Farming systems research

FARMING SYSTEMS RESEARCH

Prepared for:

PPC/CDIE
U.S. Agency for International Development
Washington, DC 20523

Prepared by:

Karen Wiese
Associates in Rural Development, Inc.
72 Hungerford Terrace
Burlington, VT 05401
(802) 658-3890

Date: December 14, 1984

AD,
Associates in Rural Development, Inc.
362 Main Street
Burlington, VT 05401 USA
Phone: 802-658-3890 Telex: 957032

December 21, 1984

Dr. J. Kenneth McDermott
2008 Powhattan St.
Falls Church, VA 22043

Dear Dr. McDermott,

I am writing in response to our telephone conversation of this morning.
Enclosed please find my draft report to PPC/CDIE on Farming Systems Research.
As I said the reason I would like you to review the paper is because I
relied heavily on your concepts to develop the FSR conceptual framework.
PPC has plans to publish the final draft as a formal discussion paper in
their evaluation series. For this reason what is said in the document
needs to be fairly firm and professionally acceptable because it will "live
forever" on the shelves in AID. Please feel free to do a content edit
from a technical viewpoint.

As we discussed, I would appreciate it if you would send me a signed AID
biodata sheet and a bill for the day you spend reviewing the paper. This
will enable us to pay you and bill AID under our work order.

I highly appreciate your assistance in this regard and look forward to
receiving your comments by mid January as there will be a meeting in Washington
on the paper at that time.

Good luck with your move to Florida.

Sincerely,

K ren iese
AAgricultural Economist

Telephone: (904) 392-1965
Cable: CENTROP FSSP
Farming Systems Support Project

INTERNATIONAL PROGRAMS
Institute of Food and Agricultural Sciences
University of Florida
3028 McCarty Hall
Gainesville, Florida 32611
Jan. 14, 1985

Ms. Karen Weise
ARD
362 Main Street
Burlington, VT 05401

Dear Karen:

Here are some notes I made as I read your paper.

1. While I think the paper is good, it gives a visibility
and emphasis to FSR which leaves me a little unsettled. You deal
fairly extensively with institutionalization of FSR.
Institutionalization of FSR separate from research and extension
seems to me to have little relevance. FSR/E should be placed in
context to fit the TIProcess and discussion needs to focus on
R/E.

In context it has tremendous potential utility. Per se it
will likely have roughly the same utility as the other disjointed
segments.

One new utility of FSR/E in context is now emerging in some
of our work. Most LDC's, especially in Africa will have to
depend heavily on imported technology from the international
technology network. That is feasible, and they don't have the
capacity to generate innovations themselves on any significant
scale. The ITN is pretty good and steadily getting better. It
can be depended on to supply a fairly good choice of technology.

What it cannot be depended on for is testing, adaptation,
etc. And, if the countries do not have that capacity they can
take much less advantage of the ITN.

Also in ITN/LDC relations, currently, the initiative is
largely with the ITN entities. LDC takes little initiative in
telling the ITN what it needs. It doesn't know and won't know
until (a) it comes to know it's farmers and (b) it starts
trying to help them solve their problems.

2. Re CSR vs FSR. People are kidding themselves when they
talk holistic. There is no way to take a holistic approach and
really emphasize it.

EQUAL EMPLOYMENT OPPORTUNITY/AFFIRMATIVE.ACTION EMPLOYER

CIMMYT talks of interactions which, if properly identified,
are the relevant parts of the whole. Dealing with interactions
is not particularly difficult.

So we are always going to be partial. The extent to which we
move more nearly to the whole will be a function of national
resources availability. They are extremely scarce now, and that
means we will be "extremely partial." The multi-disciplinary
aspect is also constrained by resources. You made this point in
the MD team being an economist and an agronomist (or a biological
researcher). Resources will keep us to CSR for a long time.

Further, it did not take genius for Syrian on-farm
researchers to learn that barley producers were interested in
barley forage as well as barley grain. So CSR does involve
livestock. CSR very quickly discovered that availability of
animal power was a prime determinant of crop production.

The major problem of livestock is feed and feed comes from
crops, chiefly crops residue.

There is another angle. FSR as an approach (not a subject
matter field OFR/FSP not NFSD) did not start with crops. Crop
researchers developed and came to use the approach. The real
question may not be "when is FSR going to address livestock" but
rather "when are livestock researchers going to embrace FSR?"
They have been so invisible because livestock research itself is
virtually invisible in most LDC's.

This brings us back to resource constraints.

You are correct. CSR vs FSR is a non-issue.

3. There is still quite a residue of NFSD but not in Africa.
I don't think it has much potential OFR/FSP will lead to "new
systems" but incrementally. Adoption of systems is not in the
cards. Farmers disaggregate technology packages.

4. I doubt that FSR projects have the well defined
characteristics you stated with Zambia, which passed your test,
it has 4 conventional researchers on its project vs 2 FSR types.
I think it is okay.

5. Extension has not been integrated. FSR people with
limited exceptions don't understand extension. Hildebrand and
Winklemann don't really believe it is necessary.

No one that I can identify (of FSR types), even understands
the extension specialist concept.

6. I think you credited Hildebrand with inventing the
"recommendation domain" concept. I think it was done by the
CIMMYT group.

Incidentally, to indicate that FSR is not as "new" as it is
thought to be in the U.S. we had "types of farming" areas back
in the 20's.

7. In one piece a statement could be interpreted that "site
selection" and "identification of recommendation domains" were
the same. That isn't accurate. Some people virtually equate
diagnosis and recommendation domain identification.

8. Shaner's 5th step, extension, needs explanation. We
asked Shaner to synthesize experience in FSR. The FSR gurus had
no experience in extension. So Shaner's "extension" step was an
extrapolation.

9. We discussed the NFSD.

10. The IARC's are an extremely valuable source of
alternative technologies that national systems can draw upon. I
expect that they would respond to problems for which they did not
have ready technology. They could aid and abet national systems
in problem identification. I don't know what they can do in FSR
per se I think CIMMYT has a good program. I'm not impressed
with IRRI. IITA is just now getting relevant. CIAT aborted its
FSR at the embryonic stage. ILCA may have something going.

11. Evaluation has us all perplexed. FSR types make
significant claims for the approach but want to take shelter
behind the long term nature of conventional research. While a
long time is needed for institutionalization, I've about
convinced myself that short run impacts should be expected of
FSR.

The Gene Ross methodology has alot of FSR characteristics.
On-farm trials to be sure technology works and then energetic
promotion. If the promotion included researcher participation
and technology performance feed back it would not violate FSR
concepts.

I'm ready to be pressed for earlier results I think. This
may be a switch from earlier observations but it's not entirely
inconsistent.

FSR does not start with technology generation. It can start
further downstream.

A University of Minnesota project in Chile with Ford support
had good results.

12. Economic feasability is almost a non-issue. There is so
much variation on both sides (costs and benefits) you've hit most
of them.

It is very clear that FSR can be very expensive if one is not
careful. But if we don't close the gap in the TIP what?

It seems to me that our job is to develop R/E systems,
practices, procedures; methods, that are economic. We know the
potential value of technology innovation.

Sustainability is another issue. Bathrick and I found the
economic returns to investment in ICTA were impressive. Yet
Guatemala was "starving" ICTA. The starving was completely
disassociated from ICTA performance.

