Journal of farming systems research-extension

Material Information

Journal of farming systems research-extension
Running title:
Journal for farming systems research-extension
Abbreviated Title:
J. farming syst. res.-ext.
Association of Farming Systems Research-Extension
Place of Publication:
Tucson Ariz. USA
Association of Farming Systems Research-Extension
Publication Date:
Physical Description:
v. : ill. ; 23 cm.


Subjects / Keywords:
Agricultural systems -- Periodicals -- Developing countries ( lcsh )
Agricultural extension work -- Research -- Periodicals ( lcsh )
Sustainable agriculture -- Periodicals -- Developing countries ( lcsh )
serial ( sobekcm )
periodical ( marcgt )


Dates or Sequential Designation:
Vol. 1, no. 1-
General Note:
Title varies slightly.
General Note:
Title from cover.
General Note:
Latest issue consulted: Vol. 1, no. 2, published in 1990.
Electronic resources created as part of a prototype UF Institutional Repository and Faculty Papers project by the University of Florida.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact Digital Services ( with any additional information they can provide.
Resource Identifier:
22044949 ( OCLC )
sn 90001812 ( LCCN )
1051-6786 ( ISSN )

Full Text
Volume 4, Number 1

for Farming Systems
Research- Extension

for Farming Systems
Research- Extension

Volume 4, Number 1, 1993

Published by
the Association for Farming Systems Research-Extension

Journal for Farming Systems Research-Extension

Timothy R. Frankenberger
Office of Arid Lands Studies
The University of Arizona, Tucson

Associate Editors
Philip E. Coyle and Jennifer Manthei
Office of Arid Lands Studies
The University of Arizona, Tucson

Production and Layout
Philip E. Coyle, Jennifer Manthei, and Sonia Telesco
Arid Lands Design, Office of Arid Lands Studies
The University of Arizona, Tucson

Ford Foundation
The University of Arizona

The Journal for Farming Systems Research-Extension is published by the Association for
Farming Systems Research-Extension (AFSRE), an international society organized to
promote the development and dissemination of methods and results of participatory on-
farm systems research and extension. The objectives of such research are the development
and adoption through participation by farm household members of improved and
appropriate technologies and management strategies to meet the socioeconomic and
nutritional needs of farm families; to foster the efficient and sustainable use of natural
resources; and to contribute toward meeting global requirements for food, feed, and
The purpose of the Journal is to present multidisciplinary reports of on-farm research-
extension work completed in the field, and discussions on methodology and other issues
of interest to farming systems practitioners, administrators, and trainers. The Journal
serves as a proceedings for the annual international Farming Systems Symposium from
which selected and refereed papers are included. It also welcomes contributed articles
from members of the AFSRE who were unable to attend the symposium. Contributed
articles will be judged by the same review process as invited articles.

ISSN: 1051-6786

The papers collected together in this issue of the Journal for Farming
Systems Research-Extension were drawn from the invited presentations of the
1992 Farming Systems Research-Extension Symposium held at Michigan
State University, East Lansing, Michigan on September 13-19, 1992. They
represent recent thinking about policy-making, farmer participation in planned
agricultural change, household food security, and the future of Farming
Systems Research in general, and the Association for Farming Systems
Research-Extension in particular. We hope that you find these papers of
The Editors

Journal for Farming Systems Research-Extension
Volume 4, Number 1, 1993


1 Challenges of Farming Systems Research and Extension
J. A. Berdegui

11 A Participatory Experiment in Sustainable Agriculture
Clive Lightfoot and Reg Noble

35 Integrating Household Food Security into Farming
Systems Research-Extension
T.R.- Frankenberger and P.E. Coyle

67 Inability of Farming Systems Research to Deal with
Agricultural Policy
Doyle Baker

87 Methods for Analysis by Farmers: The Professional Challenge
Robert Chambers

103 From Research to Innovation: Getting the Most from
Interaction with NGOs in FSRE
John Farrington and A.J. Bebbington

125 Structural Adjustment Policy Reform and the
Resource-Scarce Farm Family: The Waning Role of FSRE
T.J. Finan

139 Making the Farmer's Voice Count: Issues and
Opportunities for Promoting Farmer-Responsive Research
Deborah Merrill-Sands and Marie-Helene Collion

Challenges of Farming Systems Research
and Extension'

J. A. Berdegue2

This paper focuses on the question of present and future challenges of
Farming Systems Research and Extension (FSRE) and the role that can
be played byan organization such as the Association for Farming Systems
Research-Extension (AFSRE) in helping meet them. The basic premise
of this paper is that, in a rapidly changing world economy, the develop-
ment potential for small-scale agriculture in Third World nations is
becoming narrower each day; as a consequence, FSRE must become
more sharply focused in terms of the target populations with which we
work and, at the same time, much more integrated with other agricultural
development instruments.

The role of Farming SystemsResearch-Extension (FSRE), of course, has
always been to promote agricultural change. We have known since the
beginning--even if we have not always acted consequently-that the evolu-
tion of agriculture follows the influences and dynamics of the national and
international economies and the corresponding social and political systems. At
least in Latin America, agricultural research and extension have in the past
been asked to support the national goal of promoting food self-sufficiency, a
prerequisite for urban and industrial development. The role of agriculture was
to substitute food imports within the framework of protected and heavily
subsidized economies. This framework is rapidly fading out; what were
originally structural adjustment measures in face of the major economic crisis
of the 1980s have become permanent tenets, at least for the foreseeable future.
Competitiveness, productivity, and profitability are fundamental qualities
in countries that are rapidly embracing the paradigms of the free market and
the international integration of economies. These qualities run parallel with
the downsizing of the role of the public sector and institutions that until now

1 Keynote presentation at the Twelfth Annual Association for Farming Systems Research-
Extension Symposium, Michigan State University, East Lansing, September 13-18, 1992.
SRIMISP, Casilla 244 34, Santiago, Chile.


have had an indisputable and almost exclusive role in promoting agricultural
development and the modernization of the small-scale, resource-poor farm-
ing sector. Agricultural researchers, extensionists, and development agents are
now being asked to justify their work in terms of contributing to this process
of agricultural modernization. Even when the public sector is willing to
allocate resources to fighting rural poverty, it is made clear that this constitutes
a separate effort from those aimed at agricultural development, involving
different tools and approaches focused on different populations.
The performance criteria that today call the attention of policy-makers in
our own countries and abroad are much harder to meet than those of the first
20 years ofFSRE. Modest increments in productivity are no longer sufficient
to justify the investment of scarce resources. We are being asked to reconvert
agriculture, introduce for-export crops, generate foreign exchange, produce
high-quality goods that can penetrate and preserve new international markets,
and achieve levels of productivity and economic efficiency that permit the
survival of the farming systems in the absence ofsubsidies, tariffs, or protection
from imports.
FSRE came into existence to help small-scale farmers in the more marginal
areas participate in the processes that had been implemented during the Green
Revolution with resource-rich producers in the best agricultural regions. After
two decades the goals have become more elusive, more complex, and more
difficult to achieve. Even increasing the productivity of commercial agricul-
ture is not enough; these systems are also facing serious problems. It is in this
context that the development potential of small-scale farming has narrowed.
FSRE will have a role to play only if it can prove itself an effective tool in
promoting not the gradual and slow improvement of small-scale farming
systems, but the rapid modernization of at least some sectors of peasant
Are there reasons to believe that FSRE can play this role? If so, what are the
steps that need to be taken in order to meet this challenge? How can an
organization such as the Association for Farming Systems Research-Extension
(AFSRE) aid in the successful implementation of those measures?

The basic principle on which FSRE is built is that the process of technology
generation and adaptation should be responsive to the characteristics of the
targeted farming systems, the farmers' objectives, and the conditions under
which specific populations must practice agriculture. This principle and its
derivations are perhaps the most significant contributions of the FSRE
approach. From there, it follows that:
A systems approach is necessary because any given agricultural enterprise
is the final product of a complex and dynamic interplay of numerous

Journal for Farming Systems Research-Extension


components and processes that take place both within and outside the
The farmers' circumstances are location specific, and so should be the
research and extension process;
A diagnostic phase is required in the overall process of technology
generation and transfer in order to understand the set of conditions that
will eventually determine the adoption rates and the final impact of the
whole effort;
On-farm adaptation and testing of technology would enhance the
probability of success of any given innovation; and
Farmer participation is indispensable if the improved technologies are to
be responsive to the farmers' priorities and objectives.
FSRE is such a logical and coherent proposition that one wonders why its
success has been limited. Tripp (1991) explores two basic factors: the
institutional context and the quality ofon-farm research. In this paper I discuss
the first of these factors.
The basic principles ofFSRE mentioned above have led to manyinteresting
and positive methodological approaches. However, it has also induced most
FSRE projects to work within the framework imposed by the constraints faced
by the farmers without questioning whether such a framework allows any
development potential at all. FSRE has been an approach used mainly in
working with resource-poor farmers living in marginal areas. As a result, most
projects have faced one or more of the following conditions: unfavorable soils,
climate, and topography; lack of credit; underdeveloped or nonexistent
markets for both inputs and agricultural products; weak or incipient farmers'
organizations; economic policies that discriminate against agriculture as a
whole and against poor farmers in particular; lack of interest on the part of
national agricultural institutions that feel their scarce resources would obtain
a better return if invested in more favorable areas; and the attitude of many
farmers that, based on the experience of many generations, innovation is a
highly risky business. FSRE has frequently attempted to untie this Gordian
knot employing only two tools: the generation or adaptation of improved
technologies and the transfer of innovations.
Even in those infrequent cases in which research and extension work hand-
in-hand to solve the puzzle, it is very likely that the problem will remain.
Generating an improved technology and informing the farmers about it only
creates a potential for change. To materialize this potential, the farmers usually
require credit to purchase inputs, markets that generate a stable demand for
the additional produce, transportation and roads to take the inputs to the farm
and products to the marketplace, timely information to adjust to changing
environments, and, in general, a whole setup of support services and infra-
structures on which the success of agriculture depends. A recent study (Low
et al., 1991) reviewed the end result of 53 research initiatives in Southern

Vol. 4, No. 1, 1993


Africa and concluded that 39 failed at least partly due to institutional and
resource-availability factors.
This whole question of impact potential (from the point of view of the
researcher or extensionist) or of development potential (from the farmers'
perspective) will become increasingly important as each country struggles to
keep its head above water within the rearrangement of the world economy that
is currently taking place. It is no longer sufficient to talk about adoption rate.
Instead, we will soon be asked: Can FSRE lead to the rapid modernization of
small-scale, resource-poor agriculture?
The answer to this question has two parts. First, we should be more willing
to say that many farmers in the more marginal areas of the world cannot make
this transition in the short or even medium term. Second, we should
acknowledge that FSRE is not a powerful enough tool in many instances and
that the Green Revolution failed in those cases not because of its conceptual
or methodological approach, but because the necessary conditions for agricul-
tural development are simply not in place. Of course, there is still a role for
agricultural research and extension in such cases. Many societies are willing to
invest in the alleviation of extreme rural poverty, and even marginal improve-
ments in agricultural production and productivity can make a contribution
toward that very worthwhile goal. However, it should be clear that "develop-
ment" is not at hand in those cases, regardless of the conceptual or method-
ological approach.
FSRE has an advantage under these conditions if only because its practi-
tioners have shown that they are willing to work in these very marginal areas.
We have accumulated vast experience over the past 20 years: we understand
the farmers' logic better than others; we know how to elicit their participation;
we have developed tools and methods; and we have built formal and informal
networks of agricultural and social scientists through which information and
experiences flow more or less effectively to other areas of the world. These are
very valuable resources for a most difficult task.
On the other hand, there are many situations, involving hundreds of
thousands of farming families in dozens of countries, in which there is a clear
potential for impact. In those instances FSRE can help in bringing about a
more rapid modernization of resource-poor agriculture. However, even here
the "institutional context" needs to be reconsidered. In short, we cannot
consider institutional factors as external parameters to our projects; we need
to internalize them. This requires thinking more in terms of agricultural
development efforts that include, but go beyond, technology research and
extension. There is no reason why the basic methodology of FSRE could not
be complemented with such tools as revolving credit funds, marketing
support, in-depth training of local resource persons, development of micro-
irrigation projects, and strengthening of the local organizations of farmers and
of their leadership.

Journal for Farming Systems Research-Extension


There are at least two consequences of this idea. First, individual projects
would be more complex and expensive; thus fewer farmers could be reached
for the same amount of money and other resources. This is a question of
breadth versus depth, and it is one that is likely to trouble policy-makers who
want to see as many families benefit as possible. Second, because no single
organization is likely to have the required competence and technical expertise
in all of the components of a given project, success would necessarily rest on
the ability to develop strong and efficient interinstitutional arrangements. If
one considers the track record of FSRE in promoting effective linkages
between researchers, extensionists, and farmers the difficulty of creating these
arrangements becomes obvious. However, I refer to the Chilean experience
in order to illustrate the feasibility of such interinstitutional linkages (Berde-
gu6, 1990).
In the early 1980s, Chile's extension service was privatized. Today, small
firms made up of one or two agronomists and a few agricultural technicians
compete with each other to be assigned the funds to provide technology
transfer services to a group of peasant farmers. Several dozen NGOs, and even
local or regional farmers' organizations such as cooperatives, operate as
"private technology transfer consultant firms." The role of the national
agricultural development agency (INDAP) is to define, supervise, and evalu-
ate the work of these private firms-developing the guiding concepts,
methodologies, and procedures, and, of course, providing the system's
financial resources.
In 1992 an effort to strengthen this system involved the development of
new technology transfer modalities. One of these is the "co-financed technol-
ogy transfer program," which operates around microregional and integrated
agricultural development projects, each of which is planned for a period of
three to five years, and works with 100-300 peasant families. In this system,
INDAP provides the funds (up to 80 percent of the total cost of the project)
for the technology transfer component while the cooperating private agency
funds other agricultural development tools such as farmers' training pro-
grams, revolving credit funds, marketing infrastructure and services, and so
on. Making use ofa World Bank loan, INDAP has made available close to US$
1 million to fund cofinanced projects that will be started during the 1992-93
agricultural year. The local and regional NGOs and farmers' organizations
must come up with an additional US$ 200,000. Although the process is still
not finished, it is expected that at least 50 projects to benefit at least 1,000
peasant families will be presented to the first national competition.
This new program is based on the assumption that more complex programs
will be successful only if effective decentralization takes place, if each effort is
kept to a manageable size, and if the key actors have a clear stake in the success
of their project. The main role of the national agency is to define clear rules
of the game, and then let the local actors operate as freely as possible to the
best of their ability. In this way it is expected that each project will contribute

Vol. 4, No. 1, 1993


to putting in place the necessary "institutional factors" that are required for
technology adoption and agricultural development to take place.


As Tripp (1991:247) makes clear, institutional issues explain only in part
"why, in spite of so much investment and interest...[have] the tangible results
been so modest." The quality of on-farm research,3 Tripp argues, is respon-
sible for many of the shortcomings of FSRE projects. Many authors have
analyzed this issue in general or with respect to specific stages of the FSRE
methodology. Questions have been raised about how target areas and
populations are selected and defined; how diagnostic studies can often lead to
misleading and/or incomplete conclusions; how the design or planning stage
is very often confused with the diagnostic studies; and how little time is spent
on propositive analysis of the field data. Questions have also been raised
concerning the insufficient consideration of macroeconomic and policy
determinants; the tendency to tackle an excessive number of problems and
objectives simultaneously leading to poorly focused projects, particularly
during the on-farm research stage; and the operational weakness of the
feedback loops that are supposed to be a key component of the FSRE
There is no need to document once more these well known problems.
Instead, it is important to highlight one general argument: the methodology
of FSRE has shown a tendency to become increasingly ill-defined. For many
of us, an acceptable methodological protocol for conducting FSRE is no
longer clear. This uncertainty is a product of a positive development; more
people, with different backgrounds and interests, base their work on the
overall idea of FSRE now than in the 1970s or early 1980s. As a result, the
FSRE "movement" is today less monolithic than at its beginning. At the same
time, as more experience is accumulated, it is natural and legitimate that
different people raise new theoretical, methodological, and operational issues.
This ability and willingness to criticize and question old notions is the basic
engine that moves science and knowledge.
However, all disciplines need to be able to sort the useful from the
nonuseful, the sound from the unsound, the true improvements from the
background noise. It is not clear to me how this is done in the FSRE
movement, with its unrivaled disposition to welcome all aboard. No new idea
should be censored to begin with, but all proposals should be tested rigorously
and, above all, new developments need to be integrated effectively into the
nucleus of current thought. Otherwise, the new developments become simple
appendices that are not internalized within the conceptual and methodolog-
3 This author distinguishes between FSR as a perspective on research and on-farm research as the
type of work done by FSR.

Journal for Farming Systems Research-Extension


ical framework of FSRE. This situation not only affects the questions of
concepts and methods used to address old objectives of income, risk,
productivity, and so on. It is also present in the new or emerging issues of
gender and, more recently, ofsustainability. These concepts reflect societal
objectives that are here to stay and that need to be integrated into agricultural
research and extension, just as they must be dealt with in all areas of
contemporary life.
However, it is indispensable that these new issues be integrated effectively
and not only in words. It is also necessary that they permeate FSRE in an
efficient and synergistic manner. This has not been easy with gender issues
and, if anything, it will be even more difficult with respect to the question of
sustainability. Like farming systems themselves, FSRE needs to become not
only more productive and profitable, but also more equitable both within and
between generations. The question of sustainability, for example, inevitably
raises the dimension of time on a large scale. I wonder if most FSRE teams are
adequately equipped, conceptually, methodologically, and financially to deal
with this complexity.

Because of the problems analyzed above, as well as other considerations that
will be reviewed below, the ASFRE has reached a turning point. If we agree
that, in order to continue its development as a useful approach for the
promotion of agricultural change, FSRE must be able to tackle the old and
new problems with fresh concepts and more powerful methods, the most
important question of the Association is how it can promote this renewal. The
following four ideas are useful in charting a new course.
The AFSRE must come out strongly and convincingly as forum for change
and renewal. IfFSRE is to survive as a dynamic and creative force, it must adapt
to the new conditions that characterize our economies, and to the resulting
new demands and objectives to be met by our countries' agriculture. The
AFSRE's primary goal must be to promote and support this change and
renewal. This can be done through different activities:
First, the membership must explicitly consider this issue and define a
position on it. The efforts of the Board, symposia, journal, and newsletter,
must reflect the position of the members of the Association. Until now, it is
likely that an important proportion of the membership has seen the Associa-
tion mainly as a provider of services organizing symposia, publishing a journal,
or even providing the occasion to meet and stay in touch with friends. I
wonder how many of us think of the Association as an active leadership forum
in the development of new ideas. The members have to establish this forum
if they expect the AFSRE to play such a role.

Vol. 4, No. 1, 1993


Second, through the definition of the agenda of the symposia, the
Association can influence the topics that will be debated and the ideas that will
emerge in the future. Considering that the regional networks and associations
are now conducting their own meetings, it is necessary to define a specific
profile for the international symposia that is complementary and not compet-
itive or reiterative with the regional events. In my opinion, the international
symposia should be devoted to more global issues.
Third, the Journal of the AFSRE can also be seen as a vehicle for the
promotion of in-depth debate about FSRE and its future contributions. The
job done until now has been extraordinary, particularly if one considers the
almost total lack of resources that has constrained this effort from the
beginning. However, it is always possible to improve and, in particular, it
would be ideal ifa strong and highly qualified Editorial Board could be formed
to support the work of the Editor. If the AFSRE hopes to make a credible effort
to establish itself as a leadership forum for change, it also needs to confront
its own internal renewal.
The AFSRE must strive to integrate different currents of thought that apply
systems concepts to the problems of agricultural development. Farming Systems
Research and Extension is not really a movement on its own, but only a specific
version of a wider current of thought within the field of agricultural develop-
ment. Moreover, within FSRE one could well speak of a subsector that grew
out of U.S.-supported, U.S. university-based projects and subsequently
developed the symposia and other initiatives that eventually gave form to the
AFSRE. The topics debated, articles published, and even the people that
attend our meetings follow much too closely the evolution of the efforts,
possibilities, and needs of the original nucleus that formed the AFSRE.
One option is to formalize this more particular nature of the AFSRE,
making clear its U.S. affiliation and dependency. From there, it would even be
possible to establish collaborative relationships with other groups, and it
would also make it feasible to reduce the financial burden of publishing a
journal for world-wide distribution or organizing meetings for people from all
over the world. Not that this U.S.-based association would need to be closed
to citizens, but those of us from other countries who want to participate would
need to be able to meet the costs of doing so.
The francophone and other European schools have never been adequately
represented in the AFSRE or its predecessors. Today, out of the FSRE sector,
African, Asian, and Latin American organizations are emerging with distinc-
tive characteristics and with the legitimate need to manage their own affairs
with greater independence. In Latin America, for example, there are at least
five major farming systems networks, and the systems concept has become
integrated in a number of national and regional institutions that are simply not
part of the AFSRE. In Asia and Africa there are also regional networks that
publish their own journals, promote their own research priorities, and hold
their own regular meetings. The AFSRE should play a role in promoting

Journal for Farming Systems Research-Extension


initiatives to link these regional players formally, both within and without the
FSRE matrix. This will not bear fruit if our attitude is one of inviting others
to join us; rather, we must say that we are willing to think together with others
about new forms of association. The meeting of the 1994 symposium in
Europe would be an excellent opportunity to formalize these new links.
The AFSRE must become a truly international body. To begin with, the
AFSRE should consider becoming an international body that can integrate the
regional groups that recognize a common origin in the FSRE movement. This
means that the authorities of the Association, journals, and newsletters, and
organizing committees of its symposia should include people from Africa,
Asia, Europe, Latin America, and North America in credible numbers and in
decision-making capacities. It also means that the international symposium
should sometimes be held in areas other than the U.S. and that its frequency
should be restudied to accommodate regional events. An international
AFSRE would need to be organized on the basis of a federation of regional
initiatives because it is not likely that the former would agree to disband in
order to become part of an international AFSRE.
The AFSRE must be able to obtain adequate resources. Farming Systems
Research and Extension tells us that in designing an improved system one
must consider the constraints imposed by the socioeconomic environment.
This principle applies perfectly to the issue that we are now debating. To play
a stronger role in the promotion of new ideas, to join with other people in this
effort, to become an international forum, and to implement a specific plan
organized around the goals and objectives defined by the entire membership,
the AFSRE must obtain a minimum level of funding. Without this funding it
is unlikely that the AFSRE can support an even modestly ambitious plan for
change and renewal; but, without this plan, the whole question of the
continued viability of the Association will need to be debated in a very serious
and concrete manner.

Berdcgue, J. A. 1990. NGOs and farmers' organizations in research and extension in
Chile. Network Paper 19. Overseas Development Institute, Agricultural Administra-
tion (Research and Extension) Network, London.
Low, A.R, S.R. Waddington, and E.M. Shumba. 1991. On-farm research in southern
Africa: The prospects for achieving greater impact. Pages 257-272 in R. Tripp, ed.,
Planned change in farming systems: Progress in on-farm research. West Sussex,
England: John Wiley and Sons.
Tripp, R 1991. The limitations of on-farm research. Pages 247-256 in R. Tripp, ed.,
Planned change in farming systems: Progress in on-farm research. West Sussex,
England: John Wiley and Sons.

Vol. 4, No. 1, 1993

A Participatory Experiment in Sustainable

Clive Lightfoot and Reg Noble2

On-Farm Experimentation (OFE) requires rethinking if it is to
contribute to our understanding ofsustainability. More holistic and
farmer-participatory approaches are essential to cope with the prob-
lems presented by the diverse and complex systems common to small-
scale farming that the conventional OFE approach has not been able
to address adequately. A new protocol for OFE must involve a
resource-systems approach with the aim of improving the sustainabil-
ity of the whole farming system through increased integration and
recycling of on-farm resources,
Farmer-participatory mapping and modeling form the major manage-
ment tools for experiment and evaluation, providing a qualitative and
quantitative measure of farming system status and transformation.
Preliminary evidence presented in this paper indicates that for integrat-
ed aquaculture-agriculture systems, these tools do enable farmers to
develop a more holistic approach to management of their farms'
resources. Data presented demonstrates that household cash income
and food supply can be significantly increased if farmers' resource
management skills are improved using mapping and modeling tech-
niques combined with exposure to different systems that illustrate
possibilities for integration and recycling of farm resources.

