group fifty or sixty sample units will be safer. The
questionnaire is developed from the Informal Survey output and is
completely local specific and highly selective in content. Many of the
focii; for example one establishing farmer priorities, resource limitations
and management strategies, will be common across many target groups
and standard question sequences can be employed to elicit the information
On the other hand the content, that is the enterprises followed and
methods used, will vary between target groups.
As always with a single visit survey it is important to enlist the
support of local leaders and ensure the selected respondents are
participating in an investigation which has the backing of the
farming community. Thorough training of the enumerators in the aim
and delivery of each question,careful pre-testing of the questionnaire,
and a level of supervision which allows careful evaluation of each
completed questionnaire while still in the field, ensures the good
quality of the data Pre-programming of the farm visits and a
careful reconciling of the sampling design with the logistical
demands of the fieldwork will ensure smooth adminsitration of the
There is a value in hand tabulation and analysis of initial surveys
to gain familiarity with the approach to their interpretation. However
with the formal survey as the end of a sequence, and with much of its
content aimed at verification, it lends itself to a high proportion
of pre-coded answers and easy computer processing.
(3) COMPLETING THE ADAPTIVE RESEARCH CYCLE. Stage 2 of the adaptive
research cycle set out in Figure .1. emerges out of the diagnostic
sequence described. In the course of the sequence the biologists
have reviewed the Body of Knowledge as well as applied basic principles
of their discipline, and put forward changes in management which would
better exploit local biological potential. The economist using the
farmers' perspective has evaluated the compatability of the resource
re-allocations implied with the existing farming system.
Final output from the verification survey is the content for a set
of adaptive experiments to test possible management changes designed
to meet the priorities and fit the capabilities of target group farmers.
Study variables in these experiments are those identified management
changes which the biologist believes will improve the exploitation of
local biological potential and which, at the same time, the economist
believes will enhance the satisfaction of farmers' priorities. The
range of levels over which these experimental variables will be tested
is decided by the farmers' flexibility in managing the resources required
to.make the changes. For example, where it is decided that animal manure
is cheaply available to local farmers, and the best base for fertility
maintenance, the amount to apply and the timing of the application might
be the two major study variables.
The levels of farmyard manure to be tested will be dictated by the
typical availability of manure to target group farmers. The treatments
on timing of application will be decided by other factors, for example;
when during the rotation the manure is to be applied, when is labour
readily available during the cropping year as the carrying and applica-
tion of manure are labour intensive, and the quality,and therefore
best timing for useof the manure under present methods of storage.
Information on these facets is fed into research planning from the
survey Once the use of readily available manure emerges as a relevant
experimental variable in the course of the informal survey, further
discussions with farmers will focus on these and other facets required
for relevant experimentation on manure. All such facets will be
verified in the formal survey. Equally important to the planning of
relevant experiments is the selection of levels of non-study variables.
In seeking the expression of biological potential in traditional experi-
mentation non-study variables are often fixed at levels which will not
inhibit the effects of the study variables; ad lib insect control,
ad lib labour for weeding, optimal time and method of planting are some
common examples. Clearly these may be completely irrelevant to the
target group farmers' situation. The study variables have been selected
with the farmers' ability to change their management as a paramount
consideration. Other variables have been rejected because he is not
flexible enough to modify their management. Clearly, the experiments
must show whether productivity will increase if the changes are introduced
within the farmers' present management regime. The parts of this regime the
diagnostic sequence has identified as inflexible,are simulated in the
experiments to ensure the study management variables are tested in
the context in which they will have to perform when tried by farmers.
The control treatment in the experiment will be as close as possible
to farmers' existing management, the only relevant baseline against
which improvements in performance, from the changes proposed in the
management of the study variables, can be measured and evaluated.
The farmer management levels adopted for the non-study variables
and for the control treatment are also derived from the survey
information. Stage 3 of the cycle is the implementation of the
experiments as close as possible to the conditions under which
farmers have to operate. The survey provides information on site
characteristics and cropping history to further ensure representation
of farmers' conditions. Similarily the understanding of the farming
system, forms a context for the interpretation of the experimental results.
Each management change implies a reallocation of resources within the
system. Each re-allocation has costs and.benefits, often in both
market and non-market terms, which are peculiar to the particular
system being researched. The understanding of farmer priorities
and resource constraints given by the survey sequence allows the
identification of the costs and benefits arising from changes in
management implied in the experimental results. It ensures that
the interpretation of the experimental results wll be relevant to
the specific situation of the Target Group under investigation.
SOME ISSUES IN CIMMYT'S CONCEPTS AND PROCEDURES FOR THE USE OF FSR
IN AGRICULTURAL RESEARCH AND PLANNING
This section explores five issues in the use of low cost FSR procedures for
planning agricultural research. The first one touches on the broader
aspects of agricultural planning and policy.
(1) THE FRAMEWORK OF TARGET GROUPS AS A LINK BETWEEN NATIONAL
AND LOCAL PRIORITIES.
A framework of Target Groups of farmers provides an interface between national
and local priorities. Use of the framework as a planning link allows a marriage
of a top down flow of policy objectives and of bottom up local needs into relevant,
and therefore more effective, research.and development programmes. Figure .3.
sets out-a common pattern for deriving agricultural research and development
(1) National policy objectives dictate a commodity as a policy vehicle.
(2) Areas are selected with suitable natural conditions, sometimes where
the commodity is already produced.
(3) Areas are selected which already have infrastructure and institutions
onto which programme enabling and marketing services can readily be grafted.
Altematively the necessary infrastructure and institutions form part of the
(4) A package of practices, traditionally aimed at optimaltechnical management
of the commodity under the climate and soil conditions of the chosen area,
is specified as programme content.