One reason may have been that ICTA had not told its own
story. ICTA itself didn't know its contribution until Dave and I
did the arithmetic. This is a major item in institutionalization
for all of R/E, including FSR.

So far project design rarely addresses the management
responsibility of acquiring resources. I don't know that any
does.

These are notes that may help strengthen your paper. I'm too
lazy to comment on the good points of which there are many. I
hope it does some good. It may be though that people want the
arguments, the controversies and do not really want them settled.

Thanks for admitting my participation in the task.

Regards,

Sinceely,

Ken McDermott

LAB 005:6

I '*

PREFACE

Since the late 1970s, the U.S. Agency for International
Development (AID) has allocated a substantial amount of
development assistance resources in support of Farming Systems
Research (FSR), in addition to its substantial core support for
international agricultural research centers conducting a variety
of FSR programs. As of FY84, AID records show support of over 40
FSR projects under implementation in each of its geographic
regions of operation. The majority of these are now midstream in
implementation terms and are due for evaluation.

PPC/CDIE summarizes, synthesizes and disseminates lessons
learned from development experience for the purpose of assisting
improved design and management of projects. In this effort,
PPC/CDIE focuses on activities that represent major current and
probable future areas of program concentration and develops
guidance for evaluation that will help pertinent and useful
lessons to surface.

In September 1984, Associates in Rural Development, Inc.
(ARD), was asked to address the problem of a comprehensive
approach to evaluating Farming Systems Research projects. To
obtain the most useful information from the upcoming FSR project
evaluations, CDIE has requested ARD to develop an initial
analytical framework which is comprehensive enough to support the
following requirements:

o that the evaluations of these projects are appropriately
comprehensive in addressing the range of issues pertinent
to the holistic nature of the research itself;

o that some basic or "core" questions are asked in all
evaluations, so that PPC is better able to summarize and
synthesize evaluation findings in ways which are most
helpful to design and implementation; and

o that evaluations address the issue of macro-implications
and use of information derived from micro-level research.

In order to assist CDIE in this effort, ARD has been asked
to provide a classification or typology of FSR projects,
according to their characteristics, purposes and operations, with
an initial identification of core issues for evaluation, and a
refinement of both through their application to a selected sample
of AID-supported FSR projects.

The activities undertaken by ARD in pursuit of these
objectives centered on:

o a review of relevant literature written by agricultural
development professionals who are or have been
substantively involved with FSR at the field level;

o unstructured personal interviews with these same
field-level practitioners, as available to ARD in the
United States;

o structured personal interviews with agricultural
development professionals working directly for AID and
other U.S. government agencies directly involved in
design, implementation, management and evaluation of FSR
projects in developing nations; and

o structured review of AID Africa Bureau Project Papers and
project implementation evaluations, as available to ARD.

This report is organized around the task objectives--not
activities--as set forth above. Section 1 provides an Executive
Summary for those who do not wish to become embroiled in the
details of conceptual and issues development and presentation.
Section 2 develops the conceptual framework and background for
thinking about FSR in the larger scheme of agricultural research,
technology development and dissemination. We then suggest a
useful classification scheme for FSR projects and programs,
developed according to objectives, characteristics and activities
used by each "class" of project/program. Section 3 provides an
identification of potential issues for evaluation, which both
"cut across" the FSR portfolio and are appropriate at the
project-specific level. Section 4 presents the results of an
application of the classification scheme to FSR projects in the
AID Africa portfolio, as well as an identification of issues of
potential relevance to those projects.

LIST OF USEFUL ACRONYMS
AID U.S. Agency for International Development

ASM Agricultural Sector Modeling

CGIAR Consultative Group f6r International Research

CIMMYT International Maize and Wheat Improvement Center (Mexico)

CSR Cropping Systems Research

FSR Farming Systems Research

IARC International Agricultural Centers

ICRISAT International Crops Research Institute for the Semi-Arid
Tropics (India)

ICTA Agricultural Science and Technology Institute (Guatemala)

IITA International Institute for Tropical Agriculture (Nigeria)

ILRAD International Laboratory for Research on Animal Diseases
(Kenya)

IRRI International Rice Research Institute (Philippines)

R&E Research and Extension

TAC Technical Assistance Committee (of CGIAR)

T&V Training and Visit System (of Extension)

\ Isko TAC(_ Tech ica A s-aJ AComtmui erJL ( f C A-

1.0 EXECUTIVE SUMMARY

There has been much confusion surrounding Farming Systems

Research (FSR), primarily because it is a new research strategy

which does not have a well-established lexicon or set of

procedures and methodologies. As a result of this confusion,

many projects lay claim to the FSR label, in good faith and

otherwise, in order to secure funding from donor agencies

currently pursuing this strategy to increase the effectiveness of

developing-country research and extension systems.

This discussion paper is written in response to a need to

clarify what is known about FSR and, in light of what is known,

to examine AID Africa FSR project designs with a view toward

surfacing issues which might be relevant to future project

evaluation.

Farming Systems Research is a research "process" or
"approach" which, in the main, takes the objective of the conduct

of on-farm research for the purpose of testing, adapting,

integrating and disseminating new technologies for adoption by

farmers. Thus, FSR, in combination with strong commodity

research and extension programs, offers an increased potential

for improving the productivity of the whole farm system and,

thereby, improving the welfare of the adopting farm family and

the nation, as measured primarily by production and income

indices. FSR has emerged in response to an increasing

recognition that the existing national agricultural research and

extension systems are functioning ineffectively, in terms of

appropriate technology development and dissemination. This

ineffectiveness is due to a lack of understanding of the

production systems in which new technologies are intended to

perform and, thus, a failure to evaluate the new technologies'

performance by the criteria of those systems and the farmer

decision-maker.

All FSR projects must embody certain characteristics and

activities in order to meet the functional requirements and

organizational objectives which they were designed to address.

The core characteristics are:

o farmer participation in the research process;

o multidisciplinary research planning, design, conduct

and evaluation;

o holistic approach to research problem identification/

prioritization and technology evaluation;

o problem-solving research orientation; and

o reduction of natural heterogeneity of farming systems to

homogeneous groups of farming systems for both research

and extension purposes.

The core activities are target and research area selection,

problem diagnosis, research design and planning, technology

testing and evaluation, and extension. The organizational

requirements for FSR implementation demand full cooperation and

collaboration among commodity researchers, adaptive on-farm

researchers and extension specialists, working with the farmers'

objectives firmly in mind.

The overall objective for FSR project performance is to

integrate multidisciplinary on-farm research into the existing

research and extension system in such a way that technologies are

developed and extended to farmers in response to their perceived

needs and are compatible with their resources.

FSR is more easily conceived of than implemented primarily

because of very fundamental attitudinal and behavioral changes

which must take place before FSR can fully meet expectations set

for it. In view of this, in addition to a generally expressed

concern to maximize the use of scarce research resources to

capture efficiency gains in project implementation, several

important issues have emerged for Agency consideration. The

importance of these issues has been reinforced as a result of a

review of AID Africa FSR project design. They are:

o Given the developmental state of the art in FSR, should

we be considering institutionalization of this relatively

unproven strategy into existing national agricultural

research and extension organizations?

o Given the generally weak and ineffective nature of

developing-country, publicly supported extension

organizations, what is the role of FSR in the technology

dissemination process?

o Given the fact that the majority of FSR projects have

chosen to concentrate on cropping subsystems or

commodities to the exclusion of other important on- and

off-farm enterprises, except as those interact with

specified crops under consideration, what are we giving

up at the goal level in terms of opportunities for

benefit impact?

o Given a reluctance to include the exogenous environment

(e.g., critical support services, infrastructure,

policies, and other institutions) as variable or subject

to modification through the FSR effort, how does this

affect the opportunities presented as a result of the FSR

process?

o Are Agency expectations for FSR project performance

reasonably aligned with the resources available, the time

frame for implementation and the magnitude of the task?