Over the past 10-15 years, On-Farm Experimentation (OFE) in Farming
Systems Research (FSR) has concentrated on improving commodity yields
and other aspects of agricultural performance. Although many experiments
were conducted with a systems perspective, few tested hypotheses at the level

SKeynote presentation at the Twelfth Annual Association for Farming Systems Research-Extension
Symposium, Michigan State University, East Lansing, September 13-18, 1992.
International Center for Living Aquatic Resources Management (ICLARM), MC P.O. Box
2631, Makati 0718, Metro Manila, Philippines.


of the whole farm system. Many of these experiments were conducted with
farmers but few were designed and tested by them. Experiments of this nature
are important and necessary for disciplinary and commodity pursuits; howev-
er, they have done little to further the concepts of farmer participation and
holistic treatment of farming problems. Indeed, this imbalance of single-
enterprise, researcher-managed, on-farm experiments versus whole farm,
farmer-managed experiments has contributed to current disenchantment
with FSR. Moreover, disenchantment has spread into our own ranks. Partic-
ipatory Rural Appraisal, Participatory Technology Development, Low Exter-
nal Input Agriculture, and Agroecosystems Analysis all distance themselves
from FSR.

Research Areas and Approaches
Enter sustainable agriculture, an object of study that interests agricultural
researchers of all stripes. Economists, agronomists, and microbiologists are
busy setting up new, exciting work (Batie, 1989; Edwards et al., 1990;
Harrington, 1991).
Will FSR practitioners again mimic the research of disciplinary and com-
modity scientists or will they fashion programs that are uniquely participatory
and holistic? Our formidable task involves lengthy farm system description,
constraints analysis, design of interventions to overcome constraints, and on-
farm experiments to evaluate interventions. Nevertheless, research agendas
invariably break down into researcher-managed component technology re-
search on varieties, fertilizers, green manures, etc. Indeed, we must avoid the
fate of US sustainable agriculture research, in which "[w]hole-farm studies
and studies combining crops and livestock were comparatively uncommon, so
it usually would not be possible to analyze how production efficiencies might
be increased through integrating all components of a farm or by combining
crops and livestock" (Anderson and Lockeretz, 1992). Perhaps the new
articulation of sustainable agriculture will inspire more farmer participation in
our on-farm experiments and more holistic scope in our tests. Where in all the
rhetoric of sustainability can we find help?

Definitions and Themes
Conventional goals of sustainable agriculture invariably seek to maintain or
increase biological and economic productivity, enhance efficiency of inputs
used, increase stability of production, increase resilience to environmental
changes, minimize adverse environmental impacts, and ensure social compat-
ibility. Although these goals offer broad scope for many disciplines, they
provide scant operational guidance. Operational hints can be found in
Conway's definition ofsustainability, which is couched in terms of the farming
system's ability to "maintain its productivity when subject to stress or
perturbation" (Conway, 1986). Conway and others, most notably Altieri

Journalfor Farming Systems Research-Extension


(1987), have gone on to suggest that sustainability is enhanced through
system diversity of enterprises over space and time as well as the recycling of
energy and nutrients, which reduces the need for external inputs. Fortunately,
many farmer-participatory techniques exist for FSR practitioners to draw on
in developing new farming systems that promote these characteristics (Cham-
bers et al., 1989; Haverkort et al., 1991).
Our challenge is not only to fashion a type of OFE that will generate
farming systems that enhance diversity and recycling, but also to develop ways
to measure these characteristics. Even before this can be done, the "how do
I improve enterprise performance" perspective of researchers and farmers will
have to be changed. Sustainable farming systems are more likely to emerge
from a "how do I regenerate my farm resource base" perspective. This paper
attempts to contribute to this challenge in three ways. First, we present some
ideas on protocols for on-farm experiments in sustainable resource manage-
ment. Second, we look at strategies to assess the impact of these systems.
Third, we look at methods of impact evaluation in terms of our definition of
sustainability. We end the paper with a discussion of anticipated problems for
the years ahead.


Indigenous Categorization of Natural Resource Systems
Small-scale, low-resource farming systems are complex, diverse environ-
ments where agricultural production is risk-prone and very much dependent
on the whim of the climate, particularly the availability of water. Such farming
systems dominate the rainfed tropics, especially sub-Sahelian Africa. Howev-
er, risk can be reduced and production increased by enhancing, exploiting,
and combining the diverse elements that comprise such farming systems as
well as by introducing elements that can aid this process (multipurpose trees,
crop-fish systems, etc.).
Local knowledge is an essential prerequisite to understanding these com-
plex systems and developing management practices that will stabilize and
sustain agricultural production in such environments. Therefore, farmers are
the most appropriate people to develop new management practices and
experiment in manipulating their farming systems. Unfortunately, small-scale
farmers are often under heavy pressure from research and extension personnel
to accept management practices that are commodity-oriented and have as
their goal enterprise profitability. Such practices often result in simplification
of the environment, which invariably leads to reliance on expensive external
and often nonrenewable inputs in order to maintain production. Under the
fragile socioeconomic and environmental conditions of most small-scale
farmers in sub-Sahelian Africa, external input farming is not sustainable and
will increase rather than reduce risk.

Vol. 4, No. 1, 1993


An alternative approach is to develop a protocol for enabling farmers to
experiment in designing resource management systems that are appropriate to
their local environments. Farmers have the expert knowledge of their local
resources and the skills to initiate and manage a range of innovations concurrently.
What is needed is a methodology that enhances these abilities and gives farmers
a holistic appreciation of their environment and the potential to both integrate its
elements and incorporate new ones that aid in this process.
The first step is to establish how farmers classify and perceive their natural
resources and proceed in such a way that the farmers are able to visualize their
whole farming system and the interrelationships of its elements. Local farming
communities often classify their environment based on soil type, topography,
and water resources. Such categorization recognizes the habitat diversity that
provides a spectrum of potential for agricultural production and exploitation
of natural resources. Participatory exercises involving several groups of 20 or
so farmers are the most successful approach to understanding how rural
communities classify their local natural resource systems. Most of the commu-
nity should be involved in this activity. One such method is to encourage
villagers to express the resource classification of their surrounding environ-
ment through simple maps and transects. This provides a mechanism for rapid


Figure 1. Research Study Area.
i d,

Figure 1. Research Study Area.

Journal for Farming Systems Research-Extension


rural assessment and an opportunity for farmers to re-examine their farming
in terms of natural resource systems rather than commodities.
The most appropriate setting for resource mapping is in the village. Groups
provide a better dynamic than individuals by allowing wider discussion and

Zomba Mountain Slopes I.nda

Kalonf T uika 41 qa
Maize Village /

\Murida mbf BUnused 1'
Legend: nna L l' t l M" .t

Pa'h's'- i7 I

S houses .-$ loan)
soil leeda ning S Fat gtai es
VI inaize BaSadyad T

Lewela SandyE a / dondol ,

uLaa lyeawneoa (while sandy Cda l moo ad Villge
lrea deni lnna
-M dnds \MMundt ,a \ I an on
-.. S ^- vG"- i .. F i p,'

I Mu ) n. Ii
50.y-1 I Milnd i.e rat d

ia d
FuetwodTl C I-Y fix -blH d .en .

sed MF kundl Fu a
ki ie*osaoi d J / Maz

a nan Uninia unzia Hd\ka
LI mw I blonsaka

Y -^ W e tiLoL" Monda u aa
.aka\ tye- u (wh iod iyi flain I '10W 1o Village

Wad baa
\F u nd Iee a i \
K 4 FWl I iI-

NemaoOn Rwer

Figure 2. Resource System Map Redrawn from Original Map Made by Villagers.

Vol. 4, No. 1, 1993


consensus on the indigenous categorization of their local resources. Usually,
the most successful approach is to allow villagers to first lead the researcher
through their farming area and then afterwards draw maps illustrating the
natural resource systems identified on the walk. Maps are usually drawn on the
ground using whatever materials are available. Indigenous names for resource
systems, soils, enterprises, etc. are used as much as possible. Local terms can
be very descriptive and reveal information on the nature of the resource
system, its use and importance. Also, drawing does not require literacy.
Stones, plant and animal material, and etc. can be placed on the ground to
symbolize different resources and aspects of topography and hydrology.
Drawing also enables farmers to obtain a broad overview of their environment
that they often have not visualized in its totality before (Lightfoot and
Minnick, 1991).

Zomba Plateau


Resource Mgulugulu Kalondo Minda Litaka Lyeplliwu Lyeplliwu Chilongo
System Mwe Phirl Ya Kumusa Ndi Lyesera Wala
Water Rain Rain gs SRan -
Resource Rain Ra n
Sol Laleme Clay .aerile Ba dambo Loam Sandy
Soils Sol Loam

Dry WNe Noe sugarcane, ccuarrben, tcJunbe.'ba6anas, Ne
pppers, finger miil ra, sh be.ans. green
potatoes. poppers veetales, tomato
Crops Mae, pumpn,
Maze, cassava, Peas, mase, rnee sJnfow= oJcuni,
wet Musroms s rghum sh pl toes, ".. ,, *. e potatoes, Cass a
pigeon peas. lnre lkgor ,,. i,,. uig mel .e

Planning No cops Nov.jan. Al year All year I.- '
Eucaiy ts.
Natal gree na
woodland Dawpar
Dry -- Goaos, Cat. $ ep --------
Animals .

Zomba District, Malawi. January 1990.

Journal for Farming Systems Research-Extension


Pilot work on farmer mapping has been carried out in Zomba District,
Malawi (Figure 1). Figure 2 illustrates a map as drawn by a group of 10 men
and women villagers from this district. The broad classification of natural
resource systems is into mtunda (upland/sloping land), munda (flat crop-
land), dimba (low-lying gardens with high water table), and dambo (season-
ally flooded areas that are uncultivated). Within these broad categories, more
subtle distinctions are made based on soil type and water resources, as
illustrated in the transect in Figure 3. Areas of similar soil type are used for
different crops depending on whether they are rain- or springfed. The transect
is a composite cross section of the farming area in Figure 2 showing all the
resource system categories in relation to each other. Such mapping provides
the researcher with a detailed picture of the diversity and distribution of land,
soil, and water resources from the perspective of the rural community. Farmers
will also have physically drawn, often for the first time, a complete represen-
tation of their own farm's relationship to the rural community and its
agroecological environment.


Figure 3 (continued). Land Transect of Village in Chinseau Area,
Zomba District, Malawi. January 1990.
Vol. 4, No. 1, 1993


Bioresource Flows and Resource Systems on Farms
Having established a map and transect showing the broad resource
categories for a village and its locality, researchers then encourage interested
farmers to make similar drawings for their own farms (Lightfoot and Tuan,
1990; Lightfoot et al., 1991). The majority of farmers should be interested.
These drawings should follow a logical sequence to ensure that farmers do not
simply draw enterprise maps. Natural resource systems are sketched in first,
then their respective enterprises. Following this sequence helps farmers to
change their thinking away from enterprise profitability to rehabilitation of
the farming system's natural resource bases. Resource systems are not restrict-
ed to those within the boundaries of the farm. Common grazing land and
water sources utilized by the farming household should be included. This
provides a more accurate picture of the total resource base of the smallholding;
it is often very different from a conventional idea of a farm. Finally, the farmer
household members draw arrows on the diagram indicating movement ofon-
farm biological materials (e.g., maize bran from the cropland to chicken
house, manure from the chicken house to the vegetable garden, etc.). This
provides a very clear picture of the level of integration between and within
resource systems.
Farmers will often make drawings that cover all activities, enterprises, and
bioresource flows that occur in an annual agricultural cycle. A more accurate
picture of annual operations can be obtained if drawings are made at different
times of the year to show the seasonal variations in farming activities and flows
of resources.
Figure 4 shows a typical farmer's model from Ndoka village (Zomba
District, Malawi). This drawing is by a woman who has just started to
incorporate fish ponds on her farm. The picture shows qualitative outputs
from resource systems to the household and market with some of their related
cash flows. An interesting point to note on the drawing is that only the fish
ponds have links with other resource systems and enterprises. Agricultural
residues (e.g., maize bran, waste vegetable material, etc.) are beginning to be
processed through the pond as feed and fertilizer. No other enterprises appear
to create these types ofinterlinkages and material recycling within the farming
system. Most of the arrows on the drawing are simply showing outputs from
systems and occasional inputs (e.g., fertilizer) rather than bioresource recy-
cling flows. The latter may increase with time if the farmer starts to recycle
more on-farm materials.
Such a farm model can provide a basis for further research and be used to
make suggestions about integration and improvement in recycling of biore-
sources. Farmers can also use drawings as a management tool for indicating
areas they feel need change or further study. In this way, research agendas can
be developed through mutual cooperation between farmer and researcher. So
far, 10 farmers are using the model they have drawn both as a teaching aid on
farm management for their family and as an experimental tool for planning

Journal for Farming Systems Research-Extension


new farming strategies. Frequently, drawings are displayed prominently in
farm households as a focal point for family discussion.
Successful drawing and modeling by farmers is best achieved when 20 or so
farmers work together, allowing all farmers to have an impact within the group.
Such groups provide a valuable forum for exchange of ideas and exposure to a


Figure 4. A Typical Farm Model Showing Resource Flows,
Redrawn from Farmer's Original Drawings.

Vol. 4, No. 1, 1993


diverse array of farm management options. Dynamic group work occurred largely
because we linked with existing farmer clubs and not "project" created groups. A
very useful approach is to have farmers model their current farming system and
then possible future scenarios which include culturing fish. Group interaction
often generates intense discussion and is more likely to lead to formulation of
integrated farming strategies appropriate to local agricultural systems. The role of
the researchers at the workshop is simply to facilitate and stimulate farmer
discussions, and when necessary provide biotechnical information that farmers
may not know.

Farmer Group Workshops on Integrated Crop-Fish Systems
Modeling workshops are also a useful precursor to exposing farmers to
demonstrations of integrated aquaculture-agriculture systems and their relat-
ed biotechnology. In Malawi, farmers that have been involved in modeling
workshops on their farms have been invited to see demonstrations ofintegrat-
ed crop-fish systems at the National Aquaculture Research Station. In
December 1990, 17 farmers, who were practicing aquaculture, were invited
to see rice-fish systems at the station. At that time, no farmers in Malawi had
been practicing rice-fish culture even though many had rice paddies adjacent
to their ponds. The farmers were treating each enterprise in terms of strict
commodity production and not as resource systems which could enhance each
other if integrated.
The usual practice is to grow one crop of rice per year between December
and June. The visiting farmers were seeing rice that had been grown during
the period of the hottest, driest months of the year (August-December). This
was possible because the rice was grown in ponds. Farmers were impressed by
the arrangement whereby two crops (rice and fish) were gathered from the
same pond and rice was grown during the dry season. After viewing the
harvest, they led a workshop to discuss the demonstration and its relevance to
their farming environment. The result was a series of drawings of farmer-
designed systems (Figures 5 and 6). Common to all of these designs were
simple drainage arrangements for decoupling rice and fish. Prior exposure to
farm mapping workshops of potential or actual integrated aquaculture-
agriculture systems enabled the farmers to design elegant and simple rice-fish
arrangements appropriate to their rural environment.
Without further input or encouragement from the researchers, 11 of these
farmers experimented and developed their own unique rice-fish systems over
the succeeding 12 months. These systems all incorporate sloping pond
bottoms and drainage, which allows easy withdrawal of water from the rice
paddy and collection of fish in a deep water refuge. Now farmer-to-farmer
technology transfer is occurring such that 40 farmers that had not attended
the open day are practicing rice-fish culture. Furthermore, many farmers are
now growing dry season crops of rice in their ponds. This new development
in integrated resource management has been the result of a one-day modeling

Journal for Farming Systems Research-Extension


Rice ridges only allow
fish to move up and
down paddy

Figure 5. Drawings By Two Farmers of a Possible Rice-Fish Arrangement
(source: Noble and Rashidi, 1990).

Vol. 4, No. 1, 1993


workshop on farm in combination with a one-day demonstration of the
technology on station. Further open days are now conducted partly on station
and partly on farm so that farmers actively involved in the development of
integrated systems incorporating vegetable-fish, rice-fish, etc. can relate their
experiences to novices to the technology.
A valuable outcome of these workshops and demonstrations is that farmers
do not consider ponds simply as fish production units, but also as water storage
for irrigating crops, watering livestock, and household use. Farmers have also
realized that agricultural residues recycled through ponds produce highly
fertile mud that can be transferred to adjacent vegetable gardens and reduce
the need to use chemical fertilizers. Thus the modeling and mapping work-
shops, together with the demonstrations, provide farmers with a forum for
peer exchange and discussion about new ways to reallocate and integrate their
resources and enterprises. Farmer experimentation resulting from this partic-
ipatory process leads to increased efficiency in utilization and recycling ofon-
farm residues, rehabilitation of exhausted soils, and reclamation of marginal
land for agricultural production.
Outlet- _
Rice ridge
Low dike between pond and field
Figure 6. Sloping Ricefield Eases Driving ofFish Into Pond So Rice Can Be Harvested.
(Authors' Interpretation of Farmer's Drawing; Source: Noble and Rashidi, 1990).

Farmer Evaluation and Participatory Monitoring
As farmers experiment in manipulating resource systems and enterprises,
the problem remains as to how and what should be measured as indicators of
sustainable improvement on the farm. Any evaluation or monitoring program
will require full participation of the farmers in its design if it is to be effective
in capturing the evolution of small-scale farming systems. Farmer bioresource
models illustrating in their totality resource systems, enterprises, and their
interrelationships provide an effective format for developing a monitoring
system. Farmers can update their models on a regular basis using the initial
plan diagram of their farm. Changes in flows within and between resource
systems, outputs from them and external inputs to them, can all be captured
when farmers produce time-series models of their farming operations. Draw-
ings will include the resource flows for the time period since the last diagram
was drawn. Wherever possible, the amount of material, frequency of flow, and
its monetary value are recorded. Quantities can be given in local terms and
then converted into kilos or reasonable estimates attempted. This procedure
enables both the farmer and researcher to monitor progress and change. Such
a process also provides farmers with a tool for improving their decision making
and skills in resource management.

Journal for Farming Systems Research-Extension


Integration of enterprises and improved recycling of on-farm bioresources
should lead to increased efficiency of energy and nutrient utilization with
concomitant rehabilitation of farming ecosystems. Greater stability in agricul-
tural production should ensue with a corresponding reduction in vulnerability
to environmental perturbations. Time-series modeling of farms as mentioned
in section 2 provides one method for capturing some of these changes and
assessing whether increased diversification of linkages between elements in the
farm system really stabilizes and improves farm production.

Impact on Households
In Malawi, a joint research program with farmers has just started to assess
the impact of integrated aquaculture-agriculture development on small-scale
farming systems. Detailed results are not yet available but some preliminary
work indicates that ponds linked to rice and vegetables may generate signif-
icant improvements for the farming household and its surrounding environ-
ment. Using methodology described in section 2, farmers have been exposed
to various crop-fish integration, particularly rice-fish, as described in Noble
and Rashidi (1990). Interest in integration is intense and many farmers have
altered their farming environment too quickly for researchers to capture the
To understand the full impact of these changes, as a preliminary sample, five
farmers who are relatively new to aquaculture systems were asked to make
drawings of their farms before and after incorporation offish ponds (Figures
7 and 8). Significant alterations have taken place in farm management due to
the incorporation of ponds. Farming systems now have increased linkages
between their elements with the rice-fish ponds acting as the focus or flow of
on-farm bioresources. Marginal land has been brought into production for
rice and fish, and has increased the household income directly through sale of
fish from the revitalized dambo resource system. These patterns were com-
mon to four of the farms sampled and each case involved bioresource linkages
to adjacent dimba vegetable gardens with flows operating in both directions
between the pond and dimba systems. Although these dambo and dimba
wetlands are individually small, often less than 0.5 ha, added together they
comprise 10-20 percent oftotal land area in African savannas (Scoones, 1991).
Table 1 summarizes the information obtained from the farm drawings. In
every case, the presence of ponds results in the initiation of integration
creating linkages through recycling of farm residues. The contribution such
systems make to household nutrition has not yet been assessed, but all of the
farmers indicated that significant output from both rice and fish goes to the
household. Farmers commented that having ponds reduced their need to buy
fish and the latter were often used for reciprocal exchange of goods and
services provided by neighbors.

Vol. 4, No. 1, 1993


In three sets of drawings, farmers provided detailed information on cash
flows coming from each resource system before and after the introduction of
fish ponds (Table 2). Rice-fish ponds' contribution to gross income varies
widely, from 10 to 62 percent, reflecting differing levels of investment and
length of time the aquaculture systems have been running.
Farm A. Three borrowed cropland areas (munda) were repossessed from
the farmer by their owners. To offset this loss of productive lands, the farmer
converted marginal dambo land into an integrated rice-fish system. Decline in
income due to loss of productive munda is more than compensated by income
from the rehabilitated dambo (Table 2) even though the dambo area is smaller
in size than the original three munda areas that were lost. The farmer does
incur expenses associated with the rice-fish enterprise of about US$23, giving
a net income of US$148. However, the full value of rice and fish from the
ponds is not reflected in Table 2. The household eats produce from the ponds
and also provides free fish to relatives. The estimated market value of


e Rie. veWcI season
Prow 10 Hooding


Figure 7. Retrospective Resource Diagram Showing Farm B Prior to Having Fish Ponds.

Journal for Farming Systems Research-Extension


consumed fish alone is approximately US$43. Therefore, the total net value
of produce from the ponds is around US$190.
Farm B. Rehabilitation of the dambo area for rice-fish ponds provides an
additional income to the household equivalent to or even slightly more than
the income from the munda, raising total gross family earnings from US$135
to US$155. Again, the total value of produce from the ponds is not reflected
in the table. Consumption of rice and fish by the family was valued at
approximately US$7, raising the total value of produce from the ponds to just
over US$160/year. The farmer has only just started rice-fish and is still in the
process of developing the system.
Farm C. Marginal dambo was rehabilitated and provided additional
income and food for the household. This cash was used to purchase fertilizer
for the munda, resulting in increased crop production and more than a
doubling of income from the munda crops. Recycling of pond mud to the
dimba raises the latter's income from US$46 to US$83, although as a


Figure 8. Current Resource Flows on Farm B After Incorporation of Fish Ponds.

Vol. 4, No. 1, 1993


percentage of total income it declines slightly. On all three farms, the presence
of ponds seems to have altered the farm system and earning potential of the
household. The average annual income of US$155 increased toUS $235 after
the adoption of aquaculture. Experiment in International Living (1991)
reported that the average annual income for rural households in Malawi was
US$130. Although one cannot draw any firm conclusions from three farms,
it would appear that integrated aquaculture systems may have the potential to
significantly raise rural incomes.
Tables 3 and 4 provide more detailed data on production and income from
14 rice-fish systems in Malawi. The majority of farmers have 200m2 ponds of
which half is planted with rice. Such systems have the potential to generate
incomes ranging from approximately US$120 to US$250 per annum depend-
ing on whether farmers operate one or two cycles of rice and fish in their
ponds. This income range is just for rice-fish and does not include earnings
from other resource systems on the farm. This earning potential represents a
mean annual production of about 50 to 100 kg of rice and 25 to 50 kg offish.
The average number of people per rural household is 4.3 (NSO, 1991,
1992b). Therefore, rice-fish systems potentially can provide 12-23 kg of rice
and 6-12 kg of fish per person per annum. ICLARM-GTZ (1991) stated that
the average daily protein requirement for an adult is 35-43 g or 12.8-15.7 kg/
capita/year. Rice-fish can produce almost the total protein requirement for a
family of four.
Table 1. Comparison of Changes In Farming System Prior to and Following Integrated
Aquaculture Development, Based on Farmer's Drawings.

Marginal wet land unutilized.
No integration between resource
systems or on-farm enterprises.

Crop residues not recycled.

Water shortage for vegetable
garden in late dry season.
Household reliant on uncertain
water supplies.
Reliant either on fertilizer for vegetable
garden or overutilized exhausted
Buy fish; rarely eat fish.

Marginal wetland does not provide food
and income for household.
Rice either not grown or only one
crops of rice.

Marginal land brought into productive use.
Ponds serve as a focal point for direct or
indirect links between resource systems.

Crop residues such as maize bran and green
leaf waste used as pond input.
Ponds placed adjacent to gardens provide
water for irrigating vegetables.
Households use pond as water catchment for
domestic use and watering livestock.
Use pond as processing unit for converting
low quality crop waste into fertile mud for
transfer to garden; reduced fertilizer use.
Ready supply offish for household
consumption; rarely buy fish.
Conversion to ponds provides food
and income.
Rice-fish ponds provide two crop per year;
rice grown for first time.

Journal for Farming Systems Research-Extension


Table 2. Impact of Integrated Aquaculture Development on Gross Income of Three
Farms (in US$).
Contribution to cash income of each resource system (%).