Farm System concepts wirn that this pattern lacks a vital element. It fails
to weigh the priorities and social and economic circumstances of local farmers
in decisions on programme content,and this may be crucial to programme
performance. Farmers priorities are manifested in the resource allocation
of their existing system which may be incompatible with the resource require-
ments of programme content. Either the commodity vehicle or the management
practices recommended for its production may conflict strongly enough with
existing priorities and resource allocations consequent on these, to be
unacceptable to local farmers.
Information from the zoning questionnaire and from secondary sources on
local farmers' existing system and circumstances is used to build a frame-
work of Target Groups. It permits an evaluation of each Target Group as
a basis for the location of programmes selected also for their relevance
to national priorities.
FIGURE 3. DERIVATION OF PROGRAMMES FOR AGRICULTURAL RESEARCH AND DEVELOPMENT
NATIONAL POLICY PRIORITIES.
IN RESEARCH AND
AREAS WITH SUITABLE
OF CLIMATE & SOIL
AREAS WITH SUITABLE
OR MODIFIABLE INFRASTRUCTURE
Once the. survey- sequence: is completed for all the target groups in the
framework,its value in fitting programmes, selected to meet policy
objectives,to appropriate Target Groups is greatly enhanced. Detailed
research and development programme content, screened for relevance to
Target Group farmers' priorities and for compatibility with the existing
resource allocations these require, can be specified. In addition
understanding the system involves understanding its relationships with
the local production environment and the survey sequence also helps specify
the complementary infrastructural, institutional and policy interventions
needed to mobilise and service the programme.
Table 2 shows a hypothetical framework of Target Groups. The tabulated
information is- from available secondary sources and a zoning questionnaire.
It will gradually be supplemented by diagnostic survey reports as adaptive
research coverage of each Target Group is achieved. However even this
preliminary information is a useful aid to the decision maker. In this
hypothetical example the framework set out in Table 2 is used to demonstrate.
the selection of priority Target Groups for an Adaptive Research Programme.
Selection is based on the relative potential of the Target Groups to
satisfy a variety of a national policy objectives. The weighting of
the various objectives remains the problem of the decision maker. Knowing
the areas have very much the same climate and soils our hypothetical decision
maker would weigh several policy guidelines in reviewing this information
on the six target groups.
(1) The number of farmers to be reached by priority adaptive research
programmes should be as high as possible.
(2) Infrastructural development and institutional penetration cannot
be made cost effective, except in extraordinary circumstances,
under 25 persons per square kilometre.
(3) Urban population is growing at 10% per annum requiring rapid
increases in marketed maize for urban food.
(4) The newly established textile industry needs cotton as raw
material -much of which is presently being imported at a high
cost in terms of foreign exchange.
TABLE 2 The framework- of Target Group information as an aid in deciding
Adaptive Research priorities; a hypothetical example.
TARGET GROUP NO. 1 2 3 4 5 6
Number of farmers (000) 12 6 8 4 15 20
Population p. square Km. 30- 10 15 40 120 40
Roads and transport. weak weak weak good good fair
Extension staff number. 5 2 2 5 18 15
staff/farmer ratio. 2400 3000 4000 800 750 1300
% farmers participate, neg neg neg 5 10 3
Buying points : number. 6 2 4 5 30 20
Point/farmer ratio. 2000 3000 2000 800 450 1000
Z farmers' using. 3 neg neg 50 90 30
Credit facilities % using Neg neg neg 3 8 1
Main Starch Staple Millet Sorghum Millet Maize Maize Maize
Main Cash Sources 1 Beer Fish Beer Maize Cotton Maize
2 Beer Meat Meat Maize Beer
Present cash income status. low low fair fair good low
Area cultivated p. farm (ha) 1.0 1.0 1.0 2.0 3.0 2.0
Method of land preparation. hoe hoe ox ox hoe ox
Main planting month. Jan Jan Jan Dec Jan Jan
Main reasons for 1 Scurity (none) Draft Draft (none) Draft
cattle keeping. 2 Sales Sales Security
Peak labour 1 Dec Dec June Feb Dec Feb
Months 2 June June -
Peak labour 1 Dig Dig Pick Weed Dig Weed
Work 2 Pick Pick
His decisions would emerge from the following: Adaptive research efforts
for Target groups 2 and 3 are low priority, both are relatively small and
neither have the population density to allow the development of a viable
infrastructure and marketing system. Target groups 1, 4 and 6 offer
prospects of rapid development at low cost, expanding the farming systems
towards the extensive margin Both Target groups 6 and 1 cover large
numbers of farmers. Target groups 6 and 4 are very similar, 4 has better.
infrastructure and the sale of livestock gives higher cash incomes. For
crop development the same adaptive research effort is likely to be
relevant. All three Target Groups have a relatively low population
density and a low area cultivated per household, probably due to restricted
market contact. With draft animals already mobilised in 4 and 6, which have
the capacity to prepare of the order of 5 hectares per season per house-
hold,rapid expansion of the cultivated area can be achieved. Progress
will be slower in Target Group 1 with a poor infrastructure, no current
cash crop and no current use of animals for draft. Target group 5 farmers
have a relatively high income level from two well established cash crops
and an infrastructure developed to exploit these. However high population
density and high proportion of land used indicates a long term soil
fertility problem unless rotational and-manure regimes are working within
Decisions hypothesised are: Immediate priority is for an adaptive
research programme in Target Groups 4 and 6, it will focus on;
(1) The introduction of cotton as a new enterprise
(2) Possible interactions between the present system,
particularly the maize enterprise, and the pattern of resource
requirements to grow cotton.
(3) Maize and cotton management methods with particular concern
to alleviate the existing weeding problem. This may have
repercussions on relative time of planting and variety selection.
(4) Alternative methods of weed management, including the use
If recommendations are generated from the Adaptive Research effort
a development programme will need to provide on an increased density
of buying points to give better access to market for local farmers.
Enhanced production will render a higher density of points viable.
Target group 6 will also need more intensive extension coverage and
improved access roads.