2.0 CONCEPTUAL FRAMEWORK

2.1 The Role of FSR in National Agricultural

Technology Innovation

Farming Systems Research (FSR) is a research "process" or
"approach" which, in the main, takes the objective of the conduct

of on-farm research for the purpose of testing, adapting,

integrating and disseminating new technologies for adoption by

farmers. Thus, FSR, in combination with strong commodity

research and extension programs, offers an increased potential

for improving the productivity of the whole farm system and,

thereby, improving the welfare of the adopting farm family and

the nation, as measured primarily by production and income

indices. FSR has emerged in response to an increasing

recognition that the existing national agricultural research and

extension systems are functioning ineffectively, in terms of

appropriate technology development and dissemination. This

ineffectiveness is due to a lack of understanding of the

production systems in which new technologies are intended to

perform and, thus, a failure to evaluate the new technologies'

performance by the criteria of those systems and the farmer

decision-maker.

With the following discussion, we hope to elaborate on the

reason for and the nature of the functional and organizational

gaps whichsexist in most developing-country agricultural research

and extension systems and the role of FSR in bridging the gap.

Figures 1 and 2 present a visualization of the functions and

organizational requirements of effective national research and

extension systems. In both figures, Area A represents the

existing gap in national systems and, thus, defines Farming

Systems Research.

In order to discuss Farming Systems Research in any

meaningful way, it is essential to recognize that it is (or

should be) an integral part of the larger national research,

development and diffusion process. Therefore, as a prerequisite

to a full discussion of Farming Systems Research, it is

appropriate to briefly discuss the entire process. For this

brief discussion, we will rely heavily on concepts and

definitions elaborated by McDermott.

Figure 1 presents what McDermott conceptualizes as the

technology innovation process in agriculture. He notes that the

visualization is a vast oversimplification of a rather complex

process in that it shows the process as a simple chronology with

one discrete function following completion of the previous one.

In reality, these functions flow together and become so

thoroughly fused that they cannot be distinguished in practice

(#5, 1984). The subsequent explanatory notes will be
useful to the reader to ensure a common lexicon.

Figure 1.

The Technoloqv Innovation Process in Aariculture

Area A

Technology
Diffusion
R RH technology Technology Technology Technology Technology f n
RESEARCH and
generation Testing Adaptation Integration Dissemination Adoption
Adoption

SSCIENCE Technology Development -

ResearchlOrganization -
Extension Organization

Adapted from J. K. McDermott, "Technology of Technology Innovation," ed. George Beal et al, East-West Center,
Honolulu, from Knowledge Generation Exchange and Utilization, forthcoming.

__

Research is commonly associated with science in the search

for new knowledge. In this sense, research is an abstract

analytical process which seeks to control all independent

variables so that new knowledge can be gained regarding the

dependent variable.

Technology Generation is a creative function wherein

something new is conceptualized, designed and put together.

Technology is usually generated on an experiment station. The

reductionist approach (the "scientific method") is used and,

thus, performance can be hypothesized. It is important to note

that not all technology generation is created within any given

national system. There is heavy borrowing from other national

and international research systems, e.g., the world stock of

knowledge.

Technology Testing is the testing of a proposed, promising

technology for performance. Often the first test (screen) will

be on the experiment station. If we keep in mind that a

technology will be of value or use only in relation to a

production system, such initial experiment station screening is

not enough. The new technology must be tested in the system in

which it is expected to perform and against the performance

criteria of that system. If the public system does not perform

the second-phase testing, then either the process stops or some

other entity will pick it up.

Technology Adaptation is the function of adapting a

technology, which has been tested (and works) in one specific

production system, often with minor modifications, so that it

performs to standard in other relatively similar systems. This

is a very important function from an economic point of view in

that to bring costs and benefits in line with management

expectations and criteria, it is necessary to enlarge the range

of sites in which a technology will serve. The need to adjust

technology to a range of sites requires extension specialists to

be active participants in the technology development process.

Technology Integration is a function with multiple aspects.

Technology itself is an integration of performance character-

istics with other technologies into a single new technology,

which must then be integrated with other technologies in the

production system. Sometimes a technology is embodied in a

commodity which can only be adopted into the production system if

larger systems which serve the production system change

accordingly, e.g., market distribution systems, etc. It is

important to realize that even in the "earlier" stages of

technology generation, integration problems can be anticipated,

will be important in designing and generating the new technology,

and will provide some of the criteria for testing and adaptation.

Technology Dissemination is an extension function which

requires a conscious, active promotion campaign for the new

technology. Although "dissemination" is often used inter-

changeably with "diffusion," there is a serious distinction to be

made. One of the most effective dissemination techniques is the

extension demonstration. Extension field agents and, in some

cases (where existing and functioning), other state, parastatal

and private-sector agribusiness entities do most of the actual

dissemination. It is necessary, therefore, that these

dissemination agents be familiar with the new technology. When a

national extension service is doing the dissemination, there must

be extension specialists, who have participated in the technology

testing, adaptation and integration phases, working with the

field agents and supervising the farmer-agent demonstrations.

Technology Diffusion is a dynamic social process which

occurs in every community through farmer membership in various

social systems: family, neighborhood, church, civic

organizations and cliques. It is a natural process, the dynamism

of which emphasizes the importance of working with farmers

throughout the technology innovation process.

With the foregoing, we have conceptualized the functions of

the technology innovation process in agriculture. As one further

step in this conceptualization, we must make the system

operational by overlaying the national research and extension

organizational structuress. By doing so, we are able to clearly

match the functions of the technology innovation process with

organizational requirements. Figure 2 presents this overlay.

Note that we have added farmer participation to this visualiza-

tion to indicate which functions are valid for their involvement.

Figure-2.

DivisioLn -Qf=a ticoIns between Research and Exten.sin
by Function of the Technology Innovation Process

Technology Innovation Process
Area A

Adapted from J. K. McDermott, "Farming Systems Support Project, VITA-CAR Team Notes," 10/19/84.

The reader should not interpret the equal area allocated to
each of the functions in the technology innovation process as
being an equal allocation of effort.

All four sub-entities of the national research and extension
organizations) can be involved toward the beginning of the
adaptation function and the end of the testing function without
violating the mandate of any of the units.

Finally, in order to understand the emergence of Farming

Systems Research and its role relative to the various functions

of the national technology innovation process, we believe it is

necessary to briefly examine the efforts of the United States'

bilateral aid and scientific communities to strengthen the

indigenous capacity for technology innovation in developing

countries. Since the 1950s, the United States has been involved

in national institution-building for agricultural research and

extension. Institution-building in this context involved human

resource development, the provision of physical facilities and

equipment, administrative and management development, and the

provision of U.S. scientific counterpart personnel to work in

cooperation with host-country researchers (USAID, #1, 1982,

p. 22).