Farm A ($92)
Total earnings $134/year

Without ponds

Farm B ($83) ($32) ($20)
62 23 15
Total earnings $134/year
Farm C ($71) ($46)
61 39
Total earnings S117/year
With ponds
Dimba Munda Rice-Fish
(vegetables) (crops) Livestock integration
Farm A ($100) ($5) ($171)
36 2 -62
Total earnings $276/year
Farm B ($69) ($34) ($24) ($28)
44 22 16 18
Total earnings $155/year
Farm C ($166) ($83) ($28)

60 30 -10
Total earnings $275/year
Note: All farmers above are classified by Malawi Government as in the low-income index bracket of less
than $30/month.
*Rice/Fish integration in Farm A replaces unprofitable off-farm business.

Impact on Resource Systems
There is only indirect on-station evidence that development of integrated
crop-fish systems may have a beneficial impact on resource systems. Detailed
on-farm analysis has yet to be done. Chimatiro (unpublished data) did carry
out some on-farm studies coupled with on-station experiments that demon-
strated that ponds act as efficient converters of low-quality vegetable waste
into nutrient-rich pond mud. Chimatiro found that the ecological efficiency
with which vegetable waste was converted into fish flesh in farm ponds was
only 0.8 percent. He assumed that probably a large portion of plant material
was being incorporated into the detritivore subsystem of the pond and
accumulating in the sediments.
Chimatiro carried out a simple test of this hypothesis on-station. Sediments
from ponds that had received grass inputs for six months were used as fertilizer

Vol. 4, No. 1, 1993





to grow head cabbages on ordinary topsoil and compared with cabbages
grown on topsoil without the addition of pond mud. Cabbages grown on soil
receiving pond sediments had yields equivalent to 1.2 t/100/m2/year, a 50
percent increase over the 0.8 t/100/m2/year produced on ordinary dambo
soil. Chikafumbwa (unpublished data) demonstrated that even higher yields
of 2t/100m2/year could be achieved with Chinese cabbage. These two simple
experiments indicate that soils adjacent to ponds can be rehabilitated by
fertilization with pond mud and can produce vegetables crops.
From the Chinese cabbage experiment, approximately 0.7 t/100m2/year
of leaf waste was produced that could be recycled back into the ponds as feed
and fertilizer. This operation regenerated the pond mud, which could then be

Table 3. Average Production of Rice Grain and Straw from Rice-Fish Integrated
Ponds in Zomba District, Malawi, 1991/92.
Rice Harvest Weight
kg grain = 28 (15) [8-60]; straw = 97 (61) [26-197]
kg/100m2 grain = 26 (8) [13-42]; straw = 81 (37) [38-198]
year grain = 55 (20) [35-96]; straw =190(115) [86-478]
*mean value of rice grain =41 K ($13)/100m2/year
Pond Statistics
Total area = 259 (167) [42-751]
Rice area = 124 (91) [40-386]
% Rice = 56 (29) [16-100]
Mean growing
period (days) = 164 (27) [1143-256]
Number of ponds = 14
Legend: ( ) = standard deviation; [ ] = range; *Last harvests were sold in May 1992 (1 kg rice =
0.72 K); US$1 = 2.3 KMay 1992 (NSO 1992b).

recycled back on to the garden soils. Farmers have adopted these practices of
recycling pond mud into the dimba garden system. In the case of Farm C,
transfer of pond mud to the dimba garden has raised the income from the
resource systems by almost 20 percent. Even building a pond on marginal land
will improve the fertility of that land through the accumulation of rich pond
sediments. Coupled with the recycling of those sediments to other adjacent
marginal areas creates a real potential for system rehabilitation and possible
stabilization of agricultural production from the dambo and dimba resource

Sustainability Evaluation
Although the data from rice-fish ponds and farmers' models seem to
indicate improvement in farm production through the adoption of integrated
aquaculture-agriculture systems, there is no evidence as yet that these systems
are necessarily sustainable.

Journal for Farming Systems Research-Extension


Defining sustainability is a complex issue. As Conway (1991) comments,
definitions vary between disciplines and almost anything goes if it fits within
the broad umbrella of self-sufficiency, integration, traditional farming, etc.
Harrington (1991) cautions that approaches to evaluation and measurement
will be strongly influenced by how sustainability is defined and interpreted.

Table 4. Average Fish Production and Harvest Values From Rice-Fish Ponds in Zom-
ba District, Malawi, 1992.
Fish Production, Harvest Weight
(kg) 14 (7) [6-25]
Kg/200m2 13 (6) [5-22]
kg/200m2/yr = 27 (12) [24-44]
Fish Production, Harvest Values
Mean value of fish
/200m2/year = 82 K ($22.00) 1 harvest
165 K ($45.00) 2 harvests*
Mean price/kg = 3.03 K ($00.82/kg)
Pond Statistics
Mean area m2 211 (19) [181-232]
Mean growing period
(mo.) for 1 harvest 5 (2) [3.4-9.4]
Number of ponds 5
Legend: ( ) = standard deviation; [ ] = range; *very common ($1 = 3.7 K August 1992).
For small-scale farmers in Africa, the major problem is maintaining the
integrity of the farming system in the face of rapidly degrading socioeconomic
and ecological conditions. A reasonable measure of sustainability under these
circumstances would be the resilience of a farming system to such environ-
mental perturbations. Conway (1967) defines resilience as the ability of an
agricultural system to "maintain its productivity when subject to stress or
If one settles on resilience as the criterion of sustainability, how and what
should be measured to evaluate whether a farming system is resilient?
Harrington (1991) points out that assessing a system's resilience will depend
on developing reliable, quantifiable, and easily measured indicators of resil-
ience. Giampictro et al. (1992) emphasize that the stability or resilience of
systems relative to internal or external fluctuations is an important parameter
to assess. They take a strictly ecological approach and state that energy flow
and storage are critical parameters to measure in assessing a system's stability
and sustainability. They coin the term "biophysical capital": the ability of an
ecosystem to use solar energy to run biophysical processes that stabilize system
structure and functions (i.e., maintain resilience). Conway's definition of
resilience is based on the premise that diversity within systems will enhance
their ability to withstand perturbation. Diversity spreads risk so that if one
enterprise fails, others will still operate and thus maintain productivity of the

Vol. 4, No. 1, 1993


system. But diversity also implies links, integration, recycling, efficient use of
nutrients, etc. in order to reduce risk. Therefore, a measure of these attributes
and how they change over time may provide researchers with some guides as
to the sustainability of a farming system.
In the context of the methodology described earlier, resilience may be
measured through time-series modeling by farmers. Table 1 already provides
some of the parameters that could be used. From farm models, change in
number and type of enterprises can be monitored over time. Regular counts
can be made of bioresource flows and the quantities of materials they
represent. Outputs and inputs in terms of labor and fertilizer can be estimated
from drawings with the farmer's help. The following criteria could be used to
measure improvement in sustainability of farming systems:
1. Measure productive output of marginal lands. Increase in agricultural
output will signify rehabilitation.
2. Measure increase in number of interlinkages between and within
resource systems. Increase in linkages will indicate that more material is being
recycled within the system and there is possibly less dependence on external
3. Measure quantities of bioresources flowing between resource systems.
This will give an indication of allocation and efficiency of nutrient flow and/
or energy flow within the farming system.
4. Measure output from resource systems and see if combined outputs
remain stable or increase.
5. Measure net cash income flows on farm diagrams. If they are increasing
or maintaining their value relative to inflation, then sustainability of the
farming system has probably increased
6. Measure net flows of produce to household as food. If these are stable
or increasing, then changes in the farm system are having a significant impact
on household nutrition.
7. Measure labor allocation relative to productive output and income
generation from different resource systems and enterprises. If returns to
limited labor are increasing, then system sustainability is probably increasing.
Labor allocations may also provide some measure of equity changes.
8. Measure material output from resource systems relative to input. If
output efficiency is greater in energy or nutrient terms, then integration is
improving sustainability of the system.
9. Measure increase in flows and linkages of materials, labor, etc. between
the farm household and those of neighbors and relatives. If linkages are not
only improving the efficiency of nutrient, cash, and labor use on the farm
implementing the changes but also benefiting other socially closely linked
households, then sustainability of the system is probably increasing and equity
is improving.
This is not an exhaustive list of possibilities, but an indication of what
information could be gleaned from farmer diagrams when the farmer is an

Journalfor Farming Systems Research-Extension


active participant in monitoring change and perceives a benefit in the process.
Using time-series farmer models of their farm systems for evaluation of
sustainability also addresses the problem of dealing with real world complex-
ity. Quantification of sustainability is often rejected because researchers
choose criteria that are quantifiable but not necessarily as crucial as concep-
tually more important "non-quantifiable" ones (Harrington, 1991). This
situation is avoided with farmer models because the farmer can indicate on
diagrams non-quantifiable changes such as increased social linkages and
reciprocation with neighbors and relatives as people and/or materials and cash
flow within and between farming systems.
Farmer drawings also alleviate some of the problems'associated with a total
factor productivity approach (TFP) to measuring sustainability. TFP relates
total outputs (0) to total inputs (I) as TFP = O/I and assumes that
sustainability is increasing if O>I. This approach does not readily measure
changes in quality of the resource base or change in diversity of components
within the system, whether they be social, economic, or environmental. Time-
series models of farms, using a variety of drawings to indicate different
component changes in the farming system, have the potential to capture and
quantify many changes not picked up by a straight TFP approach.
Likewise, data collected from drawings can be converted into biomass and
energy flows which will enable one to measure the "biophysical capital," as
discussed by Giampietro et al (1992). They state that stability of living systems
is related to their ability to maintain their structure and function by replacing
the flow of energy and matter discarded by systems with new energy and
matter. In a strict ecological sense, for a farming system to be stable or
sustainable, these criteria are essential. The evaluation procedures indicated
above have the potential to measure these criteria.

Although this research may be considered to be in its infancy, several positive
signs have emerged that encourage further work. Like many others, we have
witnessed an empowerment of farmers through participatory research. Given
that farmers are the managers of natural resources, this can only be encourag-
ing. We have seen farmers change decisions away from enterprise profitability
to favor resource system rehabilitation. We have observed that managing
water and living aquatic resources provides an entry point for sustainable
farming system development. We have grasped the utility of quantifiable
sustainability indicators for comparing performance of different farming
systems. Through ICLARM's research in other countries in Africa and Asia we
have uncovered the universality of resource systems. A vital output at this stage
of our understanding of sustainable agriculture is to find commonalities using
multiregional evaluations. Commonalities will lead to generalizations about

Vol. 4, No. 1, 1993


sustainable agriculture that can then be tested empirically, and thus a set of
principles for sustainable agriculture can emerge.
On-farm experiments for sustainable agriculture do not look like conven-
tional experiments. Our experiments in resource management have a new
sequence of (1) resource inventory and analysis, (2) exposure to alternative
resource use techniques, and (3) monitoring farming system transformation.
These experiments have the objective of changing resource management
decisions in a favorable direction. Here, favorable means toward less reliance
on nonrenewable resources, less environmental degradation, increased whole-
farm profitability, reduced pollution, and reduced adverse social and econom-
ic effects. Researchers monitor indicators that we hypothesize will lead to
sustainable farming systems, a hypothesis that still has to be tested. Neverthe-
less, we do know the impact of these experiments on households and resource
systems. We do not know how long this kind of research must run for
hypotheses to be tested. Thus, we share an idea in its early stages and not a
finished product. Other shortcomings we anticipate follow.
We feel insufficient attention has been given to measurement of environ-
mental effects. If resource systems are to be valued then we need to know more
about biological and ecological processes that underlie their degradation and
rehabilitation. Concomitant with this need is a failure to analyze off-farm
social and economic effects. At some time we must work at hierarchies greater
than the farm and the resource manager. Particularly important here will be
studies of markets, policies, and equity. We do, however, appreciate that
research cannot address all issues in the same test. Nevertheless, resource
management experiments as we have envisioned them must open doors for
other issues to be studied through collaborate research efforts. If what we have
written is worth following up on, then technical issues will pale in comparison
with the institutional hurdles that face us.
Experiments of the kind described need wide-scale farmer participation.
Establishing a large population of farmers requires the skills of grassroots,
nongovernmental organizations (NGOs). We see little chance of this kind of
work being done without partnerships between governmental organizations
(GOs) and NGOs. Vital linkages within research institutions themselves
include links between on-station researchers and on-farm researchers, links
between social and biological science researchers working at higher and lower
levels in the agricultural systems hierarchy. Links with policymakers are also
vital if policies to protect and promote sustainable systems are to be formulat-
ed. Over and above all these institutional constraints we see the greatest hurdle
as the reluctance of scientists to change research styles that they have used for
years and that are reinforced by peers, journals, and professional societies.

Journal for Farming Systems Research-Extension



Thanks to ICLARM technical assistants Mr. A. Montjeza and Mr. R. Selemani
and field assistant/driver Mr. W. McLorry for data gathering on rice-fish
ponds. Thanks to Fredson Chikafumbwa for making available his material on
vegetable-fish integration. Thanks also to Emma Mafuleka, Fredson Chikaf-
umbwa of ICLARM and Mr. Kalumpha, Fisheries Department, for assistance
in farmer modeling sessions and to Albert Contemprate of ICLARM for
drawings and figures. Particular thanks should go to Nancy Axinn and Larry
Harrington for their valuable comments on the drafts of this paper.


Altieri, M. 1987. Agroecology: The scientific basis of alternative agriculture. Boulder,
Colorado: Westview Press.
Anderson, M.D., and W. Lockeretz. 1992. Sustainable agriculture research in the ideal
and in the field. Journal of Soil and Water Conservation 47(1):100-104.
Batie, S.S. 1989. Sustainable development: Challenges to the profession of agricultural
economics. Presidential address at the AAEA Summer Meeting, Baton Rouge,
Louisiana, July-August.
Chambers, R., A. Pacey, and L.A. Thrupp, eds. 1989. Farmer first: Farmer innovation
and agricultural research. London: Intermediate Technology Publications.
Chikafumbwa, unpublished data
Chimatiro, unpublished data.
Conway, G.T. 1967. The properties ofagroecosystems. Agricultural Systems 24:95-117.
Conway, G.R. 1986. Agroecosystem analysis for research and development. Bangkok,
Winrock International.
Conway, G.R. 1991. Sustainability in agricultural development: Trade offs with produc-
tivity, stability and equitability. Paper presented at the 11th Annual Association for
Farm Systems Research-Extension Symposium, Michigan State University, East
Lansing, October.
Edwards, C.A., R. Lal, P. Madden, R.H. Miller, and G. House, eds. 1990. Sustainable
agricultural systems. Soil and Water Conservation Society, Ankeny, Iowa.
Experiment in International Living. 1991. The roles of NGOs in promoting agricultural
development and natural resource management in Malawi. Proceedings of United
States Agency for International Develpment/SHARED Workshop, May.
Giampietro, M., G. Cerretelli, and D. Pimentel. 1992. Energy analysis of agricultural
ecosystem management: Human return and sustainability. Agriculture, Ecosystems
and Environment 38:219-244.
Harrington, L.W. 1991. Measuring sustainability: Issues and alternatives. Journal for
Farming Systems Research-Extension 3(1):1-20.
Haverkort, B., J. van der Kamp, and A. Waters-Bayer, eds. 1991. Joining farmers'
experiments: Experiences in participatory technology development. London: Intermedi-
ate Technology Publications.
ICLARM-GTZ (International Center for Living Aquatic Resources Management and
the Government of the Republic of Zambia). 1991. The context of small-scale

Vol. 4, No. 1, 1993


integrated agriculture-aquaculture systems in Africa: A case study ofMalawi. ICLARM
Stud. Rev. 18. 302 pp.
Lightfoot, C., and N.A. Tuan. 1990. Drawing pictures of integrated farms helps
everyone: An example from Vietnam. Aquabyte 3(2):5-6.
Lightfoot, C., R. Noble, and R. Morales. 1991. Training resource book on a participatory
method for modelling bioresource flows. International Center for Living Aquatic
Resources Management (ICLARM) Educational Series 14, 30pp.
Lightfoot, C., and D. Minnick. 1991. Farmer-first qualitative methods: Farmers'
diagrams for improving methods of experimental design in integrated farming
systems. Journal for Farming Systems Research-Extension 2(1):57-71.
National Statistical Office (NSO). 1991. Malawi population and housing census 1987.
Malawi: Government Printer, 95 pp.
National Statistical Office (NSO). 1992a. Annual survey of agriculture 1982/83-1984/
85. Malawi: Government Printer, 146 pp.
National Statistical Office (NSO). 1992b. Monthly Statistical Bulletin May 1992.
Government Printer, 27 pp.
Noble, R., and B. Rashidi. 1990. Aquaculture technology transfer to smallholder farmers
in Malawi, Southern Africa. The ICLARM Quarterly 13(4):14-16.
Scoones, I. 1991. Wetlands in drylands: Key resources for agricultural and pastoral
production in Africa. Ambio 20(8):366-371.

Journal for Farming Systems Research-Extension

Integrating Household Food Security into
Farming Systems Research-Extension'

T.R. Frankenberger and P.E. Coyle 2

Agricultural development projects can aid low-resource farmers in coping
with household food insecurity. The small-scale rainfed farming systems
of such farmers, however, are extremely complex, and so researchers and
extensionists need to take account ofmultiple production and consump-
tion linkages, coping strategies, and environmental constraints in helping
to ensure sustainable access to food for members of these farm house-
holds. To do this, a number ofindicators ofhousehold food security can
be aggregated into vulnerability maps of food insecure populations, and
these, in combination with rapid and participatory rural appraisals, can
systematically identify households at risk of food shortage. But to
effectively integrate household food security concerns into FSRE, it is also
important to elicit the active participation of farmers in problem identi-
fication and technology development through collaborative research and
on-farm experimentation

The importance of household food security (HFS) to agricultural develop-
ment efforts has been drastically accentuated by the current severe drought
now plaguing Southern Africa. The fact that nearly 18,000,000 people are
food insecure and at risk of severe malnutrition (Green, 1992) justifies the
priority households give to securing sufficient food supplies as a major
production goal. The recurrent risks associated with fluctuating rainfall and
unstable markets has led many farmers to diversify their food procurement
strategies in order to secure a wide food base and sufficient supplies (Velarde,
1991). Farmers' pursuit of these diverse strategies to meet their food needs has
significant implications for the types of interventions promoted through
farming systems research-extension (FSRE).
1 Keynote presentation at the Twelfth Annual Association for Farming Systems Research-
Extension Symposium, Michigan State University, East Lansing, September 13-18, 1992. This
paper was initially prepared for The Famine Mitigation Activity Support Project, USAID, Office
of US Foreign Disaster Assistance, Washington, DC. It was previously published in Nutrition
and household food security in farming systems research, Mansa, Zambia, August 10-14, 1992,
Ministry of Agriculture, Food, and Fisheries, Adaptive Research Planning Team, Zambia.
Office of Arid Lands Studies, University of Arizona, Tucson, Arizona, USA 85719.


This paper addresses a number of the key issues related to household food
security that have a direct bearing on the work carried out by farming systems
practitioners. It begins by summarizing a number of conceptual issues
associated with HFS that need to be taken into account in agricultural
research. This is followed by a brief discussion of the client group to which
most FSRE activities are aimed, and how their participation in the research
process is key to the promotion of HFS. This paper then concludes with a
number of suggestions of how HFS considerations can be incorporated into
the FSRE process.


Household Food Security and Livelihood Security
Food security is defined by the World Bank (1986) as "access by all people
at all times to enough food for an active and healthy life." Operationalizing
the concept at the national level is not the same as at the household level. At
the national level, food security entails adequate food supplies through local
production and food imports. However, adequate availability of food at the
national level does not necessarily translate into even distribution across the
country, nor equal access among all households.
In the past 15 years, much conceptual progress has been made in our
understanding of the processes that lead to food insecure situations for
households (Frankenberger, 1992). We have moved away from simplistic
notions of food supply being the only cause of household food insecurity to
assessing vulnerability of particular groups in terms of their access to food.
Thus, food availability at the national and regional level and stable access are
both keys to household food security (see Figure 1). Access to food is
determined by food entitlements that may include viable means for procuring
food (either produced or purchased), human and physical capital, assets and
stores, access to common property resources, and a variety of social contracts
at the household, community, and state level (Maxwell et al., 1992). The risk
of entitlement failure determines the level of vulnerability of a household to
food insecurity. The greater the share of resources devoted to food acquisi-
tion, the higher the vulnerability of the household to food insecurity.
However, household food security is but one dimension of livelihood
security. Livelihood is defined by Chambers (1989) as adequate stocks and
flows of food and cash to meet basic needs. Poor people balance competing
needs for asset preservation, income generation, and present and future food
supplies in complex ways (Maxwell et al., 1992). People may go hungry up to
a point to meet another objective. For example, deWaal (1989) found during
the 1984-85 famine in Darfur, Sudan that people chose to go hungry to
preserve their assets and future livelihoods. People will put up with a

Journal for Farming Systems Research-Extension

- Household
Food Security


Availability Stable Access
a O

SiStorage, 0
o Conservation Processing 1

Sufficient Local Supply Viable
CL (production + gathering Household Sstainabilit S aial
+ purchased + food aid) Procurement Sustainabity Sustainabi

Public Private Local
Access Non- Access Acess Access Natural Natural Informal Stable
SFood Adequacy to Prod- farm to Gatred Shed Resource Resource Social Govem-
Assets auction income credit Food Food Conser- Conser- Mecha- ment
10- vation vation nisms

Cultural Food Nutritional
Acceptability Safety Adequacy


considerable degree of hunger to preserve seed for planting, cultivate their
own fields, or avoid selling animals (Maxwell et al., 1992). Similarly, Corbett
(1988) found that in the sequential ordering of behavioral responses em-
ployed in periods of stress in a number of African and Asian countries,
preservation of assets takes priority over meeting immediate food needs until
the point of destitution. Given the importance of livelihood security to farmers
in risk-prone areas, risk avoidance and entitlement protection must be built-
in selection criteria for screening technology.

Production/Consumption Linkages
To promote HFS in ongoing FSRE efforts, it is important to understand
the complex linkages that exist between production and food consumption.
Changes in farm management and production technology have not always
improved the food consumption status of producers (Frankenberger, 1990).
Agricultural development project are not nutritionally neutral. Some impor-
tant linkages are summarized below.
Crop diversity. As small farm households become integrated into a market
economy, the production of nonfood cash crops often replaces traditional
subsistence crops. This shift from subsistence to cash cropping may result in
decreased crop diversity and a concomitant increased dependency on outside
food sources. As a result, food consumption and nutritional status may be
adversely affected. Some other consequences of a shift to cash cropping
include: 1) less land available for food crop production; 2) a breakdown of
traditional foodsharing networks; and 3) the elimination of important minor
crops and wild plants that provide essential nutrients during the preharvest
period when staple foods are often in short supply (DeWalt, 1983; Dewey,
1981; Fleuret and Fleuret, 1980; Frankenberger, 1985; Longhurst, 1983;
Messer, 1989).
Income. Household income is a major determinant of family food con-
sumption. Factors such as the control and form of income, and the regularity
of its receipt, may be equally or more important than total income in
understanding the nutritional effects of agricultural development initiatives.
When women control household income, they are more likely to spend it on
food and health care. Continual or periodic forms of income are more often
spent on food than lump-sum income. In-kind (food) income is more likely
used for family consumption than cash income. Increasing income is often
associated with the increased consumption of purchased foods, especially
foods of animal origin. Diets dependent on purchased foods, however, do not
necessarily meet nutritional needs more adequately than diets that rely on
agricultural products and wild foods (Dewey, 1981; Fleuret and Fleuret,
1980; Frankenberger, 1985; Hcrnandcz et al., 1974; Kennedy and Cogill,
1987; Longhurst, 1983; Pinstrup-Andersen, 1981; von Braun and Kennedy,
1986; Saenz de Tejada, 1989).