A second priority for an adaptive research effort would be as part of a
development effort in Target Group one. Plans for an improved infrastructure
and marketing system and more intensive extension coverage are a pre-
requisite for an adaptive research effort with this high a priority. Initial
extension effort on the mobilisation of local animals for draught purposes
would allow encourage rapid development of the system towards the extensive
margin which, is usually more easily managed and with only 20% of the land
area utilised will almost certainly be the more profitable alternative.
With a clash between millet and cotton labour requirements for harvesting
the initial focus for an adaptive research effort will be the introduction
of maize as a joint food and cash crop.
Third priority for an adaptive research effort is intensification of the
production system in Target Group 5. This would improve to second priority
in the absence of resources for a programme for infrastructure and market
development in Target Group 1. Given land preparation as a factor constraining
system development the survey phase of an adaptive research programme
will focus on management practices and varietal types to give improved
complementarity between the resource requirements of cotton and maize
as the two major resource absorbers, over the land preparation, planting
and weeding period. Also, given that 60% of the available land is cultivated,
in the absence of an animal enterprise regimes for chemical fertilisers
will be important. The relatively high levels of cash income and the lack
of extra land to move to the extensive margin indicate that cash outlays
for additional intensifying purchased inputs will be made available 5y farmers.
More certain conclusions on this would follow from the diagnostic survey
sequence in the Adaptive Research Programme. This information from background
sources and the zoning questionnaire can provide a description of local
circumstances and outline characteristics of the farming system of each
Target Group. The description is enough to allow vital local social
and economic factors to weight decisions on what policy objectives can
be achieved where, given the present circumstances of the target groups.
It gives an improved basis for the selection of broad programme content.
(2) CAN RESEARCH RESOURCES HANDLE THE DIVERSITY OF TARGET GROUPS
LIKELY TO EMERGE ?
One important use of Target Groups as a planning framework is in ranking
priorities for research attention. Not all Target Groups can be researched
simultaneously, resources wont normally be adequate. If we make assumptions
about the diversity of Target Groups and the rates of work of Farm Systems
Researchers and Agronomists we can demonstrate hypothetical rates of
coverage of small holder agricultural sectors in LDC's.
First, how many groups do we expect in the agricultural sector and what
resources are available to handle this number. The only large scale zoning
exercise done on a farming system Basis raised an average of 10,000 farms
per Target Group (5) Given a small farm population, for Kenya as an
example of 1.5 million farms, this implies some 150 Target Groups.
Continuing to use the Kenya case as an example, there are presently
twelve graduate Farming System Economists in the research services.
Marry these with twelve Agronomists as adaptive research teams to carry
out local specific survey and experimentation and make the following
assumptions about work load:
Farm Systems: Economists : With commitments to join agronomists and farmers
at experimental sites, and in the economic interpretation of experimental
results, two diagnostic surveys of new Target Groups are feasible per
professional per year.
Agronomists: Six sites are required for each target group and an
average of 2 years work are required to arrive at farmer recommendations
for each group. Each professional can supervise 18 sites, that is
3 Target Groups, each year.
Under these ass-umptions the completion of 2 years adaptive research
in all 150 target populations, countrywide, would take 9 years.
Priorities would decide which areas were covered first.
Projected coverage of 150 Target Groups by
12 Adaptive Research Teams.
YEAR 1 2 3 4 5 6 7 8 9
FSR'S NEW RD'S 24 24 24 24 24 24 6 -
COVERAGE 24 48 .72 96 120 144 150 -
AGRONOMISTS NEW RD'S 24 12 24 12 24 12 24 12 6
COVERAGE 24 36 60 72 96 108 132 144 150
The capacity of the Agronomists limits the speed
Research Programmes cover the country.
at which Adaptive
If a hypothetical 150 Target Groups with six sites required among each
are accepted as constant, Table 4 below shows the years required to
cover the sector if the number of sites supervised per agronomist and
years of work required at the site are varied.
The range of years to full sector coverage with variation in the
Number of sites supervised per agronomist
12 18 24 30
1.0 6.3 4.2 3.1 2.5
Average years 1.5 9.4 6.3 4.7 3.7
at each site. 2.0 12.6 8.4 6.3 5.0
2.5 15.8 10.5 7.9 6.3
3.0 18.9 12.6 9.5 7.5
Within a long term effort to re-orient the focus of agricultural
research to the problems of small farmers, nine years is a modest
period. If priority is given to Target Groups containing large numbers
of farmers the majority of the small farm population will be covered
much more rapidly. At the same time an assumed allocation of some 24
professionals to this Adaptive Research Programme is relatively modest
in terms of the numbers in the professional research establishments in
many LDC's. Given the stored 'Body of Knowledge' from Applied Research
results and the burgeoning role of the International Agricultural
Research Centres in Applied Research, the adaptation of knowledge to
local situations would seem a logical priority for national agriucltural
(3) THE ROLE OF THE FARM SYSTEM ECONOMISTS VIS-A-VIS OTHER SCIENTISTS
IN AGRICULTURAL RESEARCH.
Economics has been described as the science of allocating scarce
resources between competing ends to maximise utilities. The farmer
as a manager, and therefore a decision maker on resource allocation,
exactly parallels this definition. It is this congruity between
the perspective of the farm system economist and the farmer which
justifies a unique, coordinating role for the farm system economist
in planning agricultural research. The economists role vis-a--vis
other technical and social scientists in agricultural research is well
identified by the relationship their specialisations have with farmers'
decisions. The farmer coordinates and weights technical, social and
cultural factors in taking his management decisions. Biologists and
engineers working with technical factors, supplemented by sociologists
working with social factors and anthropologists working with cultural
factors make partial contributions to the decision process for research-
planning. These partial contributions are coordinated and weighted
by the farm system economist, his perspective of their relevance
representing that of the farmer.