Our successful domestic experience with the U.S. Land Grant

research, extension and education system formed the basis for our

institution-building efforts. However, according to McDermott,

we made a few errors in conceptualizing the land grant model and

its role in the national technology innovation process for its

transfer to developing nations. The essence of the problem is in

the mistaken translation of "research" and "extension" as the

functions of the technology innovation process instead of labels

for administrative organizations (1982 & #6). Thus, we

transferred administrative organizations which did not have the

mandate to perform all the necessary functions of technology

innovation and dissemination within the U.S. context and would

not in the developing-country context. This error was compounded

by the lack of explicit recognition of the integrative functions

which mediating entities (e.g., the so-called agribusiness

complex in the United States) fulfill in the technology

innovation process (Winkleman & Moscardi, 1979). And finally, as

has been pointed out by many participants in this institution-

building process, we did not properly appreciate the experimental

nature of farmers and the extent to which innovation and

experimentation occur within the dynamic social systems of

farming communities in developing countries as well as in the

United States.

We can now see the results of these "conceptual short-

comings" in the operation of the national research and extension

organizations) which developing-country governments built with

our assistance. There are national agricultural research

capacities in most developing countries which have varying

capabilities (determined by many factors outside the scope of

this paper) in conducting applied research using the scientific

method of inquiry, with a conspicuous absence of a farmer-client

orientation. Similarly, there are national agricultural

extension capacities in most developing countries which have

varying capabilities for introducing innovation to their farmer

clientele--determined in part by the limited applicability of

most experiment-station recommendations available for extension.

The resulting ineffectiveness of developing-country agricultural

research and extension organizations is of critical importance

because research, technology development and extension derive

purpose and meaning from their service to the nation in the

achievement of national development goals (Dillon, 1976).

In summary, then, it has been recognized that there are

conceptual and operational gaps in the agricultural technology

innovation process in most developing countries. In functional

terms, it is the inability, at present, to develop technologies

for the production systems in which they are intended to perform

and to evaluate the performance of those technologies against the

criteria of those systems.

It is now possible to identify, conceptually, what Farming

Systems Research is and what functions and organizational

requirements will be necessary to bridge the existing gap in the

developing-country agricultural technology innovation process.

Referencing both Figures 1 and 2, FSR can be conceptualized as

that area (A) of function and organizational operation between

the dashed lines. Generally stated, FSR is comprised of the

technology innovation functions of technology testing,

adaptation, integration and dissemination (to a varying extent).

In operational terms, FSR requires the participation of

disciplinary and commodity research teams (in Figure 2, national

subject-matter research programs), multidisciplinary field

research teams, extension specialists, extension field agents and

farmers, working in cooperation toward common objectives. The

only relevant measure of objective achievement is the adoption of

new, improved technologies by target farmers, which increase the

productivity of the whole farm system, and consequently improve

the welfare of the adopting farm family, and the welfare of the

nation, as measured by production and income indices.

2.2 FSR Objective. Characteristics and Activities

With the foregoing, we have presented a functional and

organizational explanation of Farming Systems Research, as well

as a rationale for its emergence as an integral part of a

national agricultural research, technology development and

dissemination process for developing countries. Now we intend to

more fully discuss Farming Systems Research "in the main" by

identifying the objective, a set of core characteristics and

activities which set FSR apart from more conventional

agricultural research. From this common base, we will then

identify three major "types" of FSR projects and programs

currently under implementation in developing countries. This

"classification" or typologyy" of FSR projects and programs is

based mainly on the differences in project and program purposes,

activities and major characteristics, which are clearly

distinguishable among the three types of FSR.

FSR: Objective

All Farming Systems Research projects and programs share a

common objective, which is to raise the efficiency of the whole

farm system (Norman, 1980). There is a distinction to be made

between the objective of FSR and more traditional discipline- or

commodity-oriented research. The FSR task is to identify

potential technologies which will not only increase the

productivity of a particular commodity/enterprise or even

subsystem, but will also not lower (and, it is hoped, will raise)

the productivity of all other subsystems within the whole farm

system. Conventional research is not typically concerned with

the performance of enterprises or subsystems other than the

subsystem under study. The whole farm system orientation places

greater demands for information on the research team than one of

lesser scope, which has implications for the types of activities

which must be undertaken and the disciplines which need to be

involved.

FSR: Core Characteristics

Farming Systems Research, regardless of the particular

approach followed, must have certain characteristics embodied in

the research organization and process for FSR to be distinguished

from the so-called traditional disciplinary- or commodity-

oriented research. Reference to this common set of

characteristics can be found throughout the literature on FSR

(TAC, 1978; Shaner et al, 1982; Norman, 1980; Collinson, 1982;

Gilbert et al, 1980; Harrington, 1980; Kampen, 1979). The

characteristics are:

o farmer participation in the research process;

o multidisciplinary research planning, design, conduct

and evaluation;

o holistic approach to research problem identification/

prioritization and technology evaluation;

o problem-solving research orientation; and

o reduction of natural heterogeneity of farming systems

to homogeneous groups of farming systems for both

research and extension purposes.

Probably one of the most important characteristics of FSR is

the insistence that farmers, for whom research is being

conducted, must be involved in the research process. There are

many events and ideas which have contributed to this

reorientation in research procedures. Popular opinion has held

that small farmers were a homogeneous group of inefficient

producers, bound in ignorance by the forces of tradition.

However, over the past 15 years, research results from all areas

of the developing world have shown that:

o small farmers are economically rational in their

decision-making regarding resource allocation (USAID,

#2, 1976);

o their objective is not necessarily profit maximization

(Labadan et al, 1980; Collinson, 1982);

o risk and uncertainty are dominant forces in their

decision-making (Shultz, 1953);

o the nature of these forces are primarily determined by

the economic and natural (bio-physical) circumstances of

the individual farmer (Winkleman & Moscardi, 1979); and

o the economic and natural circumstances which dictate

the type of farming system are widely diverse, even

within relatively small geographic areas (Hildebrand,

1979).

Essentially, research scientists have begun to recognize that if

they are to be successful in developing technologies for adoption

by small farmers, they must first come to understand those

farming systems, and then evaluate the performance of new

technologies against the criteria of those systems. According to

Collinson, "relevance in research demands that local farmers

provide their own perspectives (to researchers) in planning

experiments and their own decision criteria in evaluating

results" (1982, p. 3). Specifically, the farmer must be involved

in the problem-identification phase to the extent that he gives

the researchers a sufficient understanding of the whole farm

system, the problems he has in increasing the production and

productivity of any particular subsystem(s), and an indication of

the priorities for problem solving. Additionally, the farmer

must be involved in evaluation of new technologies designed by

researchers to solve his production problems. Such involvement

provides the researchers with information regarding the

appropriateness of the technology and its acceptability to the

farmer. Also, in cases where problems arise concerning

appropriateness and/or acceptability, the farmer provides

essential feedback. This feedback can be important in not only

improving the "fit" of a particular technology to a particular

system, but also in guiding future experimentation and

dissemination efforts.