Journal for Farming Systems Research-Extension


Seasonality of production. In most areas of the world there is a seasonal
dimension to agricultural production, food availability, and malnutrition.
Many farm families must cope with a cyclical period of deprivation referred to
as the "hungry season." This occurs in the weeks preceding harvest when food
stocks are low and food prices are high. Such periods of stress have a negative
impact on children's nutritional status and growth. Adults may lose as much
as seven percent of their body weight during the hungry season. This tends to
coincide with the agricultural cycle's peak labor period when a farmer's energy
expenditure is at its highest. Food shortages before harvest also coincide with
peaks in infection rates for diarrhea, malaria, and other debilitating diseases
(Frankenberger, 1985; Longhurst, 1983; Maxwell, 1984; Pinstrup-Anders-
en, 1981).
Role ofwomen in production. Women's participation in agricultural produc-
tion has an effect on family consumption and nutritional status that is closely
tied to the income earned and labor demanded by this activity. Most income
earned by women from agricultural activities is used for food purchases.
Children of working women are less likely to be malnourished than children
of nonworking women. However, activities that increase the labor-demands
on women's time may lead to changes in cooking habits, the preparation of
less nutritious and/or fewer meals, the cultivation of less labor-intensive and
less nutritious food crops, and less time devoted to breastfeeding and child
care (Fleuret and Fleuret, 1980; Frankenberger, 1985; Kumar, 1977; Long-
hurst, 1983; Maxwell, 1984; Tripp, 1982; von Braun and Kennedy, 1986).
Crop labor requirements. The introduction of new cash crops may require
more human energy than previously grown crops, and the added energy
requirements may be greater than the value of the output. These increased
energy demands could also have deleterious effects on intrahousehold food
distribution if some members of the household require a higher food intake
to meet labor demands (Fleuret and Fleuret 1980; Frankenberger 1985).
Food preferences. Improved crop varieties should have acceptable character-
istics for successful adoption by farm families. They should satisfy local tastes
in terms of flavor and texture, otherwise they are unlikely to be adopted for
subsistence and may only be produced for commercial purposes. New varieties
should also have acceptable cooking qualities. Varieties that take longer to
cook may require more fuel, water, and labor than indigenous varieties. In
addition, time- and resource-efficient preparation methods should be intro-
duced concomitantly to better ensure the use of new crop varieties to meet
consumption needs (Fleuret and Fleuret 1980; Frankenberger 1985; Tripp
Market prices. Market prices and market access can have a significant impact
on the consumption patterns of small-farm households. For example, in most
developing countries, high consumer prices coincide with food shortages in
small-farm households. In addition, government price and trade policies may
adversely affect domestic producer prices that, in turn, serve to keep the

Vol. 4, No. 1, 1993


purchasing power of farmers low. Finally, market inefficiency and periodic
market instability can place in a vulnerable position households that are
dependent on purchased food to meet their food needs (Fleuret and Fleuret,
1980; Frankenberger, 1985; Longhurst, 1988; Malambo, 1987).
By understanding these linkages, farming systems researchers will be more
cognizant of the unexpected effects that newly introduced production alter-
natives could have on consumption (Frankenberger, 1985). However, not all
linkage will be important in every context. To determine which linkages are
important, it is essential to understand the coping strategies farmers pursue to
maintain HFS.

Coping Strategies
Households do not respond arbitrarily to variability in food supply. People
who live in conditions that put their main source of income at recurrent risk
will develop self-insurance coping strategies to minimize risks to their HFS
and livelihoods (Longhurst, 1986; Corbctt, 1988). Examples of such strate-
gies are dispersed grazing, changes in cropping and planting practices,
migration to towns in search of urban employment, increased petty commod-
ity production, collection of wild foods, use ofinterhousehold transfers and
loans, use of credit from merchants and money lenders, migration to other
areas for employment, rationing of current food consumption, sale of posses-
sions (e.g., jewelry), sale of firewood and charcoal, consumption of food
distributed through relief programs, sale of productive assets, breakup of the
household, and distress migration. In general, coping strategies are pursued
by households to ensure future income-generating capacity (i.e., livelihood)
rather than simply to maintain current levels of food consumption (Corbett,
1988; deWaal, 1988; Haddad ct al., 1991). These strategies will vary by
region, community, social class, ethnic group, household, gender, age, and
season (Chambers, 1989; Thomas et al., 1989). The types of strategies
employed by households will also vary depending upon the severity and
duration of the potentially disruptive conditions (Thomas ct al., 1989).
In analyzing varieties of coping strategies, it is important to distinguish two
types of assets that farmers have at their disposal. Assets that represent stores
of value for liquidation (liquid assets) are acquired during noncrisis years as a
form of savings and self-insurance; these may include small livestock or
personal possessions such as jewelry. A second set of assets are those that play
a key role in generating income (productive assets). These are less liquid as
stores of value and are much more costly to farm household in their disposal.
Households will first dispose of assets held as stores of value before disposing
of productive assets (Corbett, 1988; Frankenberger and Goldstein, 1991). A
household's access to assets is often a good determinant of its vulnerability
(Chambers, 1989; Swift, 1989a).
Swift has also identified claims as another type of asset used by households
to assure their food security. Claims refer to the ability of households to

Journal for Farming Systems Research-Extension


activate community support mechanisms. Claims also may encompass govern-
ment support mechanisms or the international donor community (Borton and
Shoham, 1991).
Most initial responses to actual or potential food shortages are extension of
practices conducted in some measure during normal years to adapt to rainfall
variability (Longhurst, 1986; Watts, 1988). Traditional methods of handling
risk can be divided into routine risk-minimizing practices and loss-manage-
ment mechanisms (Walker and Jodha, 1986). Risk-minimizing practices are
adjustments to production and resource use before and during a production
season. These involve such practices as diversification of resources and
enterprises, and adjustments within cropping systems. Crop-centered diver-
sification can include choice of crops with varying maturation periods,
different sensitivities to environmental fluctuations, and flexible end use
products. Farmers also will reduce production risks by exploiting vertical,
horizontal, and temporal dimensions of the natural resource base. Vertical
adjustments involve planting at different elevations in a topographical se-
quence. Spatial risk-adjustments include planting in different micro-environ-
ments or in tercropping. Temporal risk adjustments involve staggering planting
times. Adjustments also may include extension of farming to marginal areas
or overuse of a particular plot, practices that can have a destructive effect on
the natural environment.
Loss management mechanisms include farmers' responses to lower-than-
expected crop production caused by natural hazards (Walker and Jodha,
1986). Reductions in crop production can be compensated for through
nonfarm income, the sale of assets, the management of stocks and reserves,
seasonal migration, and reciprocal obligations among households. Overcx-
ploitation of certain resources (forest reserves, for example) for market sale
also may be part of this loss management strategy.
In communities marked by landholding and income inequalities, house-
hold responses occur differently along the lines of wealth and access to
resources (xonghurst, 1986; Tobert, 1985). Identical climatic conditions can
affect households of varied economic levels to different degrees. Seasonal
shortages for some families produce famine conditions for others. Poorer
households, including many women-headed households, having smaller
holdings and a weaker resource base, are more vulnerable to stress than are
wealthier households, and begin to suffer earlier when food shortages hit
(Frankenberger and Goldstein, 1990). The poor resort to early sales of
livestock, pledge farms, incur debt, sell labor, and borrow grain at higher
interest rates. In essence, crop failures and other shocks reveal rather than
cause the fragile nature of HFS among vulnerable rural families. At the same
time, prosperous households buy livestock at deflated prices in conditions of
oversupply, sell or lend grain to needy farmers, purchase wage labor at
depressed rates, and purchase land (Watts, 1988). Thus, during a food crisis,

Vol. 4, No. 1, 1993


a cycle of accumulation and decapitalization can occur simultaneously within
a single community, depending on the depth of the current crisis.
Patterns of coping strategies can be diagramed to show the sequence of
responses farm households typically employ when faced with a food crisis
(Figure 2). These sequences of response are most frequently divided in the
literature into three distinct stages (Corbett, 1988). In the earliest stages of
a food crisis (stage one), households employ the types of risk-minimizing and
loss-management strategies discussed above. These typically involve a low
commitment of domestic resources, enabling speedy recovery once the crisis
has eased. As the crisis persists households are increasingly forced into greater
commitment of resources just to meet subsistence needs (stage two). There
may be a gradual disposal of key productive assets, making it harder to return
to a precrisis state. At this stage, a household's vulnerability to food insecurity
is extremely high. Stage three strategies are signs of failure to cope with the
food crisis and usually involve destitution and distress migration (Corbett,
Recent studies have found that the range of coping strategies pursued by
farm families in drought-prone areas may be changing over time (Downing,
1988; Thomas ct al., 1989). Three major trends appear to be developing.
First, risk-minimizing agricultural strategies appear to be narrowing in some
locations (e.g., in Kenya) as repeated sale and reacquisition have depleted
domestic and productive asset levels (Frankenberger and Goldstein, 1991). In
these areas, agricultural coping strategies are being replaced by strategies that
diversify income sources through off-farm employment and nonagricultural
production (Mead, 1988; Swinton, 1988). Some of these nonfarm strategies
include practices that are known to be environmentally damaging, but that
provide a last resort in crisis conditions. Second, strategies that relied on social
support and reciprocity for overcoming food deficits are eroding due to the
integration of individual households into the cash economy (Thomas et al.,
1989). Third, a shift has been observed in the responsibility for coping with
drought from the individual household and local community toward the
national government through drought and famine relief programs (Franken-
berger, 1990). This trend is due in large part to the reduction in response
flexibility of small farm households (Frankenberger and Goldstein, 1991).

Household Food Security and Environmental Degradation
Although coping strategies may be seen in the short term as functional
adaptations to uncertain conditions and hence beneficial, some commonly
practiced strategies may have dire consequences for the natural environment
in the long run (Frankenberger and Goldstein, 1991). Particularly for poorer
farmers with limited resource endowments, the process of maintaining
household viability may be exacted at the expense of the natural surroundings.
Poor people often occupy ecologically vulnerable areas such as marginal

Journal for Farming Systems Research-Extension


drylands, tropical forests, and hilly areas. As drought conditions worsen and
conditions of food insecurity persist, the range of options available to
resource-poor farmers becomes more limited and inflexible. In such situa-
tions, questions of long-term environmental sustainability become secondary.
Day-to-day survival demands the use of any food procurement strategy
The exploitation of common property resources (CPRs) is particularly
important to resource-poor farmers for meeting household food security
needs. Wild leaves, roots, grains, bushmeat, and forest products provide
additional food sources, buffer seasonal shortages, and provide alternative
sources of income (Davis et al., 1991). These resources are relied upon heavily
during times of stress (Jodha, 1986). Therefore, the degradation ofCPRs and
loss through the encroachment ofprivatized agriculture has disproportionate-
ly affected the food security of the poor (Davis ct al., 1991).

Earlier Time of Occurrence- Later

High Permanent Outmigration Low
Selling Land

Migrating for Aid

Pledging Land
Sale of Domestic Assets
Commitment Borrowing Grain or Cash from Merchant
of Domestic
Resources Selling Livestock Reversible

Dry Season Farming (Migration)
Migrating for Wage Work
Sell Labor Power
Using Stored Foods

Borrow Grain from Kin
Low lih
SUse of Famine Foodsigh

Figure 2. A Model of Responses to Food Shortage (Adapted from Watts, 1988).

Women are often more vulnerable to the effects of environmental degra-
dation than men because they are often more involved in the collection of
common property resources (Davis ct al., 1991). Since women often make a
greater contribution to household food security than men, a decline in

Vol. 4, No. 1, 1993


women's access to resources may have a significant impact on the nutritional
status of the household.
Coping strategies that may promote environmental degradation include
cutting trees to make charcoal, overharvesting of wild foods, overgrazing of
grasslands, and increased planting in marginal areas. All of these strategies may
degrade soil conditions and augment problems of soil erosion (Norman,
1991). Farmers often realize the damage their actions have on the environ-
ment upon which their livelihood depends. However, as drought conditions
worsen and food insecurity persists, the range of options becomes limited to
such desperation strategies.
Thus, vulnerability to food insecurity usually means vulnerability to envi-
ronmental degradation. However, development activities attempting to pur-
sue both household food security and environmental objectives must consider
the short- and long-term trade-offs associated with these dual objectives.
Long-term sustainable natural resource management initiatives will not be
successful if they ignore the short-term food security needs of the local
population. Likewise, sustainability will be compromised if long-term envi-
ronmental concerns are sacrificed for immediate food needs. For development
goals to be achieved, a balance must be struck between these two objectives.

Indicators of Household Food Security
As stated earlier, food availability and stable access are both critical to HFS.
For this reason, any particular monitoring system used for assessing HFS must
incorporate both food supply/production data and access/entitlement data
as part of its indicator set. Vulnerability to food insecurity is location specific,
therefore, indicators are needed that measure supply and food entitlement
changes at the local level.
A number of different indicators can be used for delineating HFS. These
can be divided into process indicators that reflect both food supply and food
access, and outcome indicators which serve as proxies for food consumption
(Frankcnberger, 1992) (see Matrix 1). Indicators that reflect food supply
include inputs and measures of agricultural production (agrometeorological
data), access to natural resources, institutional development and market
infrastructure (prices), and exposure to regional conflict and its consequences.
Indicator that reflect food access are the various means or strategies used by
households to meet their HFS needs. These strategies will vary by region,
community, social class, ethnic group, household, gender, and season. Thus,
their use as indicators is location-specific (see below). Outcome indicators can
be grouped into direct and indirect indicators. Direct indicators of food
consumption include those that are closest to actual food consumption rather
than marketing channel information or medical status (e.g., household
consumption surveys). Indirect indicators are generally used when direct
indicators are either unavailable or too costly (in terms of time and money) to
collect (e.g., storage estimates, nutritional status assessments).

Journalfor Farming Systems Research-Extension


As indicators that reflect food access, the generalized patterns of coping
strategies find practical application as tools for local food security monitoring
(Frankenberger and Goldstein, 1991). Building upon the work of the World
Food Program (WFP), there are three types of indicators that can be
monitored for changing coping responses, thus suggesting worsening condi-
tions and heightened food insecurity. Leading indicators (WFP refers to these
as early indicators) are changes in condition and responses prior to the onset
of decreased food access. Examples of such indicators are: 1) crop failures (due
to inadequate rainfall, poor access to seed and other inputs, pest damage, etc.);
2) sudden deterioration of rangeland conditions or conditions of livestock
(e.g., unusual migration movements, unusual number of animal deaths, large
numbers of young female animals being offered for sale); 3) significant
deterioration in local economic conditions (e.g., increases in the price ofgrain,
unseasonal disappearance of essential food stuffs, increases in unemployment
among laborers and artisans, unusual low levels of household foodstocks); and
4) significant accumulation of livestock by some households (due to depressed
prices caused by oversupply). Leading indicators can provide signs of an
impending problem and may call for a detailed situational analysis to deter-
mine the extent of the problem, causes, and need for monitoring. These
indicators are a combination of process indicators dealing with both availabil-
ity and access vulnerability (Frankenberger, 1992).
Concurrent indicators (WFP calls these stress indicators) occur simulta-
neously with decreased access to food. Examples of such indicators are: 1)
larger than normal numbers of able-bodied family members in search for food
or work; 2) appearance in the market of unusual amounts of personnel and
capital goods, such as jewelry, farm implements, livestock (draft animals); 3)
unusual increases in land sales or mortgages; 4) increases in the amount of
people seeking credit; 5) increased dependence on wild foods; 6) reduction in
the number of meals; and 7) increased reliance on interhousehold exchanges.
Concurrent indicators can be assessed while carrying out a situational analysis
using rapid rural appraisals. These indicators are primarily access/entitlement
related. Once the nature and extent of the problems have been confirmed,
interventions can be introduced that focus on the causes or mitigate the
Trailing indicators (WFP calls these late outcome indicators) occur after
food access has declined. They reflect the extent to which the well being of
particular households and communities have been affected. In addition to
signs of malnutrition and high rates of morbidity and mortality, trailing
indicators include increased land degradation, land sales, consumption ofseed
stocks and permanent outmigration. All of these indicators are signs that the
household has failed to cope with the food crisis (Frankenberger and Gold-
stein 1991).
An understanding of farmer coping strategies can be essential in guiding the
design and implementation of interventions to increase HFS. As Figure 3

Vol. 4, No. 1, 1993


illustrates, the types of coping strategies employed by households not only
indicate household vulnerability to food shortage, but also correspond to
different types of government and donor responses. Household coping
strategies that do not involve divestment normally indicate moderate vulner-
ability, and government/donor response is more appropriately oriented
toward longer-term development efforts. Such responses can be targeted to
enhance the long-term sustainability of HFS, especially in those areas where
vulnerability is likely to increase. In regions where divestment is beginning to
occur, household vulnerability becomes high and mitigation should be
considered the appropriate response. Mitigative interventions are those that:
1) abate the impacts of the current emergency while reducing vulnerability to
future emergencies; 2) target the conservation of productive assets at the
household level; and 3) reinforce and build upon existing coping patterns
(Hutchinson, 1991). In areas where productive asset sales and permanent
outmigration have begun to occur, the population is extremely vulnerable to
famine. Such indices would call for immediate relief action on the part of the
government and donors. Thus, an appropriately designed HFS monitoring
system could be flexible enough to serve all three purposes. Presently most
Early Warning Systems operating in Africa are only used for food aid planning
(i.e., the relief function).
Given their usefulness in identifying vulnerable households, it is important
to also recognize the limitations of these food access indicators. First,
socioeconomic variables mean different things in different contexts (Borton
and York, 1987). Researchers and development practitioners should under-
stand the locational specificity of socioeconomic variables so that they are not
misinterpreted. Second, the raw data used as indicators can be misleading.
Hesse (1987) demonstrated that regional livestock market data from Mali
could easily be misunderstood because individuals were buying and selling the
same stock repeatedly in the same day. Thus, the quality of the data needs to
be properly validated before being incorporated into a monitoring system.
Third, without a baseline for determining what is "normal" behavior for a
given population, it is difficult to make valid interpretations of trends
displayed by indicators (Borton and York, 1987). Fourth, given the locational
specificity of socioeconomic indicators, it is difficult to make comparisons
across, or to aggregate the data. This remains one of the critical areas of
research to be addressed. Because of these limitations, numerous challenges
lay ahead for those HFS monitoring systems that incorporate socioeconomic
data (Haddad et al., 1991).
To minimize inaccuracies derived from the use of socioeconomic indica-
tors, multiple indicators should be used whenever possible. The convergence
of evidence will instill confidence in those agencies responsible for addressing
food security problems. In addition, attempts should be made to pretest
indicators to determine whether local factors may distort an indicator's
validity and reliability (Haddad et al., 1991). Efforts also should be made to
limit food access indicators to a manageable number.
Journal for Farming Systems Research-Extension


Indicator Availability Sources of Information Measurement Level of Limitation
and Collection Method Aggregation

Food Supply Indicators

Meteorological Data

Information on Natural Resources
(includes grazing resources)

Agricultural Production Data
(crops and animals)

Agroecological Models

Food Balance Sheets

Information on Pest Damage

Market Information (prices)

Regional Conflict

readily government reports
available monitoring stations
satellite remote


periodic assessments
government, NGOs
satellite imagery
government and donor studies

readily government reports
available crop cutting on sample plots
remote sensing
farmer reports

not readily monitoring stations
available soil assessments

readily secondary sources
available government reports

moderately field assessments
available government reports

readily price data
available market surveys

not readily key informants
available NGOs

cumulative amount/average
change from average

dekedal values
dekadal value/previous dekadal
dekadal average/long-term
dekadal average

seasonal kg/capita
departure from average kg/capita
% change from past years

FAO Crop Specific Soil
Water Balance Model

requirements (opening stocks, production,
imports, domestic per capital
requirements, exports and closing stocks)

seasonal kg/capita for crops
% of change from last year

value of crop prices, livestock prices
monthly value/average
monthly value for previous year

# of incidents
influx of refugees

national number of stations
regional timing of rains may be
district false indicator






access to remote sensing

limited information
on other crops
besides staples

capability for analysis

nontraded crops

frequency of assessment

interpretation of
sales and price

regional collection of data
local in conflict zone



Availability Sources of Information
and Collection Method


Level of Limitation

Food Access (Effective Demand or Entitlement)

Risk Minimizing Strategies
land use practices limited

diversification of livestock limited

Loss Management Strategies
dietary change (both quantitative
and qualitative)

change of food source

diversification of income sources

access to loans/credit

livestock sales

seasonal migration

sale of production assets

distress migration




formal surveys
formal surveys

HH surveys
in-depth interviews
HH surveys

HH surveys

limited RRA
HH surveys
available market surveys
secondary data

limited RRA
HH surveys
limited RRA
HH surveys

limited RRA
HH surveys
government records

changes in crop mix
changes in time of planting
changes in livestock mix
early movement to alternative range
# animal deaths

reduction in # ofmeals
decreased dietary diversity
shifts from preferred to lower status food
increased dependence on wild foods
# of HH dependent on reserves
grain price increases
changes in petty marketing patterns
changes in wage rates
increase # of HH seeking off-farm assistance from relatives
assistance from relatives
increase # of people seeking assistance from relatives
# of people seeking credit
increase sale of livestock/season
decline of livestock prices

large # of people migrating for work

appearance in market of unusual amounts of personal and
capital goods (jewelry, farm implements, draft animals)
sale of young female animals
# of whole families moving out of area

HH/village location specific

HH/village location specific

HH/village location specific

HH/village location specific

HH/village location specific




location specific

location specific

location specific

location specific

location specific



'5 3

% \
r. SN







Indicator Availabilit

Outcome Indicators

Direct Indicators

household budget limited
and consumption surveys

j household perception limited
P of food insecurity

n food frequency assessments limited

Indirect Indicators
storage estimates limited
subsistence potential ratio readily

household food security card limited

nutritional status assessments readily

y Sources of Information
and Collection Method

national surveys

in-depth interviews

HH surveys
24-hr recall

HI surveys

HH surveys

HIH surveys

government health
formal surveys


Level of Limitation

price per unit of food or caloric per unit of food
conversion factors/capita

# of months of self provisioning from household
production and receipt of in-kind as perceived
by the household
# of meals per day
# and types of ingredients in meals
# of times per day a nutrient-poor
gruel was served as main meal

# of months food stores will last as perceived by the IIH

size of farm, expected yield and age and sex
composition of household
Amount of food produced/food required

food available from main crop compared to
HH requirements on monthly basis

arm circumference

national high cost
district z
village local population
1-HH may distort data O

village difficult to aggregate at
ItH regional/national level
limited level of precision, -
culturally specific

village difficult to obtain due to
HH cultural beliefs
difficult to aggregate
village difficult to aggregate
HH assumes all farm
land used for
food production
village only useful in areas
HH where most food is
household grown
national nutritional status
regional influenced by sanitary
local conditions, care
age assessment question


To date, few information systems are presently in place that adequately
incorporate both supply/production data and access/entitlement data in the
same indicator set. Early Warning Systems in most countries have had a food
supply orientation focusing on production data and nutritional status prima-
rily because these data arc easiest to obtain and are well suited to aggregate
analysis (Buchanan-Smith ct al., 1991). Few governments or donors are
willing to commit the time or resources necessary to obtain information on
socioeconomic indicators that are sensitive to the vulnerability of different
groups. Decentralized HFS monitoring systems would be the best means of
obtaining such information. Centralized HFS monitoring systems are likely to
experience more difficulties in adequately assessing the HFS status of local
populations (Frankenberger, 1992).
The information needs of different user groups will influence the selection
of HFS indicators and data collection methods to be used. National govern-
ments and donors require quantitative information in a centralized system to
help make informed planning and policy decisions regarding the sharing of
limited resources across regions. Local governments, farming systems re-
search teams, NGOs, and local communities require qualitative location-
specific information in a decentralized system to design appropriate
interventions. A balance must be struck between the need for data for
centralized decisions on the allocation of resources and a need for information
appropriate for decentralized HFS monitoring and interventions.
HFS information systems can be designed to take both of these concerns
into account. Using a staged process, vulnerability to food insecurity can be
mapped for a country or region to assist national governments and donors in
making decisions regarding the allocation of resources across regions. This
information can then be used to determine where more location-specific HFS
information is necessary to collect by farming systems teams. This information
could then be used for targeting development initiatives and for setting up
decentralized HFS monitoring systems.

Vulnerability Mapping
Pioneering efforts in vulnerability/risk mapping have been carried out in
Bangladesh and Sudan under WFP support (Borton and Shoham, 1991). The
USAID-funded Famine Early Warning Systems Project also has contributed
significantly to this conceptual development (Downing, 1990). Vulnerability
maps are maps which identify the areas and sectors of the population that are
most vulnerable to food insecurity. These maps highlight the regions that
need to be monitored more closely, and identify factors to take into consid-
eration in designing interventions for vulnerable areas (Borton and Shoham,
1991). An earlier version of vulnerability mapping used in the 1970s was
"functional classification" of undernourished populations as a basis for food
and nutrition planning (Joy, 1973).