The comparative advantage of disciplinary specialisation is only
effectively mobilised in interdisciplinary research. Where research.
organisation is compartmentalised each discipline is isolated and blinkered,
each is convinced that its own answers are those needed by farmers.
Interdisciplinary organisation, focused on priority farmer problems,
can Bring a synergistic effect to the research process.
(4) IF WE NEED TO UNDERSTAND A FARMING SYSTEM TO MODEL IT, WHAT DO
WE GAIN FROM MODELLING?
The procedures promoted by CIMMYT have been designed with the LDC
institutional situation in mind; limited funds and limited manpower.
The procedures allow the understanding of the system of a Target Group
of farmers over a 3 or 4 month period. Costs are low and turn around
time is rapid. A major controversy among micro-economists working in
LDC agriculture centres around the intensity of investigation and
analysis required to understand systems. A spectrum of data collection
methods is available; from single visit surveys to cost route methods
in which farmers are visited regularly over the agricultural calendar.
Similarly a spectrum of analytical tools is available with simple
cross tabulations and sophisticated variants of mathematical programming
as extremes. Thinking of extremes the CIMMYT procedures are close to
the rapid ahd cheap end of the collection and analysis spectra, with
a turn around time of 3-4 months for any one target group of farmers,
compared to 18-30 months for frequent visit data collection and programming
analysis. Intensive investigation is vastly more expensive, particularly
in opportunity cost terms.
Cheap and rapid procedures allow an understanding of between 4 and 10
target groups over the time period used to study a single target
group by very intensive methods. Where research manpower is scarce it
is assumed that the benefits from an understanding of 4-10 target groups
are very much higher than from the increased understanding of one target
group through detailed data collection and the use of sophisticated
modelling techniques for system analysis. Further, mathematical programming
cannot in practice effectively represent real life situations and the level
of control of enumerator and respondent error in cost route data collection
techniques dependent on interview is usually very poor Our conclusion
is that the sophisticated ends of the spectra of collection and analytical
methods do not offer effective operational procedures.
The understanding of a system has to be so sound to model it and to
allow a realistic interpretation of the txesults, that the modelling
itself becomes superfluous. The original understanding does the
job in a way best fitted to the manpower resources of LDC's. Cost/
effectiveness and this assessment of the existing state of the arts are
the rationale behind the CIMMYT procedures which mobilise an almost
anthropological approach to understanding farmers and their systems
(5) FARMING SYSTEMS RESEARCH: CARTE BLANCHE OR PRE-DETERMINED FOCUS ?
The issue is whether, in trying to identify farmers' needs, the
diagnostic survey sequence should or should not be pre-focussed onto
It is a major issue. There has been much criticism of the blinkered
approach in traditional experimentation due to compartmentalised research
establishments and heavy disciplinary specialisation. The breeder
believes variety is the main problem, the soil scientist fertility,
the entomologist insects and so on. The farmer will have a list of
problems, many interrelated. If research effortaddresses those well
down on his list, missing say his top three, he will regard such
changes as irrelevant to his priority needs and will be unwilling to
commit his limited resources to their adoption. The question is does
a systems approach which is pre-focussed onto one facet of the system
fall into this same trap of treating what may, for the farmer, be
secondary issues. Should the initial diagnosis of the system be completely
without prejudice as to what the key problems might be? Even with a
whole system orientation investigation does have to home in on the
problem area which is crucial to the development of the system. It will
almost always be found that one enterprise offers the best leverage on
that problem and will form the focus for experimentation and development
But investigation from a whole system orientation may focus research on
a different enterprise than investigation of a pre-determined enterprise.
Clearly, ideally, the investigation and planning process should be
initiated carte blanche with a whole system orientation and the intention
to focus research on key problems and the enterprise,or enterprises
offer best leverage on those problems.
In CIMMYT our mandate for work on maize and wheat demands compromise
to a pre-identified commodity approach. However, the funnelling process
in investigation, from general to particular, allows confirmation at
the Informal Survey stage, before costs are significant, that the
commodity pre-focus is consistent with effective leverage on the
system as a whole. It also has some operational advantages:
(1) The pre-determined enterprise orientation is particularly
feasible when research is planned in a region (across several
TG's) where the. enterprise in question is the major resource
absorber in the system. This enterprise usually offers the.
best leverage on system problems. This is particularly true
for starch staples in subsistence and semi-subsistence farming
(2) Many national research efforts are commodity oriented. The
pre-determined enterprise approach allows easier introduction
of a farming systems based research planning sequence into
existing research establishments. A whole system orientation
implies a radical re-organisation and the abandoning of commodity
programmes; too much for existing establishments to swallow.
(3) Many important policy objectives are commodity based and readily
related to commodity based research programmes.
PROGRESS IN THE PROMOTION AND IMPLEMENTATION OF CIMMYT FSR
PROCEDURES IN EASTERN AFRICA
(1) OVERALL PROGRAMME STRATEGY
The programme has followed a two stage strategy. Where national level
agricultural research administrators have shown an interest in the need
to bring research closer to the small farmer, a demonstration of the
procedures is mounted in a maize or wheat growing area of their choice.
Usually an area is chosen for which there are no recommendations, or
where administrators feel the current recommendations are not being accepted
by farmers. These initial demonstrations are organised, supervised and
reported by CIMMYT but involve country professionals in both biological
sciences and economics. In these demonstrations the CIMMYT procedures are
implemented within a homogenous target group to understand their farming
system. An adaptive research programme, tailored to produce new techniques
to solve problems and exploit opportunities in the system, is drawn up as
a conclusion of the demonstration.