Given the expressed need for improved understanding of

small-farmer decision-making, in terms of the conditioning of

their perceptions of risk and uncertainty by economic and natural

(bio-physical) circumstances in which decisions are made, the

importance of multidisciplinary research teams in FSR cannot be

overstressed. At an absolute minimum, multidisciplinary research

in the FSR context demands the collaboration of a research

agronomist and an agricultural economist in all of the activities

of the FSR cycle (e.g., site selection, diagnosis, design,

experimentation and testing). Norman has characterized the

underlying determinants of any farming system as technical and

human, with the human element including both exogenous and

endogenous factors. He claims the technical and exogenous human

factors are beyond the control of the individual farmer.

Therefore, the farmer will make decisions regarding resource

allocation within his farm system according to his perceptions of

the constraints which those technical and exogenous human factors

impose on him (1980). Whereas it is the task of a physical or

biological scientist to modify the technical environment to

provide the potential for improvement of the production system

operating in that environment, FSR practitioners claim that the

technical perspective is not enough. Collinson states that the

most important contribution of the agricultural economist is in

providing an understanding of how and why, in managing his farm,

the farmer will compromise on the optimal technical management of

any one enterprise to raise the productivity of the whole system

(1982). Thus, a collaborative, multidisciplinary research effort

can provide the combination of perspectives and abilities to

understand both the technical and the human determinants of any

given farming system, and thereby provide the possibility of

modifying constraints, resulting in the improvement of the whole

farm system.

As stated, FSR recognizes the importance of the underlying

technical and human determinants of any particular farm system as

constraints to increased efficiency of that system. Therefore,

FSR focuses on the interrelationships and interdependencies

between the technical and human elements and, as such, is more

holistic than conventional agricultural research (TAC, 1978;

Norman, 1980). Upon this much, we find general agreement in the

literature. However, it appears that there is some confusion and

disagreement as to interpretation of the word "holistic" to infer

a guiding principle and general perspective (Byerlee et al, 1982;

Simmonds, 1983) or to infer the application of whole farm

analysis and similar-systems analysis concepts (Labaden et al,

1980). In general, we believe the majority of FSR practitioners

use the term "holistic" to mean bringing a broad understanding

(obtained from both primary and secondary data) and appreciation

of technical and human elements which constrain the various on-

and off-farm production enterprises to the problem

identification, research prioritization and technology evaluation

activities of the FSR process.

FSR must have a problem-solving orientation. As we pointed

out earlier in this section, FSR is not "scientific" or "basic"

research. Rather, FSR involves technology testing, adaptation

and integration (with a varying role in dissemination). The

important point here is that FSR is not undertaken in order to

generate knowledge for the sake of knowledge. Knowledge which

is generated through FSR is applied to production problems of a

specific farming system to raise the overall efficiency of that

system, as perceived by the farm family, according to their

goals.

Once established that the economic and natural

(bio-physical) environments in which a farmer must operate are

the most influential forces in his decision-making regarding

technologies to be used in production, it is important to

recognize that there is certain heterogeneity in small farm

agriculture. Therefore, technologies must be tested, adapted

and integrated into production systems of small farmers who may

produce the same commodities but have different preferences and

operate in different market situations (Bremer, 1983). One of

the first FSR activities is, therefore, to reduce the "universe"

of heterogeneous farming systems to a set of relatively

homogeneous subgroups. The disaggregation is done first in

terms of agro-ecological systems or differences in the technical

element, then if further disaggregation is necessary, in terms

of differences in the human element (Norman, 1980, p. 8). With

this, the research process can proceed with each group in turn,

or simultaneously, depending on development priorities of the

national decision-makers and the research resources available.

One such relatively homogeneous group of farming systems is

referred to as a "recommendation domain" (Hildebrand, 1976) or

"domain of adaptation" (Zandstra et al, 1981). The terminology

implies that there is a clearly bounded area, not necessarily

physically contiguous on a map, which is defined according to

economic and natural characteristics, and for which new

technologies with a proven "fit" may be applied by all farmers in

that area.

FSR: Research Activities and Methodologies

Given the relatively recent emergence of Farming Systems

Research, the literature abounds with methodological discussion.

Much of what has been published is in the vein of documenting

the development of activities and methodologies within the

context of country or IARC project and/or program implementation

(Norman et al, 1982; Collinson, 1982; Hildebrand, 1979). Other

authors have gone a step further to make the attempt to publish a

particular standardized guide to FSR activities and

methodologies, based on a particular country or IARC experience

or collection of experiences (Zandstra et al, 1981). To our

knowledge, only one group has attempted to present general

standardized guidelines for the conduct of FSR (Shaner et al,

1982). This discussion of the activities and methodologies of

FSR in the main will be general and superficial, but sufficient

for the purpose at hand.

Norman states that "there are four successive research

stages: descriptive (diagnostic), design, testing and extension"

(1980, p. 6). Hildebrand's experience in Guatemala and the work

of CIMMYT and IRRI suggest a fifth stage, which actually comes

first, involving site selection or identification of

recommendation domain (Hildebrand, 1979; Winkleman & Moscardi,

1979; Zandstra et al, 1981).

Figure 3 presents an illustration of the sequence of the

five stages of the FSR cycle (Shaner et al, 1982). There are

several things which should be noted regarding the figure.

Experiment station and extension collaboration which interacts

with the first four and five stages, respectively, is set off to

the sides to emphasize their supporting roles. It is also

important to note the iterative nature of the activities. In the

figure, they are shown to be sequential and discrete. However,

in practice, certain activities may overlap and be repeated or

skipped altogether. Information feedback is an important element

of FSR. Finally, farmer involvement is not shown on the diagram.

Although there is some disagreement about farmer involvement in

each of the stages, it is generally recognized that farmers will

need to be involved in diagnosis, testing and extension.

Figure 3. The Five Basic Activities of On-Farm Research in FSR&D

From Shaner et al, Farming Systems Research and Development, 1982.

Each of these research stages is composed of a specific set
of activities and each activity can be conducted by a range of
methodologies, which are not agreed upon and some claim are
situation specific. For ease of presentation, Table 1 shows each
of the stages and common activities.

Table 1. Stages and Activities of the FSR Cycle

FSR Stage

Activities

1. Target and Research
Area Selection (Site
Selection)

2. Problem Identifi-
cation and Development
of Research Base
(Diagnosis)

3. Planning On-Farm
Research (Design)

4. On-Farm Research
and Analysis (Testing)

5. Extension of
Results (Extension)

Secondary data collection
Primary data collection
Data analysis
Decision

First preliminary analysis
Secondary data collection
Second preliminary analysis
Primary data collection (formal and
informal methods)
Analysis of data
Problem identification
Prioritization of well-defined
problems
Feedback of ill-defined problems for
additional data collection and
analysis

Appraisal of organizational capabilities
and resources
Appraisal of relevant shelf-technology
(local, regional, national, inter-
national sources)
Setting assumptions
Setting research priorities
Developing hypotheses for testing
Establishing research collaboration
Ex ante analyses of proposed
experiments
Considering alternative research
activities
Setting final plan for experiments

On-Farm Research

Researcher manager trials
Superimposed trials
Farmer manager trials

Results Analysis

Biological performance
Net benefit projection
Economic
Financial Feasibility
Sociocultural
* Acceptability
* Long-term versus short-term

Field demonstration plots
Multi-locational testing
Pilot production programs

Two things should be noted regarding this table:

1. activities are those used primarily in conduct of crop
and cropping systems on-farm research; livestock, perennials
and off-farm enterprises will require additional or different
activities and perhaps a different sequence.