Journalfor Farming Systems Research-Extension


Crop & Livestock Adjustments
: Diet change
SFamine food use'
Grain loan from kin
Laor sales migrationn)

Small animal sales

Cash/cereal loan
S\from merchants

I E Productive asset sales

Farmland pledging

Farmland sale

"3 ', Outmigration

Time ,

-- Relief




Adaptation Divestment ;

Diet change, borrowing, Liquid Productive'
seasonal labor migration assets assets

Figure 3. Responses to household food shortages (after Watts, 1983).
The types of coping strategies employed by households indicate household
vulnerability to food shortage, and correspond to different types of government
and donor responses. Office of Arid Lands Studies, The University of

Vol. 4, No. 1, 1993


Vulnerability to food insecurity is an aggregate measure for a given
population of the risk of exposure to different types of shocks (e.g., drought)
or disaster events (primarily supply indicators) and the ability to cope with
these events (primarily access/entitlement indicators) (see Matrix 2). Map-
ping vulnerability involves assessing the baseline vulnerability (the contextual
factors encompassing food insecurity events over the previous years), current
vulnerability (the shocks overlying the baseline), and future vulnerability
(trends associated with long-term food security risks (Frankenberger, 1992)).
A number of different approaches have been used in mapping food-related
vulnerability (Frankenberger, 1992). These include: (1) desegregating exist-
ing data on socioeconomic groups; (2) surveys that collect information
directly relevant to vulnerability; (3) using existing data on key indicators of
vulnerability; and (4) conducting rapid rural appraisals (Borton and Shoham,
1991). Combining approaches may be necessary due to quality differences in
the data. Geographic Information Systems are now being used for combining
different data sets (Hutchinson et al., 1992).
The types of information that can be used as indicators of vulnerability to
food insecurity will vary considerably between countries and regions within a
country. Some indicators may be more important than others in determining
vulnerability, so subjective weighting of indicators is often necessary (Borton
and Shoham, 1991). If weighting must be done, it is important to rely on
individuals who have local knowledge and experience in the areas to assign
these weights.
Vulnerability maps drawn up for arid/or semi-arid regions should take into
account the location of ecologically favorable areas that serve as refuge points
during drought conditions (S. Davies, Personal Communication). The over-
utilization of the resources in such areas by multiple users during times ofstress
can increase the vulnerability of the local population. Monitoring posts or
sentinel sites (Mason et al., 1984) could be established in these areas of
convergence to assess the regional impact of droughts.
Vulnerability maps have great potential for national governments and
donors in assisting with decisions regarding the allocation of resources across
regions. The development of such maps could ideally be a first step in
identifying districts or subregions where it is necessary to collect more
locational-specific HFS information for designing appropriate interventions.
Decentralized HFS monitoring systems could then be developed in these
designated areas.

A Systematic Approach to Identifying Food Insecure Households
In countries where national early warning systems already exist (e.g., crop
forecasting, food balance sheets, nutrition surveillance), information supplied
by these systems can help develop vulnerability maps for various regions.
These vulnerability maps should be based as much as possible upon both

Journal for Farming Systems Research-Extension


supply-type indicators and access/entitlement indicators to avoid designating
an area as vulnerable which may not be. Farming systems teams would not
necessarily be responsible for creating these maps, but would use them to help
target future activities. These maps would be fine-tuned as more information
becomes available from the farming systems teams and other sources.
The vulnerability maps can then be used for designating areas where more
location-specific HFS information can be gathered (Frankenberger, 1992). If
such information does not already exist, farming systems teams can conduct
RRAs to understand the local socioeconomic context and identify HFS
constraints and key indicators to be used in decentralized food security
monitoring systems.
The information gathered by the farming systems teams could feed directly
into the development ofa district or sub-regional contingency plan, consisting
of the HFS monitoring system and a set of predetermined responses that
would be implemented if and when food security conditions change. These
responses would be designed in noncrisis years, and would encompass
development-type interventions that enhance the long-term sustainability of
HFS, mitigation-type interventions that enable households to retain their
productive assets and existing entitlements, and relief-type responses if
immediate food aid distribution is warranted. Responsibilities for these
various actions will be negotiated and assigned to government agencies,
donors, and local NGOs prior to the onset of the food crisis to improve
response timing (Frankenberger, 1992). Whenever possible, participation of
local communities in information gathering and response should be encour-

Integrating HFS Considerations into Ongoing FSRE Projects and
FSRE development projects and programs should be designed in such a
way to take into account periodic shocks that may negatively impact the food
security situation of households. To prevent households from selling off their
assets and diminishing their ability to take advantage of project/program
inputs, project designs should incorporate: (1) a monitoring system with
indicators that can detect changes in entitlement and food supply; and (2)
contingency plans that protect the asset base of the project beneficiaries
during periods of stress through income transfers such as food-for-work/
cash-for-work. FSRE teams would not be responsible for implementing these
programs, but could help in designing such interventions for the area in which
they are working. Through local community participation, these contingency
plans can be designed to focus on improvements in infrastructure and/or
natural resource management that will enhance the long-term food security
of the local area.
To effectively integrate HFS concerns into farming systems research and
extension activities, it is important to first understand the nature of the client

Vol. 4, No. 1, 1993


group, and then to elicit their participation in problem identification and
technology development. The next section will address these issues.

The major clientele of FSRE are low-resource farmers. Their small-scale
rainfed farming systems tend to be more internally complex in comparison to
industrial or green revolution systems, and are often more dynamic in
exploiting unpredictable conditions. As Chambers (1991) points out, poor
people in these areas seek to multiply their enterprises to raise their income and
reduce risk. This diversity is the key component in the sustainability of their
HFS and livelihoods. Thus, many low-resource farming systems are moving
in the opposite direction to that of industrial or green revolution agriculture.
Instead of becoming more simple and uniform, they are becoming more
complex and diverse. Rather than intensifying external inputs, intensification
is more internal.
Because of this complexity and diversity, many on-station researchers do
not understand these systems well. Approaches for developing widely appli-
cable technologies for relatively simple systems in uniform environments no
longer are appropriate (Merrill-Sands et al., 1991). New approaches are
needed for identifying multiple products that can be tailored to the identified
needs of diverse clientele and production systems. In addition, the dynamic
nature of these systems requires that diagnostic updates are regularly carried
out to ensure that farmer problems and needs are taken into account.
Farming systems research has done a good job in eliciting farmer partici-
pation in the identification of the various strategies pursued to meet food
security needs. However, low-resource farmer participation beyond the
diagnostic phase has been limited. Such participation is critical in the search
for sustainable HFS interventions that are locally acceptable. This has much
to do with the fact that transfer of technology approaches are still the
dominant paradigm in most agricultural research systems (Chambers et al.,
1989). Many have argued that farming systems research or on-farm client-
oriented research (OFCOR) activities have enabled formal R&D systems to
extend to the farm. That is, on-station researcher-managed experiments are
now conducted on farmers' fields. Participatory approaches in technology
generation are limited because technology generation is still considered the
domain of the biological scientists (Knipschcer, 1989).
To enhance the active involvement of farmers in the technology develop-
ment process, participatory methodological innovations were derived under
a number of labels such as farmer participatory research (Farrington and
Martin, 1987) and farmer first (Chambers et al., 1989). Farmer-first models
call for methodological reversals in agricultural research (Rhoades, 1989;
Baker, 1991). This involves a shift away from a technology-supply orientation

Journal for Farming Systems Research-Extension

Matrix 2

Risk of an Event Ability to Cope

Shocks/Trends IlH Characteristics Access to Resources Production/Income Support Structures

Baseline Vulnerability
Crop Production andLivestockRisks composition access to land crop/livestock community support
drought episodes (age dependency access to labor production mechanisms (claims)
soil conditions ratio) liquid assets other income sources NGOs
pest infestations education productive assets seasonal migration government policies
Market Risks health status credit access to social
market infrastructure out migration common property services
price fluctuations (assets, food, resources (for wild
cash crops, livestock) foods and other
food shortages products)
access to employment food stores
Political Risks

S Current Vulnerability
Crop Production andLivestock Risks
current drought
pest attack
Market Risks
market infrastructure
price fluctuations (assets, food,
cash crops, livestock)
food shortages
access to employment
Political Risks

(age dependency
health status

access to land
access to labor
liquid assets
productive assets
common property
resources (wild
foods and other
food stores

other income sources
seasonal migration

community support
mechanisms (claims)
government policies
access to social

Future Vulnerability (trends)
Land Pressure
Out Migration


land tenure


support structure
changes ua



Baker, 1991). This involves a shift away from a technology-supply orientation
and hypothesis-deduction model to an emphasis on indigenous farmer
knowledge, innovative behavior, and farmer experimentation. Many advo-
cates of this model feel that farmer knowledge, inventiveness, and experimen-
tation have long been undervalued, and that farmers and scientists should be
partners in the research and extension process (Rhoades, 1989). Research
should be based on the problem analysis and priorities of farmers, with farmers
being the central experimenters (Chambers et al., 1989).
According to the farmer-first approach, farmers participate in the technol-
ogy development process in three ways. First, farmers are involved in the
diagnostic phase. The role of the researcher is to elicit, encourage, facilitate,
and promote the analysis by farmers. Second, farmers are provided a range of
choices of possible solutions to identified problems which they pick and
choose to suit their conditions and enhance their adaptability. This translates
into early involvement of farmers in the technology design process, especially
in screening alternative solutions. Third, farmers must be encouraged to
actively participate in experiments for site-specific technologies and adapta-
tion. This may entail improving the farmers' own capacity to carry out on-farm
trials and experimentation.
Although farmer-first approaches have provided excellent suggestions for
ways to improve farmer participation in research aimed at improving HFS,
caution must be exercised in adopting all of its recommendations in a
wholesale manner (Baker, 1991). First, farmer articulated demands nearly
always relate to short-term priorities. An exclusive focus on these priorities can
lead to an underinvestment in sustainable HFS options. Second, there are
differences in household priorities corresponding to gender roles, wealth, and
village location. Whose priorities and interests should be taken into account?
Third, quantitative measures are often needed to convince policy makers and
extension services of the value of technology options. Thus, some experimen-
tal rigor may be necessary. Fourth, farmer-first reversals in the technology
design process will be met with much resistance in national agricultural
research systems that primarily use a transfer of technology model. Such
differences in objectives and methods could reduce institutional acceptance
and researcher collaboration with on-farm research teams. Compromises may
have to be sought to gain acceptance of such new participatory approaches.

The previous sections have discussed a number of HFS issues that should be
considered in the implementation of FSRE programs and projects. This
section outlines a number of ways to incorporate these issues into ongoing
research and development activities.

Journal for Farming Systems Research-Extension


To improve the targeting of food-insecure households within a research
area, FSRE teams can work with other government agencies, donors, and
NGOs to derive vulnerability maps. The development of vulnerability maps
could be the first step in identifying districts or sub-regions where it is
necessary to collect more location-specific HFS information and target
interventions. Early Warning Systems already exist in most southern African
countries, so information supplied by these systems can help map vulnerability
for various regions. For example, in Zambia the Central Statistics Office in
collaboration with FAO and UNICEF is setting up a drought impact
monitoring system for the country. In Malawi, national household food
security surveys have been regularly carried out by the Food Nutrition and
Monitoring Unit (Ministry of Agriculture) with UNICEF support. Existing
data from these systems can be used to formulate maps to cut down on costs.
Once the vulnerability maps have been developed, areas can be targeted
where more location-specific HFS information can be gathered. Two diagnos-
tic tools that have been used extensively because of their timeliness and cost-
effectiveness are rapid rural appraisals (RRA) and participatory rural appraisals
(PRA). Although these techniques are related, they are not the same and
should not be considered interchangeable.

Rapid Rural Appraisals
RRAs have been employed in food security monitoring as a way to provide
a systematic overview of the diet and strategies for acquiring food in the target
area while using a minimum amount of survey time and resources (Franken-
berger, 1990). They can be effectively used in carrying out preharvest surveys
and food systems inquiries in the initial stages of setting up an information
system (Davies et al., 1991). Such surveys have helped identify the critical
regional food resources that need to be sustained and managed (Velarde,
1991). These surveys can also help identify food-insecure groups in detail in
order to plan food security interventions (Maxwell, 1989).
RRAs rely on multidisciplinary teams to carry out surveys and often use a
topical outline derived from secondary sources and past surveys to help guide
the interviews. Data collected may include information on food preferences,
food marketing and purchases, meals and food preparation, food storage and
depletion of food stocks, free food distribution, wild foods, seasonality of food
supply, food prices, health related food issues such as weaning practices,
specialty foods for child-bearing women, prevalent health problems, women's
involvement in agriculture and other production activities, and consumption
constraints facing households (Frankenberger 1990). Information generated
from RRAs provides insights into interregional differences in consumption
patterns rather than intravillage differences.

Vol. 4, No. 1, 1993


Despite the multiple advantages derived from RRAs, it is important to
recognize the limitations of such approaches. Researchers cannot be certain
that households interviewed in the survey are representative of most house-
holds in the region. Time constraints usually do not allow for systematic
sampling procedures to be followed. Thus, RRA techniques should be viewed
as complementary to other research methodologies such as formal surveys and
in-depth anthropological studies. RRAs can even be combined with the
formal interview process to correct biases. For example, random sampling
procedures could be introduced halfway through field visits once hypotheses
have been identified (Molnar, 1991).
Given time constraints, RRAs may also have trouble targeting the least
visible food-insecure target groups such as landless, rural poor, women, and
isolated ethnic groups. To compensate for this, RRA teams can focus on
degraded resource areas and smaller marginal farms while interviewing
households (Molnar 1991).
The major intention of RRAs, from our perspective, is to allow to
understand the diversity of farming systems and corresponding constraints
that are distributed within a given target area. Once this diversity and
complexity are understood, specific villages can be selected which are repre-
sentative of a wider array of villages so that further diagnoses can be carried
out. It is at this point that participatory rural appraisals should be conducted.

Participatory Rural Appraisal
PRAs also involve multidisciplinary teams that gather HFS information in
a systematic, yet semi-structured way; however, they tend to focus on the
village rather than the region, and community participation is considerably
more active (WRI, 1989). PRA is intended to help communities mobilize
their human and natural resources to define problems, consider successes,
evaluate local capacities, prioritize opportunities, prepare a systematic and
site-specific plan of action, and create a means for facilitating community self-
help initiatives. It brings together the development needs as defined by the
community with the resources and technical skills offered by the Government,
donor agencies, and NGOs.
A number of techniques are used in PRA to elicit farmer involvement in
identifying HFS problems and deriving possible solutions. One method
involves open-ended group discussions used to enable farmers to analyze
problems, identify research opportunities, and prioritize interventions. Such
discussions are different from many of the group discussions carried out in
RRAs because not only do they generate information, but they also allow
farmers to synthesize information and draw conclusions (Ashby, 1991). Thus,
the group's own understanding of the HFS problems is advanced.
Diagrams have also been used effectively to stimulate questions and
responses, allowing the farmers' knowledge to be made more explicit (Con-

Journal for Farming Systems Research-Extension


way, 1989). Diagrams can simplify complex information, making it easier to
communicate and analyze. Five different diagrams derived from agroecosys-
tem analysis are often used. Maps are used to identify different parts of the farm
or village and its relation to basic resources and land forms. Transects tend to
be drawn by researchers who walk from the highest point to the lowest point
in the immediate environment accompanied by the local people. Consulting
people in each zone, transects can help identify major HFS problems and
opportunities in the agroecosystem and where they are located. Calendars are
used to indicate seasonal features and changes and are useful for allowing
farmers to identify critical times in the crop production cycle with regard to
changes in climate, cropping patterns, labor access, food procurement strat-
egies, diet, and prices. The Adaptive Research Planning teams in Zambia have
developed a number of such calendars for their respective research areas
(Velarde, 1991). Flow diagrams are used to present events in a cycle of
production, marketing, and consumption. Venn diagrams can be used to
understand the institutional relationships within a village. Such information
could be critical to understanding the informal social mechanisms (e.g.,
claims) that buffer households from periodic shocks.
Through the use of RRA and PRA techniques, FSRE researchers have
begun to appreciate the complexity of the coping strategies and livelihood
systems that farmers pursue. This has led many researchers to expand the
enterprise coverage from a predominantly crop focus to a broader crop-
livestock off-farm mix. Linkages between systems at the field, farm, village,
region, and wider political economy have been identified (Zandstra, 1991).
Diagnostic techniques have also improved our understanding of the local
classification systems for plants, soils, types of land, crops, and wild plants,
facilitating better researcher-farmer communication on HFS issues. Anthro-
pologists and sociologists have played a vital role in fostering this improved
interaction with farmers.
However, it is the excessive concern of social scientists with diagnosis that
has contributed to the limited participation of farmers in other phases of the
research process. As Ashby rightly points out (1989), diagnostic research has
become a hothouse of methodology development spawning sondeo teams,
informal surveys, rapid appraisals, key informant surveys, etc. The farmer has
become an object of investigation just as plants, soils, insects, and viruses are
objects of study to be measured. Asking farmers questions has become an
industry. Thus, to involve farmers in other phases of the research process,
which is critical to HFS improvements, we must first involve the social

A major assumption that many researchers make is that farmers will adopt
the technologies that are generated by the research system once they have

Vol. 4, No. 1, 1993


been tested and meet certain production criteria. However, farmers often take
these recommendations and adapt them to suit their own resources and
purposes (Ashby, 1991). This is why it makes sense to involve farmers early on
in the testing phase so that technology aimed at improving HFS can be
adapted to their circumstances. This could speed up the technology develop-
ment process and reduce unnecessary costs for technologies that are inappro-
Collaborative farmer participation in technology design and testing is more
likely to occur when researchers are willing to allow farmers to contribute to
the conceptualization of an experimental program. For example, farmers can
be brought in at early stages to help researchers select varieties for on-farm
testing (Ashby, 1991). Such an approach was used in Uttar Pradesh, India for
screening improved rainfed rice varieties (Maurya et al., 1988). Such a menu
approach was also used in Rwanda, where farmers were invited into the
experiment station to participate in the seasonal selection ofpotatoes (Haugerud
and Collinson, 1990). During this selection process, researchers found that
they had ignored many important features of interest to farmers. HFS criteria
should be incorporated into this screening process.
When choosing participants for early screening of technologies, care must
be taken to include those members of the household with the most expertise
for the given crop or livestock species. This is especially true for interventions
that directly impact HFS. For example, since women are primarily involved
with bean production in Rwanda, they were asked to take part in evaluating
hundreds of advanced breeding lines in the national bean research program
(Ashby, 1991).
One of the major obstacles to involving farmers more effectively in the
design and early testing phases is that researchers are afraid that farmers will
lose confidence in the research system inappropriate recommendations cannot
be given. Ashby (1991) points out that farmers need to be given more
responsibility for technology design, testing and adaptation, especially if
research budgets are severely limited. In addition, farmers and private traders
can be more involved with seed multiplication in order to increase the menu
ofoptions made available to farmers. This will help cut down on the cost borne
by the research station in multiplying seed.

On-farm Experimentation
The extent of farmer experimentation has often been underperceived by
most researchers. Much more could be done to strengthen farmers' informal
R&D systems. First, social scientists could be more actively involved in
identifying the various topics upon which farmers are presently conducting
their own experiments. Revelation of these experiments will help researchers
understand which topics and HFS Problems farmers arc most interested in and
help to solve these problems.

Journal for Farming Systems Research-Extension


Second, researchers and extensionists can help improve farmers own
capacity to carry out experiments. Nonformal education and training could be
provided to farmers to enable them to understand and implement controlled
comparisons, replications, and random assignments (Ashby, 1991; Bunch,
1989). Such approaches have been successfully adopted in Colombia, Guate-
mala, and the Gambia (Ashby, 1991). In all these cases, results have shown
that farmers with primary schooling can master the major principles of
The goals of such participatory approaches is to encourage a process where
people develop their own agriculture and solve the HFS problems in a self-
sustaining way (Bunch, 1989). We must get away from the idea of providing
packages to farmers, and allow them to choose from a menu of options that
fit more appropriately with the diversity of strategies they are pursuing to meet
their food needs. Highly structured on-farm trials limit farmers' ability to
experiment with and manipulate the new genetic material (Sumberg and
Okole, 1989). It also precludes adjustments in other production practices or
exploitation of production niches which could make the new variety more

It is apparent that much intellectual progress has been made in our under-
standing of the processes that lead to food-insecurity situations for house-
holds. We have moved away from simplistic notions of food supply being the
only cause of household food insecurity to assessing vulnerability of particular
groups in terms of their access to food. For this reason, information should be
collected on factors that play a role in limiting food availability and the options
that households have for food access. The procurement strategies and social
mechanisms used by households to obtain stable access to food are diverse and
complex. FSRE diagnostic studies should attempt to understand this diversity
by'documenting the production-consumption linkages and household cop-
ing strategies that characterize an area. With such information, HFS indicators
can be identified that are location specific and can be monitored in ongoing
development activities.
To aid in targeting FSRE activities aimed at HFS problems, vulnerability
maps can be drawn up for a country or region to identify the areas and sectors
of the population that are most vulnerable to food insecurity. The develop-
ment of such maps would be the first step in identifying districts or subregions
where it is necessary to collect more location-specific HFS information in
order to design appropriate interventions.
Once a food-insecure area has been designated, diagnostic tools such as
RRAs and PRAs can be used to understand the local socioeconomic context
and identify HFS constraints and key indicators to be used in decentralized

Vol. 4, No. 1, 1993


food security monitoring systems. Contingency plans would be drawn up to
link information to response. Responses would encompass development-type
interventions that enhance the long-term sustainability of HFS, mitigation-
type interventions that enable households to retain their productive assets,
and relief-type responses if immediate food and distribution is warranted. The
FSRE teams would not be responsible for all of these responses, but would
coordinate such activities with other government agencies, donors, and local
To ensure that HFS interventions are appropriate, FSRE teams should
encourage farmers to participate in all phases of the research process. Farmers
participation in diagnosis, design, and experimentation will enable them to
deal with their HFS problems in a self-sustaining way.


Ashby, J.A. 1991. Adopters and adapters: the participation of farmers in on-farm
research. Pages 273-286 in R. Tripp, ed., Planned change informing systems: Progress
in on-farm research. Chichester, UK: John Wiley and Sons.
Ashby, J.A. 1989. The effects of different types of farmer participation on the manage-
ment of on-farm trails. Agricultural Administration and Extension 25:235-252.
Baker, D. 1991. Reorientation, not reversal: African farmer-based experimentation.
Journal for Farming Systems Research-Extension 2(1):135-147.
Borton, J., and J. Shoham. 1991. Mapping vulnerability to food insecurity: Tentative
guidelines for WFP offices. World Food Program, Relief and Development Institute,
London. mimeo.
Borton, J., and S. York. 1987. Experiences of the collection and use of micro-level data
in disaster preparedness and managing emergency operations. Report on January
workshop held at the London School of Hygiene and Tropical Medicine, Relief and
Development Institute, London.
Bunch, R. 1989. Encouraging farmer's experiments. Pages 55-59 in R. Chambers, ed.,
Farmer first: Farmer innovation and agricultural research. London: Intermediate
Technology Publications.
Chambers, R. 1991. Complexity, diversity and competence: Toward sustainable liveli-
hoods from farming systems in the 21st century. Journal of Asian Farming Systems
Association 1(1):79-89.
Chambers, R. 1989. Editorial introduction: Vulnerability, coping and policy. IDS
Bulletin 2(2):1-7.
Chambers, R., A. Pacey, and L.A. Thrupp, eds. 1989. Farmer first: Farmer innovation
and agricultural research. London: Intermediate Technology Publications.
Conway, G. 1989. Diagrams for farmers. Pages 77-86 in R. Chambers, A. Pacey, and L.A.
Thrupp, eds., Farmer first: Farmer innovation and agricultural research. London:
Intermediate Technology Publications.
Corbett, J. 1988. Famine and household coping strategies. World Development 16(9):1009.
Davics, S., M. Buchanan-Smith, and R. Lambert. 1991a. Early Warning in theSahel and
Hi. rn ofAfrica: The state of the art: A review of the literature. IDS Research Reports
No. 20. Institute of Development Studies, University of Sussex, Brighton, UK.