The initial demonstration in each interested country is fully written
up as a formal report to promote interest in the procedures on the part
of research adminsitrators and country professionals, and to provide
an example of how the procedures operate to interested professionals
elsewhere. Although CIMMYT usually carries a major part of the costs
of these initial demonstrations it has been deliberate policy that
subsequent activities be funded nationally to try to ensure that further
cooperation is more than a gesture. The Eastern African Economics
Programme annual budget of US$ 125,000 is regarded as 'priming' for
a national commitment to financing the use of the concepts and procedures
The second stage of the strategy follows where the demonstration
convinces national agricultural research administrators that the
procedures can improve the relevance and efficiency of the research
effort. Where adminsitrators are sufficiently convinced to recruit
graduate staff as trainee Farm System Economists, CIMMYT has helped
to build up their professional capacity to implement the procedures
within the research services.
CIMMYT's in service training input to develop this capacity has been
supplemented by further formal training. Where overseas universities
have been used, masters degree training has been funded by a parallel:.
Ford Foundation programme.
(2) PROGRESS WITH THE PROMOTION AND DEMONSTRATION OF THE. CIMYT
PROCEDURES IN EASTERN AFRICA; 1976 -1980.
During 1976 discussions were held with senior agricultural policy makers
and research administrators in Ethiopia, Uganda and Tanzania, as well
as in Kenya.
In Ethiopia senior research administrators showed an awareness that
experimental results may be irrelevant to many farmers. Their concern
had already manifested itself in the establishment of a Socio-Economics
Unit in the Institite of Agricultural Research. There was interest
in testing CIMMYT's procedures to promote coordination between the
long established biologists and the newly introduced social scientists.
In Uganda there was an awareness of the problems of relevancy in the
Faculty of Agricutlure, Makerere University and in the Planning Unit
of the Ministry of Agriculture. In both institutions economists were
interested in procedures for identifying farmers' problems. However
there was no concern over the issue of relevancy within the Research
Division of the Ministry of Agriculture and this, together with the
unstable political situation, deterred us from embarking on field
activities in Uganda. A proposal on the procedures being promoted
by the Programme, and on the concepts behind them, was put to the
National Crop Research Planning Committee in Tanzania in October 1976,
as a result of interest in the Programme on the part of senior research
administrators. The Committee gave CIMMYT permission to approach
Directors of Research Stations and to seek their interest in a demonstration
of procedures to improve relevancy in research. Two Directors were
approached';of the Central Research Station, Ilonga, and Uyole Agricultural
Centre, Mbeya. Both expressed concern with the problem and interest in
procedures to alleviate it.
The adoption study by Gerhart supported by CIMMYT and done in Kenya (1)
had heightened awareness of the relevancy problem among both policy
makers and research administrators.
A strong political emphasis on increasing the prosperity of the rural
population had perhaps begun to expose shortcomings in the application
of research results. Agricultural administrators were actively seeking
a re-orientation of research and development efforts to improve the
realisation of political aspirations. There was interest in the CIMMYT
procedures as a means to this end. Contact was made with both Zambia
and Malawi in the second half of 1977. In Malawi the Research Division
of the Ministry of Agriculture and Natural Resources was in the middle of
re-organisation. There was the belief among senior researchers that
experimentation was always done in the interests of the farmer and the
programme was not pressed further. In Zambia there was strong political
concern that the rural population should be drawn into the development
process, presently centred around copper and the large commercial farm
sector. Administrators were conscious that the mass of small farmers
operate under such different circumstances that a new orientation would
be necessary. There was interest in the set of CIMMYT procedures as
a contribution to a re-orientation. Recently, in September 1980,
interest has been expressed by the agricultural research services of
Zimbabwe in a demonstration of the procedures.
In Ethiopia the area chosen for demonstration of the procedures
was around Bako, 250 kilometres West of Addis Ababa and adjacent to
a major research station. With agricultural economists established
in the Institute of Agricultural Research (IAR), and one posted at
Bako, staff from there with. the addition of the Head of the Socio
Economics Section, IAR,and a Farm Economist from the College of
Agriculture, University of Ethiopia, at Debre Zeit, participated
in the demonstration of procedures. With the first two stages of
the investigation completed travel in the countryside became difficult
with the organisation of the peasant milita and the war against Somalia.
Cimmyt was asked to postpone the final stage of the demonstration of
procedures. A Food and Agricultural Organisation (FAO) Farm Economist
was appointed to IAR at about this time and once communications eased,
continued with. the implementation of a programme based on the concepts
behind the CLMMYT procedures. In Tanzania the interest of Station Director's
at Ilonga, in Kilosa District, Central Tanzania and at Uyole Agricultural
Centre, Mbeya,promoted demonstrations of procedures in both areas.
Economists were already established in the Research Division, and in
Mbeya, station economists and biologists were brought together for
the demonstration. In Ilonga economists from the Faculty of Agriculture
at the University of Dar-es-Salaam were teamed with biologists from the
Station. Reports were prepared for both demonstrations. These set out
an experimental programme for maize tailored to the needs and circumstances
of the target groups of farmers. The Report on the Ilonga demonstration
was presented to the National Crop Research Committee in October 1978.
It was well received by the committee and was given detailed consideration,
as a means of improving the linkage between farmers and research at a
workshop arranged to discuss the re-organisation of agricultural research
in Tanzania in March 1979. The procedures and concepts behind them
have been taken up in the Faculty of Agriculture, University of Dar-es-
Salaam, and at Uyole Agricultural Centre, Mbeya, a zonal center with
research responsibilities throughout the Southern Highland areas of
CIMMYT's Eastern African Economics Programme is headquartered in Kenya
where procedures were demonstrated in two areas in 1976; one in Siaya
District on the shores of Lake Victoria, the other in Kwale District
on the Coast. At this time Kenya had no farm economists in her Scientific
Research Division of the Ministry of Agriculture and these two demonstrations
brought together economists from Egerton College, and biologists from
the research services. The Siaya demonstration was written up as a formal
report and was well received by senior research administrators in 1977.
A one day seminar for senior research staff, held at the Faculty of
Agriculture, University of Nairobi in June 1977 provided the opportunity
to discuss the procedures and the concepts behind 'hem.