2. No specific references are made to on-station work which
will occur in Stage 4, if necessary.

2.3 Classification Scheme for FSR

The exact origin of Farming Systems Research has not been

possible to trace. We do know that FSR was incorporated into

the charter mandates of most of the International Agricultural

Research Centers when formed in the early 1970s by the CGIAR.

Also at that time, various development organizations and donors

were supporting the development of FSR in specific country

programs such as in Guatemala (ICTA), Mexico (Puebla Project)

and Colombia (Caqueza Project). However, to our knowledge, there

is no literature which documents the development of the FSR

"movement." What we do know is that what emerged in the thinking

of many international agricultural research organizations and

development donors as the need to incorporate agricultural

economics, ex ante, into the process of planning and evaluating

agronomic research through on-farm experimentation has blossomed

and undergone a process of mutation since those early days.

Currently, there appear to be three major "types" of FSR

projects and programs under implementation in developing

countries. Norman Simmonds, in his 1983 report to the World

Bank, provides labels for these three types of FSR. There is an

alternative FSR classification suggested in the literature

(i.e., upstream, downstream) (Norman, 1980; Gilbert et al, 1980;

Harrington, 1981; TAC, 1976). However, we believe that upstream

applied research and downstream FSR are complementary functional

parts of the same process and thus not separate classes of FSR.

Thus, in this regard, we will rely on Simmonds' classification of

FSR and encourage the reader to examine that report for more

details. The three types are:

o Farming Systems Research/Sensu Stricto,

o On-Farm Research/Farming Systems Perspective, and

o New Farming Systems Development.

Farming Systems Research/Sensu Stricto is academic research

on farming systems as they exist; their description, analysis,

classification and understanding. In practice, there can be

varying degrees of depth, but typically such research goes deeply

into the agriculture, economics and social context of the system

under study. This type of research is good for the purposes of

academia, but much less so for the purposes of agricultural

technology development and dissemination. Further, this type of

FSR cannot meet the functional and organizational requirements we

posited as necessary to bridge the hypothesized gap in the

national technology innovation process. With regard to increas-

ing whole farm system productivity, FSR Sensu Stricto could only

meet a portion of the requirements for objective achievement. In

terms of FSR core characteristics, FSR Sensu Stricto does not

embrace a problem-solving orientation, although it could display

the others. Finally, in terms of activities, FSR Sensu Stricto

will probably never get beyond the diagnostic phase. For all

these reasons, although we accept the fact that FSR Sensu Stricto

is being practiced, we reject it as an acceptable approach to

meeting developing-country research and extension needs.

On-Farm Research with a Farming Systems Perspective

(OFR/FSP) begins with collecting just enough information (in the

FSR Sensu Stricto mode) on the farming system to define the

research necessary for stepwise modification and improvement of

that farming system. Most of the research conducted in this

style will be on-farm, in recognition that technologies intended

to change the performance of a particular system need to be

adapted to the circumstances of their farmer users, and that

experiment station results can by no means always predict the

experience with a technology at the farm level. In practice,

OFR/FSP meets all of the requirements of FSR, but is perhaps less

holistic than some would like.

New Farming Systems Development (NFSD) begins with some

degree of information on a particular farming system, will use

some degree of on-farm research and is oriented toward the

complete replacement of the existing system with radically new

systems. NFSD shares the general FSR objective and can display

all the core characteristics. The activities may also be

similar to OFR/FSP but will vary as to relative emphasis and

methodology. Systems simulation and on-station research are

typically the most important activities. We readily recognize

that there may be circumstances which will dictate (by reason of

environmental degradation, population pressure, etc.) such

radical change. However, New Farming Systems Development will

require enormous investment of undoubtedly scarce resources and

total political commitment in order to succeed. For this reason

work in NFSD will probably be highly site dependent and not

undertaken as a rural development option too often. Finally, we

do not believe that NFSD will be able to bridge the R&E

functional and organizational gap hypothesized at the beginning

of this section.

Figure 4 is a box diagram-type visual presentation of these

three types of FSR, their objectives, major activities and

characteristics. There are no horizontal connecting lines,

indicating there is no connection among these three types of

research as they have nothing in common apart from a focus on

farming systems as the subject of the research and data

collection/analysis activities, though with the latter, there is

a significant difference in degree and purpose.

Figure 4. Farming Systems Research Classification

Activities:

* Data Collection
* Data Analysis
* On-Station Experiment
* On-Farm Experiment
* Technology Evaluation
under Farmer Experience
* Some Level of Technology
Dissemination

Major Characteristic

Technology Testing
and Adaptation

New Farm Systems
Development

Objective: Replace
Existing Farming System
with Radically Different
System

Activities:

* Data Collection
* Data Analysis
* Some On-Farm Experimentation
* Majority On-Station
Experimentation
* Prototype Development

Major Characteristic

Systems Simulation
and On-Station Testing

We would make one comment regarding the terminology chosen

to identify the second class of FSR. There are almost as many

labels for this type of FSR as practitioners and/or authors on

the subject. For example, Byerlee prefers "On-Farm Research with

a Farming Systems Perspective;" Collinson prefers "On-Farm

Adaptive Research;" Hildebrand prefers "Farming Systems Research

and Extension." The chosen terminology is that of Byerlee et al,

because it stresses the importance of on-farm research relative

to other research activities conducted in the FSR cycle while

maintaining the whole farm research orientation. Each of these

labels has its own emphasis. However, at this level of

classification the differences implied by these various labels

are not significant. While they may differ slightly, they all

share the same objectives, activities and characteristics. Thus,

this class of FSR is clearly distinguishable from the other two

we have suggested.

3.0 ISSUES FOR FSR PROJECT/PROGRAM DEVELOPMENT

AND IMPLEMENTATION EVALUATION

In the foregoing section we have presented a conceptual

framework for thinking about Farming Systems Research in terms of

its functional role and organizational requirements in the

national agricultural technology innovation process of developing

countries; the major objective for FSR; the outstanding

characteristics which differentiate FSR from other types of

research; and a suggested classification scheme which

distinguishes among different FSR approaches currently under

implementation around the world. With this conceptual framework

serving as "common ground," and focusing the discussion on

OFR/FSP, this section identifies eight major issues for

consideration by AID in future FSR project and program

development and implementation evaluation. These major issues

represent broad areas of concern, expressed by development

professionals in the FSR literature which we consulted, and

confirmed by personal interviews with FSR practitioners and

AID/Washington personnel with varying FSR program management

responsibilities. As presented, these issues in the broad

context represent concerns for project and program development or

design. The discussion of the issues surfaces areas of concern

with implications for implementation evaluation. Briefly stated,

these issues are:

1. Should FSR be institutionalized into existing national

agricultural research and extension services?

2. What is the role of the International Agricultural

Research Centers (IARCs) in FSR development at the

national level?

3. What is the role of FSR in the national technology

extension process?

4. What is the role of FSR in providing linkages between

national research and extension organizations and other

national agricultural planning, development and service

organizations?

5. Should FSR take more than a field crop orientation in

setting the research agenda?

6. How should AID consider the question of FSR cost in

relation to concepts of benefits, effectiveness and

efficiency?

7. In the design of FSR projects are time and objective

properly aligned?