Journalfor Farming Systems Research-Extension


Davies, S., M. Leach, and R. Davies. 1991b. Food security and the environment: Conflict
orcomplementarity. IDS Discussion Paper. Institute for Development Studies,University
of Sussex, Brighton, UK.
DcWalt, K. 1983. Income and dietary adequacy in an agricultural community. Social
Science and Medicine 17(23):1877.
Dewey, K 1981. Nutritional consequences of the transformation from subsistence to
commercial agriculture in Tabasco, Mexico. Human Ecology9(2):151.
Downing, T.E. 1988. Climatic variability, food security and smallholder agriculturalists
in six districts of central and eastern Kenya. Unpublished PhD. dissertation. Clark
Downing, T.E. 1990. Assessing socioeconomic vulnerability to famine: Frameworks,
concepts, and applications. FEWS Working Paper 21. USAID, Washington, DC.
Farrington, J.,and A.M. Martin. 1987. Farmer participatory research: A review ofconcepts
and practices. Agricultural Administration (Research and Extension) Network. Dis-
cussion Paper No. 19. Overseas Development Institute, London.
Fleuret, P., and Fleuret. 1980. Nutrition, consumption and agricultural change. Human
Organization 39(3):250.
Frankenberger, T.R. 1985. Adding a food consumption perspective to farming systems
research. Report prepared for USDA, Office of International Cooperation and
Development, Nutrition Economics Group, Washington, DC.
Frankenberger, T.R. 1990. Production-consumption linkages and coping strategies at
the household level. Paper presented at the Agriculture-Nutrition Linkage Workshop,
Bureau of Science and Technology, USAID, Washington, DC.
Frankenberger, T.R 1992. Indicators and data collection methods for assessing household
food security. University of Arizona, Office of Arid Lands Studies, Tucson, USA.
Frankenberger, T.R., and D.M. Goldstein. 1990. Food security, coping strategies, and
environmental degradation. Arid Lands Newsletter 30:21-27.
Frankenberger, T.R, and D.M. Goldstein. 1992. Linking household food security with
environmental sustainability through an Analysis of coping strategies. In K. Smith,
ed.,Growing Our Future. New York: Kumarian Press.
Green, R.H. 1992. Southern Africa: That the people may be fed. Paper presented at the
Drought and Famine in southern Africa Conference, Institute of Development
Studies, University of Sussex, Brighton, UK.
Haddad, L., J. Sullivan, and E. Kennedy. 1991. Identification and evaluation of
alternative indicators of food and nutrition security: Some conceptual issues and an
analysis of extant data. Washington, DC: International Food Policy Research Insti-
Haugerud, A., and M.P. Collinson. 1990. Plants, genes and people: Improving the
relevance of plant breeding in Africa. Experimental Agriculture 26:341-362.
Hernandez, M., et al. 1974. Effect of economic growth on nutrition in a tropical
community. Ecology of Educational Nutrition 3:283.
Hesse, C. 1987. Livestock marketdata as an early warning indicator ofstress in the pastoral
economy. Pastoral Development Network, Discussion Paper No. 24f. Overseas Devel-
opment Institute, London.
Hutchinson, C.F. 1991. Famine and mitigation. In Famine mitigation: Proceedings of
workshops held in Tucson, Arizona, May 20-May 23, 1991 and Berkeley Springs, West
Virginia, July 31 -August 2, 1991, US Office of Foreign Disaster Assistance, Washing-
ton, DC and Office of Arid Lands Studies, Tucson, Arizona, USA.

Vol. 4, No. 1, 1993


Hutchinson, C.F., P.T. Gilruth, R.A. Hay, S.E. Marsh, and C.T. Lee. 1992. Geographic
information systems applications in crop assessment and famine early warning. Univer-
sity ofArizona, Arizona Remote Sensing Center and Advanced Resource Technology
Program, Tucson, Arizona, USA.
Jodha,N.S. 1986. Poor in dry regions ofIndia. EconomicandPolitical Weekly21(27): 1169-
Joy, L. 1973. Food and nutrition planning. Journal of American Economics 24.
Kennedy, E. and B. Cogill. 1987. Income and nutritional effects of the commercialization
ofagriculture in Southwest Kenya. IFPRI Research Report No. 63. Washington, D.C.
Knipscheer. H.C. 1989. The roles of CRSP social scientists in technology evaluation and
generation. Pages 249-256 in The socialsciencesin internationalagricultural research:
Lessons from the CRSPs. Boulder, Colorado, USA: Lynne Rienner publishers.
Kumar, S.K. 1977. Composition of economic constraints in child nutrition: Impact of
maternal incomes and Employment in low-income households. Ph.D. dissertation,
Cornell University, Ithaca, New York.
Longhurst. R. 1983. Cash crops, household food security, and nutrition. IDS Bulletin
Longhurst, R 1986. Household food strategies in response to seasonality and famine.
IDS Bulletin 17:27-35.
Longhurst, R. 1983. Agricultural production and food consumption: Some neglected
linkages. Food and Nutrition 9:2.
Malambo, L.M. 1987. Rural food security in Zambia. Ph.D. dissertation, Michigan State
University, East Lansing, Michigan, USA.
Mason, J.B., J.P. Habicht, H. Tabatabao, and V. Valverde. 1984. Nutritional surveil-
lance. World Health Organization, Geneva, Switzerland.
Maurya, D.M., A. Bottrall, and J. Farrington. 1988. Improved livelihoods, genetic
diversity and farmer participation: A strategy for rice-breeding in rainfed areas of
India. Experimental Agriculture 24(3):311-320.
Maxwell, S. 1989. Rapid food security assessment: A pilot exercise in Sudan. RRA Notes
5. International Institute for Environment and Development, London.
Maxwell, S. 1984. Health, nutrition and agriculture: Linkages in Farming Systems
Research. IDS Discussion Papers 198:32.
Maxwell, S., and M. Smith. 1992. Household food security: A conceptual review.
Institute of Development Studies, University of Sussex, Brighton, UK.
Mead, D.C. 1988. Nonfarm income and food security: Lessons from Rwanda. Pages 331-
338 in G.D. Mudimu and K.H. Bernsten, eds., Household and nationalfood security
in southern Africa: Proceedings of the fourth annual conference onfood security research
in southern Africa, October 31 November 3, 1988. University of Zimbabwe/Michigan
State University Food Security Research Project, Department of Agricultural Eco-
nomics and Extension, Harare, Zimbabwe.
Merrill-Sands, D.M., S.D. Biggs, R.J. Bingen, P.T. Ewell, J.L. McAllister, and S.V. Poats.
1991. Institutional considerations in strengthening on-farm client-oriented research
in national agricultural research systems: Lessons from a nine-country study. Ex-
perimental Agriculture 27:343-373.
Messer, E. 1989. Seasonality in Food Systems: An anthropological perspective on
household foodsccurity. Pages 151-175 in D.E. Sahn, ed., Seasonal Variability in
Third World Agriculture. Baltimore and London: International Food Policy Research
Institute and The Johns Hopkins Press.

Journal for Farming Systems Research-Extension


Molnar. A. 1991. Rapid Rural Appraisal methodology applied to project planning and
implementation in natural resource management. In J. Van Willigen and T. Finan.
eds., Soundings: Rapid and reliable research methods for practicing anthropologists.
National Association for the Practice of Anthropology, Bulletin 10.
Norman, C.W. 1991. Soil conservation: Using farming systems development as an aid.
Paper prepared for the Farm Management and Production Economics Service,
Agricultural Services Division, Food and Agriculture Organization (FAO), Rome.
Pinstrup-Anderson, P. 1981. Nutritional consequences ofagricultural projects: Conceptu-
al relationships and assessment approaches. World Bank Staff Working Paper No. 456,
Washington, DC.
Rhoades, R.E. 1989. The role of farmers in the creation of agricultural technology. Pages
3-9 in R. Chambers, A. Pacey, and L.A. Thrupp, eds., Farmerfirst: Farmer innovation
and agricultural research. London: Intermediate Technology Publications.
Saenz de Tejada, S. 1989. Food consumption and its relation to production: A survey in
Lesotho. University of Arizona, College of Agriculture, Nutrition in Agriculture
Cooperative agreement, Tucson, Arizona, USA.
Swift, J. 1989. Why are rural people vulnerable to famine? IDS Bulletin 20(2):8-16.
Thomas, R.B., S.H. Paine, and B.P. Brenton. 1989. Perspectives on socio-economic
causes of and responses to food deprivation. Food and Nutrition Bulletin 11:41-54.
Tobert, N. 1985. The effect of drought among the Zaghawa in northern Darfur. Disasters
Tripp, R. 1982. Farmers and traders: Some economic determinants of nutritional status
from northern Ghana. Food and Nutrition 8(1):3.
Velarde, N. 1991. The Zambian farming systems approach to studying household food
security. Paper presented at the Working Seminar on Dependency on Forest Foods for
Food Security, March 13-15, 1991. Swedish University of Agricultural Sciences,
International Rural Development Centre, Uppsala, Sweden.
von Braun. J., and E. Kennedy. 1986. Commercialization of subsistence agriculture:
Income and nutritional effects in developing countries. IFPRI Working Paper No. 1
on Commercialization of Agriculture and Nutrition, Washington, DC.Buchanan-
Smith, M., S. Davies, and R. Lambert. 1991. Aguide to famine early warning and food
information systems in the Sahel and Horn of Africa: A review of the literature. IDS
Research Reports No. 21. Institute of Development Studies, University of Sussex,
Brighton, UK.
de Waal, A. 1989. Famine that kills: Darfur Sudan 1984-1985. Oxford, UK: Clarendon
de Waal, A. 1988. Famine Early Warning Systems and the use of socio-economic data.
Disasters 12(1):81-91.
Walker, T., and N. Jodha. 1986. How small farm households adapt to risk. In P. Hazell,
ed., Crop insurance for agricultural development: Issues and experience. Baltimore,
Maryland: The Johns Hopkins University Press.
Watts. M. 1988. Coping with the market: Uncertainty and food security among Hausa
peasants. Pages 260-290 in I. De Garine and G.A. Harrison, eds., Coping with
uncertainty in food supply. Oxford: Clarendon Press.
World Bank. 1986. Poverty and hunger: Issues and options for food security in developing
countries. Washington, DC: World Bank.
World Resources Institute (WRI), Center for International Development and Environ-
ment. 1989. Participatory rural appraisal handbook: Conducting PRAs in Kenya.
Washington, DC: World Resources Institute.
Vol. 4, No. 1, 1993


Zandstra, H. G. 1991. Approaches in farming systems research and extension: Problems
and improvements. Journal of the Asian Farming Systems Association 1(1):101-111.

Journal for Farming Systems Research-Extension

Inability of Farming Systems Research to
Deal with Agricultural Policy'

Doyle Baker2


The premise ofthis paper is that lack ofattention to agricultural policy and
support systems has constrained the potential impact of farming systems
research-extension (FSRE) and has been a factor contributing to its
marginalization. The paper starts with a historical overview, stressing
FSRE's failure to activate a policy dimension, and gives several reasons
why FSRE practitioners can no longer ignore policies and support
systems. The paper then turns to strategic priorities for FSRE in the
1990s. The top priorities identified are: (1) renewed commitment to in-
depth household and village studies, (2) design and evaluation research
related to support systems and farmer programs, (3) increased attention
to governmental and societal goals in research priority setting, and (4)
taking greater responsibility for formulating research policy and ensuring
research system sustainability. The paper ends with discussion of three
crucial issues affecting prospects for operationalizing a policy dimension
in FSRE: uncertain demand for policy research, research portfolio
opportunity costs, and implications for FSRE multidisciplinarity.


A critical weakness of farming systems research-extension (FSRE) has been its
inability to deal with agricultural policy. Policy and support system issues have
not always been ignored but, for the most part, a passive approach has been
followed, treating factors exogenous to farming systems as parameters for
farming systems analysis. A passive approach toward policy and support
systems might have been desirable during the early days of FSRE, in order to
establish a niche within technical research institutes, define identifiable
boundaries, and secure FSRE project funding. However, it was based on
short-run thinking and reflected unfounded assumptions about technology
supply options, severity of constraints imposed by inappropriate policies, and
poorly performing support systems.
1 Keynote presentation at the Twelfth Annual Association for Farming Systems Research-
Extension Symposium, Michigan State University, East Lansing, September 13-18, 1992.
Agricultural economist, International Institute of Tropical Agriculture; currently leading the
Economic Analysis Unit, Institute of Agronomic Research, Cameroon.


The premise of this paper is that retaining a self-imposed limitation to
concentrate on adaptive research has reduced FSRE's potential impact and has
been a factor contributing to its marginalization. In order to reestablish
credibility, efforts to incorporate a policy dimension in FSRE will be required.
This paper undertakes a review and reassessment of FSRE research priorities
as a step toward this end. The paper starts with background information on
the role of policy and institutional analysis in FSRE. The following section
gives a critical appraisal of FSRE's passiveness with respect to policy-oriented
research. The remainder of the paper turns to strategic considerations in
revitalizing FSRE and increasing impact. The fourth section proposes prior-
ities for FSRE in the 1990s. The fifth section identifies critical issues affecting
prospects for adding a policy dimension to FSRE.

This section shows that the importance of the policy and institutional
environment was recognized during the early stages of farming systems
research methodology development. However, a policy dimension was never
operationalized as an integral component of FSRE. In fact, prospects for
policy and institutional analysis declined in the late 1980s due to FSRE
retrenchment. For the last several years, there has been increased emphasis on
assessments of adoption and potential impact, and these studies have provided
a weak policy linkage. Nevertheless, current thinking remains that FSRE
should play a minimal role in policy analysis, largely restricted to provision of

Policy and Institutions in FSRE Methodology
Most FSRE analytical frameworks, such as those developed by David
Norman and the CIMMYT economics program, included policies and
institutions as important external or exogenous circumstances affecting
farming systems. Norman (1980) identifies three sets of exogenous factors:
external institutions; miscellaneous factors such as population density; and
community structures, norms, and beliefs (Figure 1). Policy linkages enter
primarily through the first factor. Norman emphasizes in particular the role of
input and output markets. Gilbert, et al. (1980), presenting the same model,
argue that farming systems research is a vehicle for linking micro and macro
perspectives, and that modification of policies can be one of the most
important factors in efforts to improve circumstances of small farmers. They
mention policy and planning units as a potential institutional linkage for FSRE
and observe that linkages with such units could enable FSRE to address
improvements that are outside mandates of technical research institutes.
The CIMMYT manual (Byerlec et al., 1980; CIMMYT, 1988) identifies
markets and institutions as external economic circumstances affecting farming

Journal for Farming Systems Research-Extension


systems (Figure 2). Policies, which they define as actions and rules of
governments, shape the external and (presumably) internal economic envi-
ronment in which farmers operate. Several direct and indirect linkages
between policies and production decisions are described, as well as linkages
between policy and research. The CIMMYT manual also points out that
information from FSRE can be useful for stimulating changes in policy.
Although policies and institutions were prominently included in FSRE
analytical models, the primary function of FSRE was to design and screen
technologies.3 Exogenous circumstances were universally characterized as
parameters for technology development research rather than leverage points
for the improvement of farming systems. These parameters might change
during the planning stage of technical research, as was pointed out in the
CIMMYT manual, but efforts to initiate change in the policy environment
were not viewed as a core FSRE activity.
Aside from a predetermined focus on technical change, FSRE methodol-
ogy had at least two additional defining features that made the activation of
a policy dimension unlikely. The first was commitment to a strategy of
modifying technologies rather than modifying farmer circumstances.4 A
second was commitment to focused problem solving, minimum investments
in data collection, reliance on rapid appraisal survey methods, and quick
results.5 Taken together, these features created a strong presumption against
longer term investments in data generation for policy analysis and reform.
Thus, a basic tension with respect to policies and institutions was inherent in
SThere were several reasons for FSRE emphasis on technical change research. The most important
was the fact that early FSRE methodology development took place mainly in the International
Agricultural Research (IARC) system. Not incidental, however, were interrelated beliefs in the
need for technical change along with the promising prospects for this change. The former
derived from a decade of disappointing experiences with community development and integrat-
ed rural development. There was a great deal of interest in identifying an engine of development
(Eicher and Baker, 1982), and both Schultz (1964) and Mcllor (1976) provided convincing
arguments that technical change by small farmers is an important precondition for growth.
While dramatic yield increases immediately received much attention, it was only somewhat later
that distributional and risk consequences were appreciated, and latter still that adequate
recognition was given to the important role of policies and support systems.
4Disillusioned with technology transfer, FSRE pioneers were convinced that the problem was
inappropriate technologies, not unresponsive farmers. Thus, the top FSRE priority was
developing technologies corresponding to farmers' objectives, resources, and managerial
capacities. This objective, however, implied a goal of ensuring that technologies be consistent
with the external economic circumstances that determine farmers' opportunity sets and
5 This emphasis had two major roots, among others. One was disappointment by agricultural
development economists with long-term investments in data collection during the 1960s and
1970s (Eicher and Baker, 1982) and consequent interest in beginning to make "bread and
butter" contributions (Collinson, 1981). Another was concern that donor agencies would find
the approach too complex, long term, and costly (Gilbert, Norman, and Winch, 1980) because
of references to holism and use of systems terminology. In the practical world of research,
FSRE's quick turnaround orientation certainly was, to an extent, induced by USAID, which in
turn was operating under a congressional mandate to identify quick solutions.

Vol. 4, No. 1, 1993

Human Technical

n Chemical
Exogenous Endogenous Physical Bioligical
S -I Mechanical

Figure 1. Adapted from Gilbert, Norman, and Winch (1980).



Climate Biological
Climate Biological


Journal for Farming Systems Research-Extension









Figure 2. Adapted from CIMMYT (1988).


FSRE methodology.6 Policies and institutions were recognized to have an
important impact on farming systems, and the objective was to help farmers
by using a systems perspective. Policy reform and institutional performance
were written off as someone else's responsibility.

Broadening and Differentiation
During the early 1980s, when FSRE was initiated in many National
Agricultural Research Systems (NARS), the narrowly focused adaptive re-
search model was dominant. This was primarily due to predetermined
mandates and institutional location in technical research institutes.7 Soon,
however, it became apparent that greater contributions were expected of
FSRE in NARS than had been the case in IARCs.8 This led to a credibility
dilemma, as characterized by Norman and Baker (1986). They argued that a
growing group of "purists" worried that unacceptable compromises, made in
pursuit of time- and cost-efficient research, could do more harm than good.
Although Norman and Baker spoke critically of the gap between FSRE
philosophy and practice, they said donors and technical scientists probably
would react negatively to calls for broadening, longer term payoffs, and
pursuit of an independent social science agenda. They concluded that FSRE
practitioners must accept a narrow orientation toward technology screening
and introduction, and modestly proposed inclusion of whole farm studies as
a complement to rapid appraisal and baseline surveys.
A major consequence of the FSRE credibility dilemma was that FSRE's
agenda was broadened only with respect to implementation of technical
change research. Additional attention was given to noncrop production
farming systems components and intra- and interhousehold dynamics.9 Rural
institutions and policy reform were not included as part of the expanded
agenda. By the mid-1980s it was common practice to methodologically

Certain tensions and inconsistencies in FSRE methodology, including those related to policy,
can be traced to the rapid spread of FSRE during the early 1980s. It was necessary to forge a
consensus on defining features and guiding principles well in advance of adequate field testing
and assessment ofalternative approaches. Limited experiences and insights, built on foundations
no stronger than the experiences of any of hundreds of FSRE teams during the past decade, then
carried great weight. Much early methodology development actually derived from logic,
intuition, and reactions against negative past experiences as much as it did from positive new
Two additional, practical reasons why policy reform did not receive more attention were: (1)
widespread belief that policy reform was politically sensitive and driven by interests and pressures
beyond the capability of FSRE teams to influence, and (2) most FSRE economists had farm
management and production economics backgrounds-not policy, trade, or marketing.
FSRE in national settings often was assigned the task of identifying options for marginal areas
and resource-poor farmers. As Norman and Baker (1986) pointed out, this naturally led to
consideration of many factors affecting rural welfare. Not incidentally, it soon became apparent
that overly optimistic estimates had been made about technology supply options, and that
sustained efforts would be required.

Vol. 4, No. 1, 1993


differentiate between FSRE and policy-oriented research using a farming
systems perspective. Methodology manuals developed by the Farming Sys-
tems Support Project (FSSP), for example, used David Norman's (1980)
distinction between FSRE and FSIP-Farming Systems Approach to Infra-
structural Support and Policy-and stated, without explanation, that FSIP
was beyond the scope of its manual on diagnosis in farming systems research
(FSSP, 1987). Merrill-Sands (1986), in a frequently cited article, identified
several types of farming systems research-all distinguished from farming
systems adaptive research. The most directly concerned with policy input was
Farming Systems Research and Agricultural Development (FSRAD).10
The key point is that despite substantial broadening in FSRE's agenda and
methods during the early and the mid- 1980s, policy and institutional research
was not activated. Instead, alternative, complementary approaches were
defined into existence-such as FSIP and FSRAD. If anything, this had the
effect of further diminishing prospects for activating a policy dimension in
FSRE. No attention was paid to FSRAD, nor was consideration given to its
potential value versus more narrowly focused FSRE.

Retrenchment and Impact Focus
Although a FSRE policy dimension was not activated on a large scale during
the 1980s, there were isolated attempts at conducting farm-based policy
research. Martinez (1986) discussed the concept of "farm-based policy
research" at the 1986 farming systems symposium. Yates et al. (1988)
illustrated use of farm-based policy in Haiti. One of the most extensive
commitments to farm-based policy research was by the Agricultural Technol-
ogy Improvement Project (ATIP,1986) in Botswana (from 1982 to 1990)."
During the second half of the 1980s, debate resurfaced about FSRE's
failure to go farther in policy-oriented research. Davidson (1987), for
example, proposed that FSRE include wider policy issues, while Herdt (1987)
9 Selection of different themes each year for the farming systems symposium gave a practical
stimulus to field researchers to broaden their approach. Indicative of growing interest in
household studies was a Rockefeller-sponsored conference on intrahousehold dynamics held in
Bellagio, Italy in 1985. Several of the papers at that conference, later published in Moock
(1986), illustrated the complexity of farm-household systems. Competitive selection for the
Intra-Household Dynamics and Farming Systems Research and Extension Case Studies Project
also made it clear that the ideal in FSRE was evolving during the mid-1980s.
1In FSRAD, according to Merrill-Sands, farming systems are placed in a broad economic and
policy environment, and the objective is long-term development through institutional and
economic reform, as well as technological change. Merrill-Sands notes that farming systems
continue to be the main unit of analysis, but the boundaries of analysis are expanded so that
linkages between the farming systems and its social, economic, and political environment receive
as much emphasis as technical change.
Several types of farm-based policy analysis were carried out: on tractor use, cart inclusion in a
lending program, intervention in the local trading system, design and implementation of farmer
assistance programs, and extension service performance (ATIP, 1986). Experimental data were
used to develop arguments in favor of inclusion of double plowing, new donkey harness, and
a hand rotary injection planter in a farmer assistance program (Worman et al., 1991).

Journal for Farming Systems Research-Extension


warned against treating FSRE as a substitute for policy analysis. A question
that will forever remain unanswered is what direction FSRE might have taken
if financial support had continued for FSRE activities. By then, FSRE was
under challenge. Interest in FSRE had declined among donor agencies and
agricultural economists due to belief that FSRE was not generating sufficient
returns to justify further investment. Lack of impact was attributed in part to
weak extension linkages but also to a narrow focus that treated too many
potential leverage points as fixed.
During the last five years, tremendous emphasis has been given to achieving
and documenting short-run impact, both in FSRE projects and institutional-
ized programs. There has been a retrenchment with respect to investments in
farming systems diagnosis, and a shift in emphasis to developing better
extension linkages and participatory research methods. FSRE's retrenchment
and concern with impact has had mixed implications for FSRE-policy linkages.
On one side, there has been a decline in household and village studies that
might have contributed insight for farm-based policy analysis. On the other,
there has been a surge of interest in adoption and impact studies. As Norman
(1991) pointed out, adoption studies do provide an opportunity to analyze
institutional and policy constraints. Perhaps the best examples of this over the
last several years have been a series of mathematical programing analyses
carried out by various researchers currently or formerly affiliated with Purdue
University.12 Many adoption studies, however, have provided little input to
policy because they have been based on ad hoc selection of a limited range of
variables related to personal and technology characteristics.

Current Thinking
Despite growing concerns with FSRE impact, current thinking on FSRE-
policy linkages is not substantially different now than it was in the mid-1980s.
As reflected in Tripp ct al. (1990), the prevailing consensus among FSRE
practitioners is that on-farm research can provide relevant data for policy
analysis, but FSRE should not be a nucleus for agricultural development policy
formulation, and-reviving Herdt's (1987) argument-should not try to
substitute for conventional policy research. Tripp et al. do emphasize the need
to develop better communication channels so policy formation processes take
better advantage of data and insights generated by FSRE teams. This is not

12 Mathematical programming was used by Sanders and others in Burkina Faso, for example, to
assess impact of technology and, through sensitivity analysis, to incorporate resource access and
pricing perspective (Sanders, 1989; Sanders et al., 1990). Ngambeki, Deuson, and Lowenberg-
DeBoer (1990) used a programing model to evaluate potential impact of various technology
packages in northern Cameroon, including a comparison ofan FSRE-developed package and an
extension-recommended package-both versus existing farmer practices. Savadogo and Brandt
(1988) is a good analysis showing linkages between income, food demand, and technologies in
Burkina Faso. Lopez-Pereira et al. (1990) modeled new technologies and used sensitivity
analysis to assess effects of price changes.