As a result of the interest among senior agricultural administrators
in Zambia a demonstration of procedures was implemented near Serenje,
a small town In Central Province 400 kms north of Lusaka. The demonstration
brought an economist from the Rural Development Studies Bureau, a research
unit attached to the University of Zambia, together with biological
scientists from the Central Research Station, Mount Makulu. The field
work was carried out in the first half of 1978 and the final report was
accepted by a meeting of the Programme Steering Committee in December
1978. The Committee charged CIMMYT with responsibilities for development
of a capacity to implement the procedures within the research services.
(3) PROGRESSS INLBUILDING CAPACITY TO USE CIMMYT PROCEDURES
CIMMYT is helping to build up capacity to use FSR procedures in
planning and1 interpreting adaptive research in Kenya and Zambia. The
Scientific Research Division, Ministry of Agriculture, Government
of Kenya, established thirteen posts for Farm Systems Economists,
including a 'enior professional as Economic Adviser to the Director
of Research, by October 1979. Agricultural graduates were recruited
to these posts of whom four had completed Masters Degrees and one
is a Ph.D, candidate in Agricultural Economics. Six of the seven
graduates recruited with Bachelors degrees have started Masters
training and four of these have completed requirements. Since under-
taking the training commitment in July 1978 six week long workshops,
the more recent ones for agronomists as well as economists, have been
undertaken jointly by CIMMYT and the Scientific Research Division
in the Kenyan Ministry of Agriculture.Regular visits are paid to
research stations to help trainees plan their work and to monitor its
progress. The first aim of the training and work programmes has been to
build up the capacity of the Farm Economists to understand farmers'
present management strategies in term of their priorities and constraints
which also restrict the changes in management which are relevant and
feasible for farmers. On the whole good progress has been made with
this. The best trainees have a grasp of the spectrum of methods
available for data collection and analysis and a good idea of the
relative suitability of the alternative methods for various situations.
Most trainees are now capable of making a contribution to planning
experiments relevant to an identified target group of farmers and of
interpreting the results in the light of the circumstances of that group.
Some of the Farm Economists have been able to use their knowledge of
farmers' situations, gained from survey work, to propose specific
adaptive experiments directed to relieving major constraints in the
systems of identified target groups of farmers.
The second aim of the training and work programmes has been to build
up the credibility of the Farm Economist with the biological researcher.
Here the programme has met with limited success. The failure to
establish a strong interaction seems to be related to two sets of factors:
(1) A lack of interest in relevance to farmers problems on the
part of the biological researchers.
(2) Features of research organisation which inhibit both
the effective use of Farm Economists and a strong interaction
These factors are discussed further in describing the evolution of
strategies for implementation of the procedures in the final section
of the paper.
In Zambia the programme Steering Committee directed that two Farm System
Economists be recruited and two new graduates joined the research
services in this capacity in July 1979. CIMMYT has been responsible for
helping to plan and supervise their work programmes and has sponsored
their participation in the Workshop Programme mounted in Kenya. Planning
is underway in Zambia to adopt a two level hierachy for Applied and
Adaptive research, to increase the establishment of Farm System Economists
and pair them with Farm System Agronomists in Adaptive Research Teams
to be based locally across the country. It is proposed that the coordinator
of these teams will be a Senior Farming Systems Economist under the
Assistant Director of Agriculture (Research).
STHE EVOLUTION OF STRATEGIES FOR INTRODUCING FSR PROCEDURES INTO
RESEARCH ESTABLISHMENTS FROM EXPERIENCE IN EASTERN AFRICA.
In this last section I trace the evolution in our thinking on ways of
introducing FSR procedures for improving relevancy into research
establishments. Kenya was the first country to ask for help in
building up a capacity to apply the procedures devised by CIMMYT
for planning relevant experiments.
(1) INITIAL STRATEGIES to introduce the procedures into the
Kenya agricultural research services were based on experience
elsewhere, as well as on the organisational implications of the
Experiences elsewhere had emphasised two points: First, historically
economists interested in agricultural technology had dwelt on
exposte criticism of agricultural researchers; coming back after
the event and telling biologists they had got it wrong. Not only
was this not constructive but it also built up an antipathy in research
establishments to these 'commentators'. In the light of this sort
of background it was felt important to emphasise a positive approach,
presenting the use of FSR as a development in research procedures,
not as a remedy for its failings.
Secondly, in the past economists attached to agricultural research
establishments had been isolated in their own units or section.
With the essentially interdisciplinary nature of the procedures
such isolation was seen as anathema to effective implementation.
Two strategies adopted to forestall isolation were; to avoid setting
up separate Economics Units on research stations and to establish a
direct professional link with senior research administrators. Past
experience had seen research station based economists passing their
operational and methodological problems through three levels of the
technical establishment for decision. At each level an attitude of
'what is this odd fellow doing in agricultural research anyway'
often brought decisions inimical to his effective operation.
A senior economist close to the Director of Research, to handle
professional queries in liaison with the Director, was seen as
a necessary counter to the scepticism of administrators with a wholly
The procedures have very strong organisational implications. The need
to examine farming situations, to decide what experimental content
is relevant to the needs of farmers and within their capacities to
absorb, and to provide the context for proper interpretation of the
results, has uncompromising organisational implications.
(1) Adaptive experiments can only be done effectively with a particular
farm situation and thus a particular target group of farmers, in mind.
(2) Disciplinary based organisation is not conducive to effective
adaptive experimentation. A compromise is required on specialist
technical viewpoints if research is to address farmer's needs.
(3) To form the basis of good recommendations to farmers, experimental
results must be achieved under conditions as close as possible to those
under which the farmer will implement them.
The full implications, presented cold, could be unacceptable
to an existing establishment. The new procedures were treated as a set of
innovative components with the re-organisational implications to be
introduced one by one as the programme developed, the easiest first.
In Kenya major research stations have national responsibilities on a
crop basis, smaller stations tend to have zonal responsibilities.