8. Are there circumstances when FSR should not be considered

as an AID program option in national agricultural

development strategy implementation?

1. Should FSR be institutionalized into existing national

agricultural research and extension systems?

Much has been written recently (Bremer, 1983; and in Flora

#8: Mbalu, 1984; Baker, 1984; Hodgens, 1984; Freseco, 1984) and

there are an increasing number of voices within AID calling for

the institutionalization of FSR into national agricultural

research and extension systems. It seems that there are several

important questions which arise in connection with the issue of

institutionalization. There are several different approaches to

FSR, each of which has objectives, characteristics, activities

and methodologies unique unto itself. Further, within each of

these major approaches there are seemingly an infinite variety of

activities and methodologies employed to achieve common

objectives. Under this situation, what exactly is it that should

be transferred and institutionalized? Secondly, decisions to

"institutionalize" are not typically taken in the "to do" or "not

to do" framework. Rather, these decisions are made in a

framework which usually involves consideration of the extent to

which "x" can assist an organization in achieving its mandated

objectives in comparison with alternatives. In the case of FSR,

it is difficult to arrive at an answer to that question, even

after a country has had some project experience with FSR.

Illustrative of this is the experience of the University of

Hawaii:

"... senior decision makers wanted to see demonstrated

results, that could be measured in terms of increased

production and income generation potential, before

committing themselves to major policy and organizational

changes necessary for the institutionalization of

FSR&D..." (Harold McArthur, in Flora, #8)

It is possible to put the question differently, in a deductive

logic, "if-then" framework as follows:

If it is true that in the majority of developing countries

there is an effective gap in the national agricultural technology

innovation process (as suggested in section 2), in terms of

necessary functions which are not being performed and a lack of

organizational collaboration required for performance of these

necessary functions in order to develop appropriate technologies

for the improvement of local farming systems, and if FSR can

bridge this gap by providing for the required organizational

collaboration and performance of necessary functions resulting in

improved appropriate technologies, then FSR should be

institutionalized. However, the problem in arriving at the

conclusion is in verification of the second "if" clause. Has FSR

proven itself capable of bridging this gap? We will not attempt

to provide an answer, but suggest that it is critical in terms of

deciding whether or not to institutionalize FSR.

Another question in relation to the issue of institutionali-

zation of FSR is whether there are alternatives. In many

countries the national (publicly supported) capacity for

agricultural research is weak and ineffective and/or nonexistent.

This means that they do not have the human, physical and

financial resources to effectively conduct applied research

(technology generation). In this situation, to add the

additional functions of technology testing, adaptation, integra-

tion and dissemination is not realistic. Is it possible to look

to other agents of change within the society as having the

capability and potential willingness to perform these functions?

Simmonds points out that in nearly every society there are four

agents of agricultural technological change:

o farmers;

o public RD&E systems;

o government-supported, financial and commercial agencies;

o commercial entities.

He notes that all four are capable of active, adaptive,

on-farm research activities which can result in technological

change (1983). When we speak of institutionalization of FSR

within the context of AID projects and programs we typically mean

trying to create the conditions which will enable the publicly

supported agricultural research and extension (e.g., Ministries

of Agriculture and Extension) to perform these functions.

Perhaps providing support for other change agents to perform the

FSR functions is an idea worthy of consideration in cases where

the existing institutional base is weak and not well integrated

into society.

In certain country situations where the government has

demonstrated its intentions to re-orient its national

agricultural research and extension system to incorporate FSR,

institutionalization may be a proper consideration by AID for

support. It should be kept in mind that institutionalization is

not a "one-shot process" and does not occur in a relatively short

time frame. Many actions, many of which involve substantial

government policy and procedural changes, must take place.

Additional financial resources will be needed to pay for

additional personnel, equipment, facilities and logistical

support. Experience with developing-country national budget

allocation procedures tells us this is not a rapid or certain

process. Probably most important will be the recruitment and

training of nationals from both research and extension

organizations in the methods of FSR. Given AID's concern for the

recurrent cost element of proposed projects and programs, a

careful analysis will be required.

Finally, in those situations where the host-country

government is ready, willing and capable of taking the necessary

steps required to institutionalize FSR into the existing research

and extension system, what management and administrative

reorganization will be required? The answer to this question is,

of course, country specific. However, there are several

considerations which center on the need to integrate and fuse FSR

throughout the administrative and management structure, rather

than simply appending another organizational unit onto the

existing structure. The need for organizational integration

derives from McDermott's observation that the functions of the

national technology innovation process are not sequential,

discrete actions, but in practice are so thoroughly fused and

interdependent that they are difficult to distinguish (#5, 1984).

This observation speaks to the need for thorough collaboration of

the various entities (program administration and management at

all levels, disciplinary and commodity specialists working in

support of adaptive research and extension specialists (and vice

versa) and extension specialists working in support of extension

field agents (and vice versa)).

Another aspect of this need for integration of FSR into the

existing institutional structure is security. If this

integration does not occur and FSR is simply an isolated

appendage on the existing structure, there can be no security in

terms of decisions regarding timely budget allocation, personnel

assignment and logistical support requirements. Further, in

times of budget austerity the probabilities are quite high that

an isolated FSR unit would be the "first to go." Hildebrand's

experience in Guatemala during the early years of ICTA are

testimonial to these kinds of integration problems.

"One of the most difficult budget problems involved

SER staff located in the regions. In 1977, the Perito

Agronomos were transferred to the regions and the

budget that was deleted from the Central SER group was

supposed to be added to the regional budgets. This,

however, was never accomplished so that we were put in

the position of begging from other regional programs.

Funds were already tight and the need to share with SER

did little to create goodwill...In 1977 and again in

1978, attempts were made and orders given to the Regional

Directors to create specific budgets for SER regional

personnel, but that was never accomplished. The result

was that the Regional Directors, seeing that the SER

personnel were budgeted through their regional Technology

Testing Teams, wanted to use them as if they were their

own staff...It is evident, that budgeting procedures can

have important positive and negative effects on attempts

at integration of the social sciences with the biological

sciences in an agricultural institute." (1979)

In summary, we believe that a thorough institutional and

organizational analysis should be conducted of the publicly

supported agricultural research and extension system in a

developing country before serious consideration is given to FSR

project development. If the assessment indicates a functional

gap which could be closed by institutionalization of OFR/FSP and

which could cause the existing infrastructure to be more effec-

tive in technology development and dissemination, and if there

are no acceptable alternatives to publicly supported OFR/FSP,

then institutionalization is probably a reasonable option.

2. What is the role of the International Agricultural Research

Centers (IARCs) in FSR development at the national level?

According to the 1978 special report of the Technical

Assistance Committee (TAC) of the Consultative Group for

International Agricultural Research, there is no conflict between

the operational mandate of the IARCs and their development of FSR

programs. In fact the TAC reporting team recommended that "all

commodity/regionally oriented IARCs, except ILRAD, should have a

clearly recognized program in, or orientation to, FSR" (TAC,

1978, p. 60). They recognized that most FSR in the IARCs has

been concerned with the cropping subsystem (with a specific

commodity or resource focus), rather than the whole farm system

and that this emphasis should be continued. The review team also

recommended that the IARCs limit their FSR activities to genera-

tion of technologies and methodologies which can be generalized,

rather than to concern for location-specific practices, whereas

national FSR programs are obliged to deal with location-specific

problems. In fact they stressed the point that development of

FSR methodologies should be the major output of the IARCs.