Vol 4, No. 1, 1993


quite the passive role of the early 1980s, but stops well short of endorsing a
FSRAD or FSIP orientation.


Some care is needed in appraising FSRE's passive attitude toward policies and
support systems. FSRE was developed as an approach for adaptive research,
not as a comprehensive development philosophy or strategy. In terms of the
narrow goal of better research methods for screening technologies, FSRE
accomplishments have been substantial, as a number of assessments have
noted (e.g.,Tripp etal., 1990). The fact that technical change and impact have
been less than desired does not mean that FSRE has failed in its defined
domain. Nevertheless, after more than a decade of substantial investment in
FSRE activities, questions can legitimately be raised about the sufficiency of
FSRE's scope and contributions-and about the continuing gap between
farming systems analytical frameworks and FSRE in practice. This section
identifies three reasons FSRE has had less impact than it might otherwise have
had. All stem, to an extent, from lack of attention to policy and support
systems. The reasons are: (1) restricted focus on incremental change, (2)
failure to make any substantial input to policy reform, design or implementa-
tion, and (3) inadequate attention to changing development needs. The
implication is that FSRE practitioners should no longer ignore policy and
support systems even if they want to operate mainly in a restricted FSRE
framework as opposed to a more comprehensive FSRAD framework.

Focus on Incremental Change
FSRE was motivated by belief that adaptive research focused on incremen-
tal technical change could in aggregate substantially contribute to agricultural
sector output, employment generation, and farmer welfare. Both logic and
empirical evidence suggest that FSRE's preoccupation with incremental
technical change has represented a serious underestimation of the range and
severity of policy and institutional constraints affecting development progress
and farmer welfare.13 Even after it became apparent that constraints from
inappropriate policies and support systems prevented significant impact, there
was a failure to adjust strategic orientation and decrease resources devoted to
"tinkering" (Tripp et al., 1990) with existing systems.

13 Timmer (1991), for example, says that most development analysts believe inappropriate policies
are the major constraint limiting the impact of agricultural development. Summarizing findings
from an important conference on accelerating food production in Africa, held in Zimbabwe in
1983, Mellor, Delgado and Blackie (1987) say the evidence is clear that the extent to which food
production growth can be increased depends on appropriate government policy. Several papers
presented at the 1984 conference on crisis and recovery in Africa point to the importance of
policy and structural constraints (Rose, 1985). Michael Lipton (1985) explicitly argued that the
reason returns to African agricultural research have been dismal is an inadequate policy

Journal for Farming Systems Research-Extension


Several negative consequences resulted because FSRE left too many factors
There has been an overemphasis on adaptive research relative to in-
depth diagnosis (Byerlee and Tripp, 1988).
Many teams have identified a relatively small number of trial themes and
then wasted scarce resources on minor trial variations over a number of
years even when itwas apparent that ultimate adoption and impact was
likely to be quite small. Long-term investments in fertilizer levels trials
in high rainfall zones and tied ridging and double plowing in low rainfall
zones are examples of these.
Much FSRE data has not related to important decision issues and
consequently good use has not been made of FSRE investments in data
collection (Ellis, 1992).
Efforts at institutionalization have focused almost exclusively on re-
quirements for adaptive research. There have been only a few isolated
attempts to introduce institutional changes to better accommodate a
policy linkage role. This has prevented effective feedback even in cases
when useful insight was generated through on-farm research.

When the full range of consequences are considered, FSRE's commitment
to incremental change within existing constraints became a "slippery slope"
to an inability to deal with policy.

Little Policy Input
As mentioned above, there is widespread agreement among leading FSRE
practitioners that FSRE should be providing information for use in policy-
making processes, even if FSRE teams are not in a position to actively
participate in policy making. The research record is not good in this regard.
Providing feedback on policies and support systems has not been treated as a
core activity-even though a change in policy may easily have bigger impact
than a minor change in technology. FSRE research papers based on baseline
studies and rapid appraisal surveys often include descriptive findings that have
implications for policy, but rarely have investments been made in empirical
research specifically directed at improving policies, programs, and support
system performance. Consequently, there have been several missed opportu-
nities in which FSRE teams could have provided needed empirical information

framework. In another study, Lecaillon ct al.(1987) analyzed interrelations between economic
policies, exogenous factors, and performance in six low- income countries over the period of
1960 to the early 1980s. They found substantial evidence that policies had a major effect on the
pattern of development. They stated that if there is one lesson from their comparative study, it
is that governments have to encourage agricultural development through appropriate institu-
tions and policies. Another conclusion from the study was that individual policy measures-such
as investing in adaptive technical research-need to be reinforced by complementary measures
and institutions if they are to be effective.

Vol. 4, No. 1, 1993


and a farm-level perspective to policy design, implementation, and monitor-
ing. Three key areas in which FSRE missed opportunities are the following:
Structural adjustment programs nearly allgo beyond market liberalization.
Adjustment programs often address other determinants of farmers' opportu-
nities and incentives-such as land tenure systems, farmer organizations and
cooperatives, and marketing boards-about which there is little agreement
among policy analysts (Timmer, 1991). Empirical data are needed to assess
possible outcomes of such structural changes, and FSRE teams are well placed
to provide data and insights by virtue of already being in the field with close
farm-household contacts.
Many food policy analyses and prescriptions are carried out at an aggregate
level, with much reliance on assumptions or "stylized facts." As Michigan State
University researchers have found in their food security research, microlevel
consequences for particular subgroups of households often differ from
aggregate effects, and various stylized facts often do not correspond well with
reality (Weber et al., 1988). FSRE field level data and perspectives again could
have provided valuable input into food policy debates.
Costs of many farmer assistance programs are higher than they need to be.
Appropriate targeting can greatly increase efficiency and effectiveness (Pin-
strup-Andersen, 1985; Gittinger, et al., 1987), as can studies directed at
reducing implementation errors (Mellor and Mudahar, 1992). Again, FSRE
field teams with their on-farm research skills and farmer contacts are well
placed to provide feedback on targeting and implementation guidelines.

Ignored Changing Development Needs
A reasonable argument can perhaps be made that FSRE need not assume
responsibility for policy formulation. However, a key goal of FSRE has been
to develop technologies consistent with external economic circumstances.
Those circumstances have changed dramatically during the last 15 years, but
FSRE for the most part has stayed with its initial focus on staple food
production by resource-poor farmers. Several different criteria should have
been taken into account when determining the research agenda.
During the past decade, a "silent revolution" (FAO, 1987) has been going
on in the agricultural policies of most countries. As part of broader structural
adjustment programs, agricultural prices have been freed, subsidies phased
out, and parastatals disbanded. In consequence, farmers' incentives and
opportunity sets have changed dramatically, putting tremendous pressure on
managerial capacity to handle adjustments in farm-household strategies.
Adaptive trials on staple food crop varieties have at best a minimal role in
helping rural households cope with their new circumstances.
Facing financial crises, many governments have reassessed agricultural
sector goals defined during the 1970s, which emphasized food self-sufficiency
and concentration on resource-poor farmers. The standard goal of self-

Journal for Farming Systems Research-Extension


sufficiency now is often accompanied with goals related to foreign exchange
generation, diversification, and employment stimulation-and occasionally
sustainability. In response to changing goals, research systems have been asked
to help reduce costs and raise profitability of exportable crops and to facilitate
diversification. Again, FSRE has largely failed to respond.
Complex linkages between technical change, development, and farmer
welfare have been identified in several agricultural development and food
policy analytical models.14 These analytical models make it clear that different
types of technologies have different distributional consequences and effects
on risk depending on the policy and institutional environment. It seems likely
that FSRE impact could have been increased, even in the face of severe support
system constraints, if priorities had been adjusted to take into account various
forward and backward linkages and multiplier effects-rather than only direct
impact on farm productivity.

Vernon Ruttan (1982) provides an appropriate guideline for determining
research priorities. He proposed that research resources be directed toward
releasing the physical and institutional constraints on production that are most
inelastic. Ruttan, at another place in the same book, cautioned that the fact
that research is conducted with skill and imagination is not evidence that it is
worth doing. Taking both pieces of advice into consideration, it seems
apparent that FSRE priorities for the 1990s need to be somewhat different
than they were in the 1980s. This section points out four high priorities
directed at improving policy linkages and hopefully increasing FSRE impact
and credibility: (1) investing more in household and village studies; (2)
initiating some research on farmer programs, support systems, and rural
infrastructure; (3) paying more attention to governmental and societal goals
in technical research priority setting; and (4) taking greater responsibility for
research policy and research system sustainability. These priorities will have to
be weighed against current and other prospective priorities in each research

14 Mcllor (1976) and Hirschman (1977) have, for example, emphasized backward and to a lesser
extent forward linkages for employment generation. Mellor also has stressed multiplier effects
of certain types of technologies. Timmer et al. (1983) give an excellent overview of economic
theory on linkages between production and marketing. Analytical frameworks are also covered
in Gittinger, Leslie, and Hoisington (1987) and Stevens and Jabara (1988). Eicher and Staatz
(1990) includes a number of papers addressing linkages by leading development economists.
Theories and empirical evidence for Africa, Asia, and Latin America are covered in Martin

Vol. 4, No. 1, 1993


Household and Village Studies
Anthropological and microeconomic household studies provided substan-
tial insight about development dynamics in the decades preceding the
emergence of FSRE. Findings related to the efficiency of small farm units
clearly played a role in reorienting agricultural policy. During the mid-1970s,
as FSRE methodology was being developed, longer term investments in
household and village studies fell out of favor. Major concerns were the time
and cost involved, as well as the lack of practical results (Eicher and Baker,
1982). Despite the prominence of rapid appraisal diagnosis in FSRE method-
ology, there were some FSRE projects and programs that invested in multiple-
visit, microeconomic and anthropological studies of farm households and
village institutions (e.g., Moock, 1986; Poats et al., 1988; Feldstein and Poats,
1989). Many FSRE practitioners to this day feel that insight from such in-
depth studies form the real core of FSRE's potential contribution to develop-
Tripp et al. (1990) justifiably warn that there is often limited capacity to
take advantage of insights generated through in-depth household and village
studies. Nevertheless, there is a need now for FSRE researchers to increase
their own understanding of the complex set of factors affecting farmer
welfare-even if there is little prospect that non-FSRE researchers and
planning units will take full advantage of results. The initial strategy of FSRE,
focused on resource-poor farmers and incremental change, has not worked.
There is an urgent need to end rigid prescriptions and stop focusing on short-
run issues stemming from predetermined mandates. Instead, as many devel-
opment specialists have argued, there is a need for longer term vision and
big-picture thinking.15 A strength of FSRE is its multidisciplinarity and
systems perspective. Building on this, FSRE researchers can work on develop-
ment paradigms that cut across neoclassical and political economic perspec-
tives, integrating perspectives from social sciences other than economics. This
can he a unique, identifying contribution that could help establish FSRE
credibility in the long run. Making this contribution will require more
substantial investment in household and village studies in order to establish a
firm empirical basis for vision, strategic thinking, and development planning.

Programs and Support Systems
It is not entirely clear why FSRE has given so little attention to support
systems. Isolation from policy formulation processes is only a partial excuse.
In actual fact, FSRE teams are in a unique position to provide farm-level

15 For example, Eicher (1990) said there is a need to identify the "tough agrarian" problems that
need to be addressed by research. Timmer et al. (1983) also emphasized the need for long-run
vis.s n of how food systems evolve under alternative policy environments. Even among FSRE
spe..alists, there have been calls for "new realistic paradigms for agricultural development"
(Roling and Fernandez, 1990).

Journal for Farming Systems Research-Extension


insight into the performance of farmer assistance programs and support
systems. If reasonable communication channels can be established with policy
and planning units, it would certainly seem that FSRE teams could make
substantial contributions by adding support system studies to their research
portfolios, even as a minor component of overall programs. The major priority
would appear to be performance of input and output markets and how these
affect farmers' incentives to adopt technologies. Planning, policy, and statis-
tical units often do not have institutional capacity to evaluate marketing
margins, input availability, and price fluctuations. A second priority is research
on extension service activities and performance. In countries where there are
farmer assistance programs, monitoring of these would also be a fruitful
possibility for FSRE contribution.
Research investments directed at improving farmer programs and support
systems would not have to be large. Experiences in Botswana, for example,
show that valuable insight can be generated through special subject surveys
using both rapid appraisal techniques and single-visit formal questionnaires.
There are three primary contributions farm-level surveys and farmer feedback
can make to improve policies and programs related to support systems: (1)
targeting, (2) implementation guidelines, and (3) evaluation of distributional
consequences. FSRE already is committed to improving technical research
through these same contributions. Policy formulation has need of similar
input from FSRE field teams.

Governmental and Societal Research Goals
Farmer goals do not encompass or reflect all potential contributions
research can make to social and economic development. During the 1990s,
many countries will continue to face economic crises that will not be solved
by increased food production. Agricultural research systems need to help
improve the profitability of export crops and create options for diversification.
As Davis, et al. (1987) argue, decisions about research resource allocations
should be based on national priorities within the existing policy setting. FSRE
teams are not an exception, and need to give increased attention to efforts
needed to solve national problems and crises.
In practical terms, FSRE teams should shift some resources to work on
industrial crops and perhaps specialty crops such as fresh vegetables and fruits.
Post-harvest technologies would appear to be another long-neglected priority
area. In making decisions from a larger, societal perspective, FSRE practitio-
ners need to pay more attention to analyses of comparative advantage and
research spillovers.16
16 Analyses using domestic resource cost (DRC) ratios, policy accounting matrices (Monke and
Pearson, 1989), and social accounting matrices (Pyatt and Round, 1985) are quite useful for
evaluating research options from a societal perspective. Enterprise data generated by FSRE field
teams can be valuable inputs into these analyses. An economic framework for evaluating research
spillovers is given in Davis, Oram, and Ryan (1987).

Vol. 4, No. 1, 1993


Research Policy and Sustainability
Byerlee and Franzel (1990) argue that FSRE researchers do not have a
comparative advantage for conventional policy research. This is true for the
most part but almost certainly does not hold for research policy.17 FSRE teams
have several unique advantages for research policy formulation. For example,
they form a bridge among major actors in technology development and
dissemination-commodity researchers, farmers, and extension services. Also,
they have field insight and experiences based on surveys and on-farm tests, a
systems perspective, and familiarity with the importance of problem solving
and priority setting.
From a practical standpoint, economists are still a scarce resource in many
NARS. Even if economists are concentrated in FSRE units, they should
assume responsibility for providing economics input into research policy
formulation. The main additional tasks to take on with respect to research
policy are: (1) analyses of linkages between national development policies,
farming systems dynamics, technical change requirements, and farmer wel-
fare; (2) research budgeting and programing, taking into account ex ante
assessment of research profitability; and (3) monitoring and impact evaluation
of all programs, not just FSRE activities.
The most compelling reason for FSRE teams to take greater responsibility
for research policy is self-preservation. At all times, success in farming systems
research depends greatly on technology supply options, so FSRE teams have
a strong incentive to help research systems work better. During times of
economic crisis, agricultural research funding often suffers as is now the case,
for example, in Cameroon. Scarce research resources often need to be directed
at protecting genetic material and maintaining minimal progress in core
commodity research programs. Thus, prospects for FSRE funding depend on
sustainability of funding for research systems. Unless overall research policy is
appropriate, it will not matter if correct priorities have been established for


Several problems are likely to be encountered in attempts to reorient FSRE in
the directions suggested above. Three stand out: (1) ensuring there is demand
for policy input and developing a strategy for effectively making input; (2)
projecting and assessing research portfolio opportunity costs; and (3) coping
with effects on multidisciplinarity.

17 Research policy encompasses decisions and operating procedures governing research resource
allocation, structures, and activities.

Journal for Farming Systems Research-Extension


Demand for Policy Research
Client demand has always been an important issue in FSRE. Reorienting
research so that farmers are viewed as clients was identified by Norman (1980)
as a defining feature of FSRE. During the past decade, farmer demand did not
always drive FSRE as much as it might have, but there still was demand for
certain types of research derived from institutional mandates or project goals.
Where is the demand for input to policy, farmer assistance programs, and
programs to improve support systems?
On relatively rare occasions, policy and planning units turn to FSRE field
teams to provide information and insight. This was the case, for example, in
Botswana, when several farming systems teams were asked to collaborate on
a field-level assessment of a major drought relief program. There was interest
in improving targeting and implementation guidelines, and when results of
the FSRE survey became available, several useful policy modifications were
quickly implemented. That was the exception; in nearly all other cases farm-
based policy research in Botswana ended up in research monographs. Insights
and recommendations on support systems design and targeting had little
impact largely because FSRE teams were isolated from on-going policy
dialogues. The lesson from Botswana is that shifting FSRE resources into
policy feedback will be of little benefit unless there is effective demand for
research results.
Michigan State University food security researchers have identified lack of
effective demand as a general problem in policy-oriented research (Weber et
al., 1988). Drawing on nearly ten years of food security research in Africa,
Weber et al. say that lack of demand was partly due to politics and ideology,
but equally important was the failure of policy analysts to respond to perceived
needs and to provide timely results. The Michigan State researchers concluded
that achieving impact from policy research requires investment of resources in
activities that increase demand for results. Two means of doing this are
workshops and working papers.
FSRE teams have an added disadvantage relative to food security research-
ers-their isolated location. A key issue to be addressed before there is
reallocation of resources to activate a policy dimension in FSRE is what are the
viable institutional options for establishing linkages and entering dialogue
with policy makers.

Research Portfolio Opportunity Costs
There is no doubt that household studies based on multiple-visit surveys or
in-depth participation observation are expensive relative to farmer-managed
trials. Special subject surveys directed at farmer programs and support systems
are not expensive, but often have only short-run impact. Due to expense and
uncertain impact of household and support systems studies, an important
issue is the opportunity cost with respect to testing and liaison activities.

Vol. 4, No. 1, 1993


Obviously, opportunity costs are less if there is a great deal of complementarity
between technology assessment and policy feedback, as would hopefully be
the case for household studies. Other factors also enter the equation. For
example, Norman and Baker (1986) argued that opportunity costs for farm-
based policy research are lower in areas with harsh climates because prospects
for technical change are limited. In areas with equable conditions, opportunity
costs could be quite high.
The key point is that priority setting and research resource allocation will
be substantially more complex if an attempt is made to broaden the scope of
FSRE. There is a weak empirical base for making assessments ofeither research
costs or benefits. Research impact studies are needed for making informed
decisions, as well as for justifying research and ensuring research system
sustainability. Some attention needs to be given to cost-efficient methods for
long-term household studies.

Multidisciplinarity and Skills
Multidisciplinarity has long been both a strength and problem of FSRE: a
strength in the sense that multiple disciplinary perspectives have been used to
address problems; a weakness because representatives of different disciplines
often have not worked well and easily together. In cases in which there has
been close collaboration, or even interdisciplinarity, this has usually been
because of shared commitment to the goal of testing and disseminating
technologies. Relations often have been less easy when social scientists have
concentrated on survey diagnosis, leaving testing activities to their technical
science colleagues. When considering an increased policy dimension in FSRE,
two major problems are likely to be encountered with respect to multidisci-
1. Even the modest agenda of priorities proposed above will force FSRE
social scientists to spend less time collaborating on on-farm tests, and perhaps
even on extension liaison activities. The big question is whether the gains from
new contributions will offset the costs of less interdisciplinarity.
2. Effective policy research requires reasonable knowledge of economic
theory and analytical frameworks. Effective FSRE adaptive research requires
some knowledge of several disciplines. Given limited human capacities and
opportunities to gain needed experience, it is not certain that FSRE social
scientists will be able to do policy-oriented research without substantial
retooling in economics.

Journal for Farming Systems Research-Extension


Certain points made in this paper should encounter few objections. For
example, it is hard to envision serious opposition to the importance assigned
to agricultural policies and support systems. Similarly, the fact that FSRE has
had only a modest impact on farmer welfare seems unquestionable. But should
the two be linked? This is where reasonable doubts can be raised. FSRE staked
out a domain and can not be held accountable for the multiplicity of factors
that have constrained progress over the past 10-15 years. Therefore, making
a call for activating a policy perspective in FSRE would be ludicrous if this were
really just a way of saying we should have been putting our efforts into policy
reform rather than FSRE. That is definitely not the intent of the paper.
To an extent, both "FSRE" and "policy dimension" are used as proxy
terms. Criticisms of FSRE are mainly objections to predetermined mandates
to investment in adaptive research, and to self-imposed limitations. The plea
for a policy dimension is really a call for modifying existing FSRE portfolios
in certain specific directions that hopefully will increase impact and credibility.
Two of the proposed priorities have drawbacks that require caution. For
household studies, it is cost and weak capacity to use the information
generated. For research related to support systems, it is effective demand.
Nevertheless, investments in developing better understanding of system
dynamics and in identifying support system options for improving farmer
welfare may be needed to re-establish credibility. The third and fourth
priorities have few obvious negative elements and should receive immediate
attention. FSRE is a public sector supported activity, so its portfolio needs to
address societal goals besides food production by resource-poor farmers. In
the light of economic crisis and competition for limited public resources,
FSRE also needs to contribute all it can to research policy and research systems
Any change in orientation has opportunity costs and uncertainty. If FSRE
had an outstanding record of helping farmers and contributing to national
development goals, then there might be a presumption against change. The
record, however, is not outstanding, and it is time to explore additional

ATIP (Agricultural Technology Improvement Project). 1986. Farming System research
activities at Mahalapye: Summary of activities, 1982-85. ATIP Research Report
Number 1. Ministry of Agriculture, Gaborone, Botswana.
Byerlee, D., et al. 1980. Planning technologies appropriate to farmers: Concepts and
procedures. Mexico, DF: CIMMYT.
Byerlee, D., and S. Franzel. 1993. Institutionalizing the role of the economist in national
agricultural research institutes. CIMMYT Economics Working Paper 93-01. CIM-

Vol. 4, No. 1, 1993


MYT, Mexico, DF.
Byerlee, D., and R. Tripp. 1988. Strengthening linkages in agricultural research through
a farming systems perspective: The role of social scientists. Experimental Agriculture
Collinson, M. 1981. Micro-level accomplishments and challenges for the less developed
world. Pages 43-53 in G.L. Johnson and A. Maunder, eds., Rural change: The
challenge for agricultural economists: Proceedings of the seventeenth ICAE. Banff,
Canada, September 3-12, 1979. ICAE, Westmead, UK
CIMMYT. 1988. From agronomic data to farmer recommendations: An economics
training manual. Mexico, DF: CIMMYT.
Davidson, A.P. 1987. Does farming systems research have a future? Agricultural
Administration and Extension 24:69-77.
Davis, J., et al. 1987. Assessment of agricultural research priorities: An international
perspective. Canberra: Australian Centre for International Agricultural Research.
Eicher, C. 1990. Building African scientific capacity for agricultural development.
Agricultural Economics 4:117-143.
Eicher, C.K., and D. Baker. 1982. Research on agricultural development in sub-Saharan
Africa:A criticalsurvey. International Development Paper 1. Michigan State Univer-
sity, East Lansing, Michigan, USA.
Eicher, C.K, and J.M. Staatz, eds. 1990. Agricultural development in the third world
(second edition). Baltimore: The Johns Hopkins University Press.
Ellis, F. 1992. Agricultural policies in developing countries. Cambridge: Cambridge
University Press.
Farming Systems Support Project (FSSP). 1987. Diagnosis in farming systems research
and extension. FSRE training units: Volume 1. Farming Systems Support Project,
Gainesville, Florida, USA.
FAO. 1987. Agricultural price policies: Issues and proposals. FAO Economic and Social
Development Series 42. FAO, Rome.
Feldstein, H., and S. Poats. 1989. Working together: Gender analysis in agriculture.
Volume 1: Case studies. West Hartford, Conneticut: Kumarian Press.
Gilbert, E., et al. 1980. Farming systems research: A critical appraisal. Rural Development
Paper 6. Michigan State University, East Lansing, Michigan, USA.
Gittinger, J.P., et al. 1987. Food Policy: Integrating supply, distribution, and consumption.
Baltimore: The Johns Hopkins University Press.
Herdt, R. 1987. Whither farming systems. Paper presented at Seventh Annual Farming
Systems Research-Extension Symposium, Little Rock, Arkansas, USA.
Hirshman, A. 1977. A generalized linkage approach to development, with special
reference to staples. Economic Developmentand CulturalChange25 (supplement):67-
Lecaillon, J., et al. 1987. Economic policies and agricultural performance of low-income
countries. OECD, Development Centre Studies, Paris.
Lipton, M. 1985. Research and the design of a policy framework for agriculture. Pages
78-93 in T. Rose, ed., Crisis and recovery in sub-Saharan Africa. Paris: OECD.
Lopez-Pereira, M., et al. 1990. Farming systems and adoption of new agricultural
technologies: An economic evaluation of new sorghum cultivars in Southern Hondu-
ras. Journal for Farming Systems Research-Extension 1:81-103.
Martin, L., ed. 1992. A survey ofagricultural economics literature, Volume 4: Agriculture
in economic development, 1940s to 1990s. Minneapolis: University of Minnesota Press.