Stations are organised internally on disciplinary lines. In mid 1978
the move began to introduce economists on to research stations avoiding
the establishment of 'economics units'. The following description of
the initial strategies followed is from the working paper, written
in 1978, to guide implementation.efforts.
" Initial objectives will be:
(1) To teach the FSE the tools of his trade and expose him to
(2) For the FSE to gain credibility with the technical scientists
on research stations that he can improve the relevancy of the
experimental programmes and research recommendations to farmers'
needs and to national priorities.
(3) To foster the ideas of target group focused experimentation
and farmer participation in the research process.
The work programme strategy will depend on the local situation. So far
in Eastern Africa technical scientists have little or no idea of how a
FSE can help the relevancy of their work. Indeed relevancy, except
as 'a good thing' is a hazy concept, a prescriptive mentality often still
dominates technical research work. On the whole strategies can be divided
positive and negative. Positive strategies should normally be the basis
of the effort made at research stations. Unless used carefully negative
strategies will enhance the barriers between technical and economic
researchers. However, if used carefully, negative strategies essentially
criticism of past or present work can increase the awareness of the
technical scientist of what is and what is not relevant. There strategies,
with. emphasis to be placed where possible on the first, positive one are
(1) To interest individual technical scientists on stations in the
circumstances and priorities of farmers who will be offered the results
of their research. The FSE begins to feed information on priorities
and circumstances of target group farmers to selected individual
scientists outlining the implications for experimental content.
Technical scientists who appreciate the need for relevancy and
express on interest in co-operation are the ideal.
(2) To review past research results and, by economic interpretation,
demonstrate how modified recommendations are more consistent
with farmer circumstances. Though a critical strategy, it has
proved positive in practice. It can give immediate benefits in
revising extension recommendations either by omitting components
which are unacceptable to target group farmers, or by modifying
treatment levels to bring them within the reach of farmers. The more
relevant improvement packages build credibility for the economist
with the extension service and with research adminsitrators. It
is less painful if the results and recommendations reviewed are
from thework of past researchers and not present incumbents.
(3) Questioning the relevance of ongoing research programmes.
(a) By asking who, which target group of farmers, is the work
being done for. Making a strong issue of local specificity
(b) By asking whether target group farmers will be able to
absorb the level of changes implied in the management variables
Both types of questioning are immediately critical of current research
efforts and therefore of the station.management and technical scientists.
The questions are justified if relevancy is a major issue and sympathetic
individual technical scientists cannot be identified. Both types of
questioning should be specific, and reviewing experimental programmes
and investigating target group circumstances will be a pre-requisite
for the FSE before embarking on detailed discussions along these lines
with technical scientists.'"
(2) PROBLEMS EMERGING FROM EXPERIENCE IN EASTERN AFRICA.
Experience in institutionalising the procedures over the last two
years have picked large holes in these strategies. At the same time
key issues in introducing FSR procedures have crystalised. Care was
taken through out the demonstrations of procedures to expose national
research administrators to the concepts and methods. However the
authority of Research Station Directors, both to influence research
planning within disciplines and to control operations was badly
underestimated. It had been foreseen that the budgetary needs for on farm
trials would be an obstacle to this component of procedures being
introduced, but in fact budgetary factors also proved a barrier to
Farm Economists moving off their stations and working among farmers,
and against taking biologists with them. Station Directors were loathe
to release transport and travel funds in what appeared to be dispropor-
tionate amounts, to new, junior research officers. Despite briefing
sessions on all stations at which economists were introduced, Directors
were often not clear or convinced of the role of the trainees. At the
same time they identified 'money related' jobs on the stations which
they felt to be the proper responsibility of an economist. These range
from costing the station dairy-herd, or station vegetable production,
to assessing future market prices for a new crop currently under
observation. An economist can do these jobs, the danger-is:their
detracting from his role of improving relevance in planning and
interpreting adaptive research.
Working links with biologists proved to be very difficult to
establish. Several facets inhibited cooperative research efforts.
(1) The biologists prime interest is his professional status within
his discipline. This is reflected in the execution of closely
controlled experiments demonstrating his disciplinary competence.
Dilution of this specialist orientation, as is implied by cooperation
with economists, threatens his professional peer group status.
There are similar penalties in working close to the farmer
where science is less pure.
(2) Technical researchers are locked into sets of specialist
experiments which are programmed and budgeted over a period
of years. Additional commitments are difficult to reconcile
with these programmes and supplementary budgeting is awkward.
(3) When any particular set of ongoing experiments is concluded
the established mechanism for deciding new commitments comes
into operation. Attempts to override this by the economists
trainee creates conflicts with established channels and his
interventions, given his junior status, are usually squashed.
(4) This junior status of the trainee economist exacerbates the
impression of 'unjustifiable intrusion' in two ways:
Cooperative efforts with 'senior' biologists are awkward, the
trainee is very much the underdog. The trainee has a poor
command 'of his new profession and cannot put his case in a
convincing way. The would be marriage has proved an extremely
difficult one to consumate.
Overall, the introduction of FSR procedures into research planning and
interpretation tends to shift the economist and agronomist, perhaps
perceive traditionally as playing service roles to disciplinary
researchers, to a central, pivotal role. The establishment feels threatened
and the social scientist seen as the intruder, is rejected.
(3) EVOLUTION IN IMPLEMENTATION STRATEGY
Lessons have been learned from this experience. Of first importance
is the recognition of the need for flexibility and a pragmatic approach
to particular institutional situations and to the personalities
dominating such situations. A major point of strategy is to focus
on research. establishments where there is already a strong awareness
that relevance is a problem.
Within such establishments where authority is strong, institutional
change may be a valuable vehicle for bringing in FSR procedures.
Where direction is weak or conservative, or where organisation is poor,
procedural innovations are seen as an added source of confusion a nuisance.
In such circumstances only a bottom up approach, working through the
station and individual scientist level, seems feasible.