Journal for Farming Systems Research-Extension


Martinez, J. 1986. Toward farm-based policy research: Learning from experience. Paper
presented at Sixth Annual Association for Farming Systems Research-Extension
Symposium, Manhattan, Kansas, USA.
Mellor, J. 1976. The new economics of growth: A strategy for India and the developing
world. Ithaca: Cornell University Press.
Mellor, J., C.L. Delgado, and M.J. Blackie. 1987. Accelerating food production in Sub-
Saharan Africa. Baltimore: The Johns Hopkins University Press.
Mellor, J., and M. Mudahar. 1992. Agriculture in economic development: Theories,
findings and challenges in an Asian context. Pages 331-542 in L. Martin, ed., A survey
of agricultural economics literature, volume 4: Agriculture in economic development,
1940s to 1990s. Minneapolis: University of Minnesota Press.
Merrill-Sands, D. 1986. Farming systems research: Clarification of terms and concepts
Experimental Agriculture 22:87-104.
Monke, E.,and S. Pearson. 1989. The policy analysis matrixfor agricultural development.
Ithaca: Cornell University Press.
Moock, J., ed. 1986. Understanding Africa's rural households and farming systems.
Boulder, Colorado, USA: Westview Press.
Ngambeki, S. et al. 1990. The impact of farming systems research on agricultural
productivity: The case of North Cameroon. Journal for Farming Systems Research-
Extension 1:125-152
Norman, D. 1980. The farming systems approach: Relevancy for the small farmer. Rural
Development Paper 5. Michigan State University, East Lansing, Michigan, USA.
Norman, D. 1991. Farming systems research in a declining donor environment. Staffpaper
92:2. Department of Agricultural Economics, Kansas State University, Manhattan,
Kansas, USA.
Norman, D., and D. Baker. 1986. Components of farming systems research, FSR
credibility, and experiences in Botswana. Pages 36-57 in J. Moock, ed., Understand-
ingAfrica's ruralhouseholds and farming systems. Boulder, Colorado, USA: Westview
Poats, S., et al. 1988. Gender issues in farming systems research and extension. Boulder,
Colorado, USA: Westview Press.
Pinstrup-Andersen, P. 1985. Food process and the poor in developing countries
European Review of Agricultural Economics 12:69-85.
Pyatt, G., and J. Round, eds. 1985. Social accounting matrices: A basis for planning.
Washington, DC: The World Bank.
Rose, T., ed. 1985. Crisis and recovery in sub-Saharan Africa. Paris: OECD.
Roling, N., and M. Fernandez. 1990. Farmers technology criteria and institutional
assumptions: Exploring mis-anticipation for points of leverage. Paper presented at
Tenth Annual Association for Farming Systems Research-Extension Symposium, East
Lansing, Michigan, October.
Ruttan, V.W. 1982. Agricultural research policy. Minneapolis: University of Minnesota
Sanders, J. 1989. Agricultural research and new technology introduction in Burkina Faso
and Niger. Agricultural Systems 29:139-54.
Sanders, J., et al. 1990. Developing new agricultural technologies for the Sahelian
countries: The Burkina Faso case. Economic Development and Cultural Change 39:1-
Savadogo, D., and J. Brandt. 1988. Agricultural development principle: Economic theory

Vol. 4, No. 1, 1993


and empirical evidence. Baltimore: The Johns Hopkins University Press.
Schultz, T.W. 1964. Transforming traditional agriculture. New Haven: Yale University
Stevens, R.D., and C.L. Jabara. 1988. Agricultural development principles: Economic
theory and empirical evidence. Baltimore: The Johns Hopkins University Press.
Timmer, C.P., ed. 1991. Agriculture and the state: Growth, employment and poverty in
developing countries. Ithaca: Cornell University Press.
Timmer, C.P., et al. 1983. Food policy analysis. Baltimore: The Johns Hopkins University
Tripp, R., et al. 1990. Farming systems research revisited. Pages 384-99 in C. Eicher and
J. Staatz, eds., Agricultural development in the third world (second edition). Balti-
more, Maryland: The Johns Hopkins University Press.
Weber, M., et al. 1988. Informing food security decisions in Africa: Empirical analysis and
policy dialogue. American Journal of Agricultural Economics 70:104-52.
Worman, F., G. Heinrich, and D. Norman. 1991. Some organizational considerations in
implementing FSRin a harsh environment: Increasing farmer involvement in Botswa-
na. Journal for Farming systems Research-Extension 2:119-137.
Yates, M., et al. 1988. Fertilizer in Les Cayes, Haiti: Addressing market imperfection with
farm-based policy analysis. CIMMYT Economics Working Paper 88/01. CIMMYT,
Mexico, DF.

Journal for Farming Systems Research-Extension

Methods for Analysis by Farmers:
The Professional Challenge'

Robert Chambers2

Until recent years, professional attention has concentrated on improving
analysis of farming systems. The challenge now is to document and
explore the scope of methods for enhancing farmers' own analyses.
Scientists and extensionists have informally been developing and using
such methods. Farmer participatory research (FPR) and participatory
rural appraisal (PRA) are complementary and overlapping sources of
experience. FPR methods are more verbal and observational, while PRA
methods are more visual, including participatory mapping, analysis of
aerial photographs, matrix scoring and ranking, flow and linkage dia-
graming, seasonal analysis, and trend diagraming. The strength of these
visual methods is clear. Farmers have a greater capacity to diagram and
analyze than most outsiders have supposed, and farmers are proving to be
good facilitators of analysis. The professional challenge is further to
develop, spread, test, and improve farmers' analysis through these and
other methods.

It has become a common belief that farmers know more, and know it better,
than the agricultural professional community once thought. Farmers' knowl-
edge does have limitations, and scientific methods and knowledge remain
enormously powerful and effective in some domains; but it is evident that
farmers have access to information that is available to scientists only at high
cost. So as we emerge from a phase of professional arrogance, an increasing
number of scientists and extensionists accept and seek to exploit the strengths
of farmers' knowledge, and to learn from and with them. On-farm research
has become common. Methods for on-farm research have been developed and
published. Farmer-designed and farmer-managed trials are less rare. In many
places, agricultural scientists and extensionists have been developing ap-

1Keynote presentation at the Twelfth Annual Association for Farming Systems Research
Extension Symposium, Michigan State University, East Lansing, September 13-18, 1992.
Institute of Development Studies, University of Sussex, Brighton, BN1 9RE, UK.


preaches and methods for working more closely with farmers. But where is this
going? What is next?
Farmers' capabilities extend far beyond mere knowledge. They have always
observed, experimented, adapted, improvised, and, in Paul Richards' (1989)
word, "performed." In all these activities they also analyze. The theme of this
paper is that one challenge now facing the agricultural professions is the
development and dissemination of approaches and methods that farmers can
use to enhance their own analysis.
This has been something ofa blind spot. On-farm and with-farmer research
has spread and generated a substantial literature, but how much has been
written on methods for analysis by farmers? With rather few exceptions, some
of which are cited below, the Consultative Group on International Agricul-
tural Research (CGIAR) has not recognized farmers' analysis as a relevant
subject. A letter written about five years ago to all the Directors-General of the
International Agricultural Research Centers (IARCs) asking them for details
of methods being used or developed for analysis by farmers drew a blank: as
I recall, not one reply was received. Until recently, the almost universal
assumption has been that "we" investigate, collect data, analyze and then
prescribe, for "them"; the literature and training have focused on improving
"our" understanding and analysis rather than "theirs."
Enhancing farmers' own analysis has recently begun to receive more
attention. No doubt the literature on farmer-designed and farmer-managed
on-farm research is one source of this attention. Lightfoot et al. (1991), for
example, present ways in which farmers can be enabled to choose between
potential research topics. Additionally, Quiros et al. (1990:23) point out that
"open-ended evaluation stimulates the farmer to think. It can be used to help
the farmer marshal thoughts in a sequence, to recall past evaluationss, and to

Table 1: Research and Extension: Some Dominant Beliefs and Modes, 1950-2000.
Explanation Prescription Key Socio- Methods Label
of Non- Activities Economic
Adoption Research
Farmers' Extension Teaching Diffusion, Question- Diffusion
C ignorance education adopters, naire surveys research
laggards, etc.
SFarm-level Removal of Input Constraints, Question- Farming
o o Constraints Constraints Supply Farming naire Surveys, Systems
Systems On-Farm Research
S Inappropriate Farmer Facilitation Participatory Discussion, Farmer
o\ Technology participation approaches observation, participatory
and diagraming research,
methods by and with Farmer-
Farmers First,
PRA etc.

Journal for Farming Systems Research-Extension


Table 2: Two Modes of Intervention.

Extractive Enabling

Outsiders' roles in diagnosis Obtain and analyze data Facilitate farmers' own analysis
Outsider's role
with technology Prescribe and transfer Search and supply
Farmers are... Passive Active
Farmers... Provide data observe, analyze
Adopt Demand
Follow instructions Test, experiment
Analysis is by... Us Them

get into the swing of thinking critically about technology." Table 1 presents
an interpretation of changes in professional views and norms-past, present,
and future. A brief caricature such as this both simplifies and distorts. Readers
will judge whether the shifts shown are significant and substantially true. The
line for the 1990s in this table extrapolates hopefully from present trends.
In the context of analysis, two polarized modes of intervention can be
distinguished: extractive and enabling. Their characteristics are suggested in
Table 2. The dominant assumption has been that analysis is mainly by "us."
The question now is how much future analysis should be by "them."
The thesis of this paper is that such analysis is desirable and can be
enhanced. Its desirability is based on the following: the comparative advantag-
es of farmers' knowledge; the complexity, diversity, and variability of farming
systems; the scarcity of scientists and extensionists; and avoidable past
wastefulness in agricultural research and extension. The hypothesis is that if
farmers can be enabled to analyze their own farming systems, they will have
a huge head-start in the range, intimacy, and accuracy of knowledge. They will
filter out or ignore masses of irrelevant information, know and promote their
own priorities, and be committed to action. That farmers' analysis is possible
is self-evident. They do it all the time to survive. Whether and how this analysis
can be enhanced is the subject of this paper.

Two streams of innovation that are contributing to methods for farmers'
analysis can be noted: Farmer Participatory Research (FPR) (Farrington,
1988; Farrington and Martin, 1988; Amanor, 1989), also sometimes charac-
terized as "farmer-first," and Participatory Rural Appraisal (PRA) (RRA
Notes, passim; Mascarenhas, 1992; Chambers, in press). These both overlap
with, and have antecedents in, applied anthropology, Farming Systems

Vol. 4, No. 1, 1993


Research (FSR), and Rapid Rural Appraisal (RRA). PRA also owes much to
agroecosystem analysis (Conway 1985). Both FPR and PRA are evolving and
spreading methods for analysis by farmers. FPR methods have tended to be
more verbal, with observations and discussions tending to take place over
longer periods of time. PRA methods have tended to be more visual, with a
tendency to concentrate on a shorter time period. Some contrasts between
these approaches are summarized in Table 3.

Table 3: Contrasts Between Dominant Modes in FPR and PRA.

Scope Agriculture only Natural resources, health,
community planning,
agriculture, poverty, programs, etc.
Main activities On-farm research Appraisal and diagnosis
and trails
Mode of interaction More verbal More visual
Analysis often through... Dialogue Diagraming
Assessments often using... Absolute measurement Relative comparisons

PRA methods have been used in an FPR context (e.g. Pretty, 1990; IIED
and FARM Africa, 1991; Guijt and Pretty, 1992), so it makes no sense to strain
at contrasts, but FPRand PRA could share more. The menu of methods that
follows draws on both, but more on PRA; it invites comments and additions
from other practitioners.


A new literature review to follow that of Kojo Amanor (1989) would elicit
much more than can be presented here. In PRA, and perhaps also in FPR,
there is currently an explosion of innovation, much of which passes unreport-
ed or exists in the greyest of grey literature. What follows has the limitations
of a personal review of some evidence and experience. Let me request
corrections and further information from readers. Most of these methods have
been developed to enable rural people to present and share information with
"us," rather than to enhance their own analysis. Only gradually has the extent
to which these methods may provide farmers with analytical tools been
recognized. Even when the outsider's objective is to extract information,
people themselves learn in the process. However, the potential for farmers to
use these methods entirely on their own is only just beginning to be explored.
For purposes of presentation, approaches and methods can be described as
falling into two clusters, one more verbal and the other more visual. In the
more verbal mode analysis and communication are through discussion, often

Journal for Farming Systems Research-Extension


supported by observation and demonstration. In the more visual mode
analysis and communication are also based on participatory diagraming. A
third cluster of approaches and methods will not be considered here. This
approach enables farmers better to test, experiment, and evaluate, whether
through on-farm research or through improvements to their own experimen-
tal practices (e.g., Bunch, 1989). This includes farmer-designed and farmer-
managed on-farm research.

Verbal Analysis: Discussing and Observing
This cluster of approaches and methods is mentioned briefly for the sake of
completeness and because of complementarities with visual diagraming, but
is not considered in detail since it is the better known approach. Farmers'
analysis through discussion and observation can be facilitated in several ways.
First, farmers' groups have long been used in agricultural extension, and
have generated a substantial literature. In recent years there have been
innovations in group management and gains in understanding (e.g., Knip-
scheer and Suradisastra, 1986; Norman, et al., 1989; Ashby et al., 1989;
Ashby, 1990). Groups present both problems and opportunities that are well
discussed in these sources. One special form of group is the innovator
workshop (Abedin and Haque 1989) in which farmers who have been
innovating come together, share experiences, and visit one another' farms.
Analysis by groups rather than by individuals has also been a trend in PRA.
Second, sequences have been used in discussion and analysis. One example
is crop biographies, where a long discussion covers the history of a crop in an
area (Box, 1989). Another is a more general sequence as used by World
Neighbors in West Africa, and described by Peter Gubbels (1988) as gener-
ating "an informative and analytical debate," starting by comparing past with
present agriculture, and proceeding to ask about problems and difficulties,
what has been tried to overcome them, and what the experience has been with
these solutions. The power and utility of verbal participatory analyses such as
this is not in question, especially when combined with experimentation and
observation. Recent developments, however, suggest that they can be en-
hanced and accelerated through a repertoire of participatory diagraming.

Visual Analysis: Making and Discussing Diagrams
Participatory visual analysis, making and discussing diagrams, has devel-
oped rapidly in the past three years. Diagrams, it used to be thought, were for
"us," or for "us" to use to explain things to "them." A recent review (Bradley,
1992) has found that almost all the literature concerns "our" communication
with "them," not their communication with us or their analysis among
themselves. Sets of participatory visual methods are outlined below. Others
not covered here are walking transects and observation (Mascarenhas, 1991),
well-being and wealth ranking (Guijt, 1992), and other methods that have not

Vol. 4, No. 1, 1993


yet been fully written up. These include Judith Appleton's technique for
women's daily time-use analysis, and Venn and "chapati" diagraming for
institutional analysis.
Participatory mapping and modeling. Participatory mapping and modeling
(Mascarenhas and Kumar, 1991) are perhaps the most widespread methods in
PRA. In mapping, rural people, in groups or individually, make one or more
maps on the ground or on paper. There are three main types: social maps,
which typically show settlements, households, and people (with seeds, or with
symbols on cards); resource maps, which show natural resources and (locally
defined) ecological zones, often those of a village or community area; and
farmhand homegarden maps, in which a farmer draws her or his farm or home
garden. In modeling, three-dimensional models are made of an area, using a
mixture of materials, in India often including rangoli (colored) powders.
Modeling has been used for participatory watershed and natural resource
management planning especially in India, Nepal, and Vietnam.
Participatory mapping and modeling have increasingly been used to
support and strengthen farmers' own analysis. With models, farmers plan by
studying and discussing models, and then mark proposals for action directly
on them. Models are made to show past, present, and future conditions. These
can then be compared, contributing to awareness, analysis, and action. NGOs
in India such as the Aga Khan Rural Support Programme-India (AKRSP) in
Gujarat and MYRADA in Xarnataka have facilitated maps of farms and fields
drawn by farmers that are then retained in the villages and used for planning
and monitoring. Farmers near Chotila in Gujarat have made a complicated
map that they retain and update ofaquifers and wells below a percolation tank.
Farmers have drawn maps of their farms and recorded yields in bushels shown
by numbers of seeds stuck onto the paper. Progress with soil and water
conservation has been recorded on village participatory maps.
Analysis of aerial photographs. This has deliberately been placed below
farmers' mapping because it could be too attractive (for "us") as an alternative
to the participatory mapping in which farmers are presenting their reality more
freely. Nevertheless, the evidence to date is that farmers in different parts of
the world have little difficulty, and take much pleasure, in interpreting aerial
photographs. They have been used in a participatory mode in countries as
diverse as Burkina Faso, Ethiopia, Kenya, Nepal, New Zealand, Papua New
Guinea, the USA, and Zimbabwe. In Papua New Guinea they have been used
for the mapping of clan boundaries and as a focus for discussion and resolution
of land right questions (Mearns, 1989). In Ethiopia aerial photography has
been used for participatory planning of land use allocations (Sandford, 1989).
1:5,000 is a good scale for farm-level detail. Sandford (1989:18) found that
settler farmers in Ethiopia:
immediately recognized that this was a photograph of their land...could without
difficulty indicate the boundaries of their land on the mosaic (correcting in the

Journal for Farming Systems Research-Extension


process some errors made by the consultants)...had no difficulty in recognizing
features such as ponds, swamps, woods, their own huts, thrashing floors, tracks,
areas under crop etc...could take one to any spot on their land shown to them on
the mosaic...could identify on the mosaic their position at any point of a walk
round the land.
Recently, experiments in India have found that villagers can be competent
in interpreting remote sensing false-colour imagery (Sam Joseph, personal
communication), but this has not yet to my knowledge been used in a truly
participatory and empowering mode, and may be limited by scale and lack of
Matrix scoring and ranking. Matrix scoring and ranking is a versatile cluster
of methods applicable in many fields. A family of comparable items are chosen.
In agriculture these have included types of crop, varieties of a crop, types of
animal, varieties of an animal, vegetables, fruits, fodder plants and sources,
trees, soil types, and individual fields. Criteria for assessing the items are
identified through straight discussion and listing, through pairwise compar-
isons, or through asking what is good and what is bad about each. The
resulting matrix is drawn on the ground or on paper. In scoring, seeds or other
counters give values in each box; in ranking, the items are ranked ordinally.
Matrix scoring is proving usually more powerful and popular than ranking.
Scoring is usually either by free scoring, or by scoring out of a fixed number
like 10, 5, or 3.
Two applications of matrix scoring and ranking deserve mention. The first
is to enable farmers to decide on their preferred portfolio. In May 1990,
Shyam Lal Arya of Nagargaon village near Nainital in India matrix ranked
seven horticultural trees for four criteria, then indicated the number of each
sort of tree he had, and then decided how many he would now like of each
(Table 4). A similar method is used by AKRSP in social forestry; women and
men in separate groups matrix score about a dozen trees each, using seeds on
the ground, and then allocate 100 small stones to the different trees to show
the proportions they would like in the nursery. The men's and women's
proportions are then reconciled in discussion. As in these examples, making
the matrix sharpens awareness and informs and improves decision-making.
The second application of matrix scoring may have wide application in
agriculture, enabling farmers to analyze varieties and then communicate their
priorities to scientists. The first farmer to do this was G.B. Ramana, Chikaban-
ihatti Village, Jagalur Taluk, Chitradurgar District, Karnataka, India, with
matrix scoring for locally known varieties of sorghum. He identified ten
criteria (e.g., Table 5). He then used tamarind seeds to score the varieties
against the criteria. When this was completed he was given 50 more seeds and
asked to score the characteristics of a "wish" variety that he would like in
addition to those varieties already available. The method has since been
repeated in Nlaphkhane and Mapoka Villages near Francistown, Botswana. A

Vol. 4, No. 1, 1993


male farmer in consultation with women matrix-scored varieties of sorghum.
A senior scientist who jointly facilitated the analysis was surprised and
impressed, and said that the farmer's preferences could be taken account of in
the breeding program. A group doing the same for trees changed their scores
four times before agreeing. In both cases the matrix scoring was taken with
seriousness and deliberation.
It remains to be seen how widespread this, and similar methods, will
become. They appear not only to enhance farmers' analysis, to provoke
revealing debate, and to provide an agenda for discussion, but also to provide
an accessible means for farmers to communicate their priorities to extension-
ists and scientists.
Farmers'flow, linkage, and causal diagraming. Perhaps the most striking
development has been linkage diagrams. These are diagrams drawn on the
ground or on paper that show flows, causal relationships, or other connec-
tions. Such diagrams, drawn by "us," are part of the toolbox ofagroecosystem
analysis from which so much participatory diagraming comes. But it has been
only gradually that we have come to realize the capacity that rural people have
not just to understand, but to make and use linkage diagrams. A varied and
impressive collection of such diagrams is to be found in the IIED report
Participatory Rural Appraisal for Farmer Participatory Research in Punjab,
Pakistan (Guijt and Pretty, 1992).
The classic case of participatory causal systems diagraming is the analysis of

Table 4. Matrix Ranking and Choice of Horticultural Trees.
Apple Khum- Peach Pear Plum Lemon Malta

resistance 7 3 6 2 4 1 5
value 1 2 3 5 4 7 6
resistance 6 3 7 2 5 1 4
Use of
fodder 3 1 2 -
Number of
trees at present 50 15 50 50 20 30 5
Number of
trees desired 0 40 0 65 45 70 6
Preferences for
new plants 2 3 4 1 5

Journal for Farming Systems Research-Extension

Analyst: Shyram Lal Arya
Facilitators: Alok Kumar and Sukhbir S. Saudhu; May, 1990.


Table 5. Matrix Scoring of Five Varieties of Sorghum Plus a New "Wish" Variety.
CH 59 Kesam Awalturi MHS 51 CHS-5 "Wish"

Taste 2 5 3 9 6 6
Yield 4 3 2 9 6 6
Good price 4 3 3 8 5 6
Fodder 3 5 5 3 5 5
Disease resistance 4 5 5 3 3 5
Ratooning 4 8 2 4
Many food uses 3 6 5 3 4
Good for health 4 5 2 5 3 4
Low input cost 2 3 4 3 2 3
Analyst: G.B. Ramama, Chikabanihatti Village, Jagalur Taluk,
Chitradurgar District, Karnataka, India.

farmers' problems in Eastern Visayas in the Philippines, in which farmers were
helped to develop and analyze systems diagrams for factors affecting the
central problem of the cogon weed (Imperata cylindrica) (Lightfoot et al.,
1989). Diagraming was used to indicate the relative importance of different
factors, and to focus discussion on alternative actions. The participatory
diagraming of bioresource flows has been pioneered by farmers and the
International Center for Living Aquatic Resources Management (ICLARM)
in countries as diverse as Bangladesh, India, Malawi, and Vietnam (Lightfoot,
1990) and is the subject of an ICLARM video Pictorial modelling: a farmer-
participatory method for modelling bioresource flows in farming systems,
starring farmers in Malawi (Lightfoot, Noble, and Morales, 1991).
Participatory causal diagraming can also show impacts of recent changes.
In March 1991, Savasi Bhura, a farmer at Gadechi village, Surendranagar
District, Gujarat, drew an impact diagram in a matter of about 25 minutes.
This showed the effects of an irrigation scheme that had come to his village
three years earlier. It included physical, biological, and social effects, and other
important effects such as the scheme's impact on school enrollments. Linkage
diagraming of farming systems (Lightfoot et al., 1991) also appears to have
potential. The ability shown by farmers, whether literate or illiterate, to draw
such diagrams has been remarkable (e.g., Guijt and Pretty, 1992).
In sum, forms of participatory linkage diagraming known to date include:
bioresource flow modeling, especially for aquaculture; causal diagraming;
impact diagraming; and farm profile and systems diagrams, including internal
and external farm linkages.
Seasonal analysis. Seasonal diagraming has become a standard method in
PRA. Analysts are invited to distinguish seasons or, more often, months, and

Vol. 4, No. 1, 1993