Ideally the two; top down with authority, and bottom up through
individual researchers are complementary.
A clear distinction has emerged between applied and adaptive research.
Applied research is the solution of technical problems on research
stations organised on disciplinary and commodity lines. Adaptive
research is a selection and testing from the range of potentially
relevant technical solutions of that part or whole solution which is
within the resource and management capabilities of target group farmers
for whom the particular problem addressed by the solution is a priority,
A good example of the applied,adaptive distinction, and the dilemma
of having no basis for the choice of relevant solutions, is semi arid
areas with reliable starch supply as the major problem for farmers
who nevertheless late plant their major start staple, maize. They
'should' plaint dry to give the crop full use of the limited moisture.
However their oxen are in such poor condition at the end of the dry
season that they cannot break the land until their condition improves
with new grass brought by the same rain which softens the land and
reduces the power required to break it 'Plant on time before the
rains' does rot address their problem it address the management compromise
of late planting, which they are forced into by their circumstances,
the poor condition of their animals.
Applied research is a range of fields has generated a variety of
potential solutions to their real problem.
(1) Animal Production
(a) Feed through the dry season
(i) by planting grass types that will carry over the dry season
(ii) by planting productive grass types in local low lying
areas where residual moisture persists.
(iii) Specifying purchased feeds to be used.
(b) Prov:-sion of AI service to reduce the need to carry male
calve-s through for breeding purposes and thus reduce the
(c) Cows used for draught to reduce the conflict between priorities
for cxen for draft and for cows for milk, to further reduce
the stocking rate.
a) Yokes giving improved power transfer
b) Equipment for breaking the land requiring less power.
3) Plant Breeding
a) A shorter term maize variety which can be planted late and still
escape the periodic cut off at the end of the rains which brings
b) Alternative species known to be less susceptible to drought.
a) Minimum tillage techniques which, with or without the use of
herbicide, will allow early establishment.
b) Land preparation at the back end of the previous rains, while
the land is soft and animals in good condition will allow timely
In contrast to the bald, uncompromising and unhelpful advice to plant
early, all these techniques are relevant and bear on the problem.
However, these are only possible, or potential solutions. Only an
understanding of the target group farmers' present system, through
adaptive research, allows the selection and subsequent local testing,
of apparently feasible solutions. Feasible solutions have resource
demands within the reach of target group farmers and resource
allocation requirements which are compatible with the resource
allocation requirements of other competing farmer priorities. The
example demonstrates the strength of the case for an institutional
adaptive research function and a two level hierarchy in research
organisation. Within a two level hierarchy every research worker
should know his role is either applied or adaptive. Current vagueness
allows sloppy generalisations about working for farmers when the real
aims are either papers for journals or a quiet life!
Operationally a two level hierarchy offers the opportunity to build
the adaptive research effort on young professionals trained from the
start in this approach. It gets pver the senior biologist/trainee
economist conflict and avoids the need for the new discipline of economics
to break in to the on going disciplinary programmes. Among the young
professionals all are learning their roles simultaneously, this avoids
the imbalance created where a new economist has not built up the
experience to demonstrate his usefulness and argue his case to the
A revised strategy then is to establish adaptive research teams,
who build up their experience together, drawing on the existing
body of knowledge and older disciplinary oriented specialists for
potential solutions to identified system problems. Once established,
adaptive teams begin to trickle unsolved technical problems, important
for farmer development, back to applied specialists. The trickle
builds up to a flow until technical problems, identified on farms,
preoccupy both Adaptive and Applied levels in the research hierarchy.
At this point the appropriate manpower balance between the two levels
will have identified itself.
(4) OUTSTANDING ISSUES IN MOBILIZING FSR FOR AGRICULTURAL DEVELOPMENT
There are major outstanding issues in mobilising the utilisation
of the FSR procedures which could well fill further papers in this
series. Most important in encouraging the use of FSR procedures
in agricultural research, is gearing the incentive system for
scientists more closely to the benefits for farmers arising from
Training for would be researchers, and perhaps more generally for
would be third world agriculturalists, is a second urgent issue.
All professionals working in smallholder agriculture need a systems
perspective. University agricultural training at Bachelor level should
include courses on understanding how small farmers operate their systems
as a foundation for relevance in research, programme planning and
Finally it should be emphasised that while CIMMYT's immediate interest
is in the application of FSR procedures to improve the relevancy of
research recommendations, the understanding of the farming system
necessarily includes an understanding of the relationships between the
farm and its production environment. Much of this environment is made up
of infrastructure, delivery systems and policy variables, all areas of
government intervention. The potential of FSR as a link between local
and national priorities and between the farm and its infrastructure,
giving insights into both policy and programme requirements, needs urgent
1. Gerhart. J. 1975. The Diffusion of hybrid corn in Western Kenya
Abridged by CIMMYT.
2. Kiray. M & J. Hinderink. 'Interdependencies between Agroeconomic
Development and Social Change. Journal of Development Studies IV.4 (1968).
3. A Series of CIMMYT sponsored and abridged Adoption studies, including
Gerhart referenced above: 1975 & 1976.
a. Winkelmann. D. The adoption of new maize technology in Plan Puebla
b. Demir. N. The adoption of new bread wheat technology in selected
regions of Turkey.
c. Vyas. V.S. India's high yielding varieties programme in wheat
1966-67 to 1971-72.
d. Gafsi S. Green Revolution: The Tunisian experience.
e. Colmenares J.H. Adoption of hybrid seeds and fertilizer among
Columbian corn growers.
f. Cutie. T.J. Diffusion of hybrid corn technology; the case of El
4. CIMMYT Economics Program. 1979. Planning Technologies Appropriate to
farmers; Concepts and Procedures. (Draft for circulation and comment.)
5. CIMMYT Eastern African Economics Programme 1979. Report No. 4. Deriving
Recommendation Domains for Central Province, Zambia.