Methodological development of FSR was directed to encompass base

data analysis, on-farm studies and research station studies.

Finally, the TAC reporting team recognized that training

national-level scientists was the major channel for the trans-

mission of FSR methodology and technology from the IARCs, and

recommended strengthening of the IARC FSR training activities.

Indeed, throughout the literature training emerges as a

major need in national programs which are considering FSR, as

well as a major cooperative linkage between the IARCs and

national programs. CIMMYT's interest in the development of FSR

procedures relates directly to the Center's association with

national programs--as a producer of intermediate goods (elements

of new technology, procedures and training) which national

programs apply in forging new technologies (Winkleman and

Moscardi, 1979). Similarly, a basic premise in the establishment

of ICRISAT was that the Institute would serve to strengthen and

support national research programs through: the provision and

exchange of germplasm; assistance in training; promotion of

scientific interaction; and provision of technical assistance.

"ICRISAT scientists have an important role in providing practical

training to local technicians and, in some cases staff

scientists" (ICRISAT, 1981, p. 13). Collinson, working since the

mid-seventies in East Africa for CIMMYT,:claims that one of his

first objectives was to "teach the farm systems economist the

tools of his trade and to expose him to agronomicc) experimental

methods" (1982). He also made the observation that in-service

and short-term training have preoccupied CIMMYT staff since 1979.

On the basis of the heavy demand for training in the procedures/

methodologies of conducting the diagnostic survey and adaptive

research, CIMMYT will need to expand its training activities and

work closely with FSR programs sponsored by various donors.

Norman, who has also spent a considerable amount of time

conducting FSR in Africa, claims that

"One of the major problems involved in establishing FSR

programs is the lack of agricultural scientists and

social scientists with FSR training or experience...

If FSR programs are to grow and be effective, they must

be staffed by individuals with training and experience in

FSR which is largely unobtainable outside of existing FSR

programs." (Gilbert et al, 1980, p. 77)

The authors go on to say that the nature of FSR, namely its

locational specificity and the need to modify methodologies to

suit local conditions, strongly favors training on location. The

main problem is lack of experience with interdisciplinary

research. They argue against formal degree training because of

the predominant requirement of degree programs for disciplinary

competence and the tendency toward increasing specialization.

However, they also argue that in order to participate fully in

multidisciplinary research, each FSR member must be firmly

grounded in at least one discipline.

Hence, we can see that the nature of FSR, e.g., location-

specific, practical/hands-on and ground-level or real problem-

solving setting, argues strongly for training assistance provided

by various IARCs. It also seems clear that the need is not,

necessarily, for formal academic degree programs in U.S. or

third-country universities. It is noted throughout the

literature that there really isn't any motivation among

university administrators to train and grant degrees in FSR

because disciplinary competence is their collective teaching

objective, and the multidisciplinary "competence" required for

analysis in a whole farm framework goes above and beyond the

application of concepts, principles and tools of a single

discipline. As of 1982, there were only five U.S. universities

having even expressed interest in such training (Shaner et al,

1982). However, a similar issue could be raised with regard to

IARC training programs based on limitations imposed by their farm

subsystem or commodity orientation. Such a limited focus (less

than whole farming system) could lead to limited perspective in

the diagnostic analysis of the technical and human environment

and in the consideration of opportunities. We do not have enough

information on the exact methodologies and procedures employed by

the IARCs, individually, in their FSR programs to judge the

validity of this issue.

This issue, however, of subsystem, commodity or resource

limitation should also be raised with regard to the transfer of

methodologies to national FSR programs. Is there something about

this limitation in perspective which imposes undesirable

constraints in the application of a more holistic or systems

perspective in the diagnostic and evaluation stages of the FSR

cycle? We believe that it would be very worthwhile for AID to

commission several very experienced field practitioners of FSR,

representing various disciplines and extension, to examine these

methodologies more closely.

Another element of FSR project implementation which must be

considered a constraint by AID is the lack of trained and

experienced U.S. FSR practitioners. The need for experienced FSR

technical assistance is another area where the IARCs could

be/have been of valuable support to national program development.

Policies of individual centers will have to be examined as to

whether or not they allow for such temporary transfers of their

scientific personnel. However, such prohibitive policies are not

cast in concrete. Given the recognized importance of FSR and

location-specific, on-farm research to the continuing progress of

the work of the centers (TAC, 1978; Simmonds, 1983; Shaner et al,

1982; ICRISAT, 1981; Zandstra et al, 1981; Morris, 1984), we

think that mutually acceptable cooperative agreements could be

drawn among individual centers, host-government agencies and AID.

There are cases in precedent (CIMMYT) which can be used as

prototype or model agreements.

Finally, we would like to emphasize the IARCs' mandated role

in the generation of prototype technologies or technology

components for transfer to and adoption by national programs in

support of their activities in technology development (Winkleman

and Moscardi, 1979; TAC, 1978; ICRISAT, 1981). This is the

so-called "shelf-technology" issue. Are there technologies which

are ready for transfer to national programs and which are

suitable for applied and adaptive research at the national level?

We know that there are many cases of such transfer having

occurred. However, we are not secure in the knowledge that this

process is systematic, continuous, responsive, reliable and

timely. As Hildebrand has noted from his experiences with

technology development in Guatemala, "The frequently conceived

notion that a pool of technology is available and waiting for

extension techniques to get it into the hands of small farmers is

mostly invalid" (1979, p. 22). We recognize the fact that his

comment was made with specific reference to national experiment

station research results/recommendations, but we think it applies

in the larger sense as well. In fact, this question is of such

magnitude that AID's Bureau of Science and Technology has

notified Congress of its intention to fund an entire project

designed to examine the extent of shelf technology in detail.

The point is simply that the IARCs could go a long way toward

speeding up the process of research planning and design within

national FSR (and disciplinary/commodity programs) if this type

of comprehensive information were available continuously,

reliably and systematically to national-level researchers in

their deliberations and considerations of alternatives and

opportunities.

In summary, then, we believe that the IARCs have a very

positive, supportive role in national-level FSR program

development which could be exploited by AID in the areas of

training, transfer of appropriate FSR methodologies, technical

assistance and technology transfer.

3. What is the role of FSR in the national technology

extension process?

The extent to which FSR should be firmly integrated into

the national technology extension process is without doubt one of

the most troublesome issues concerning FSR development. If the

conceptualization of the national technology innovation process,

as presented in section 2, is valid, then FSR must not only have

forward links into the basic and applied research, or science and

technology generation functions, but strong links into the

function of technology dissemination. However, there are several

problems in creating the research-extension linkage which are

important and demand consideration by host-country governments

and donor organizations promoting FSR.

First, extension organizations are very weak and

ineffective, if not inoperable, in most developing countries.

Second, there are no clearly recognized methods which link the

development of a technology with the requirements for its

appropriate extension to those production systems in which it is

intended to perform. Lastly, there has been little attention

paid to the form of the research output expected and extension

strategies compatible with this output. If the expectations of

developing-country governments and donor agencies for FSR are to

be met (i.e., the development and adoption of appropriate

technologies which result in significant increases in farm

production, productivity, and farm family and societal welfare),

then it seems clear the problems facing extension organizations

need to be dealt with immediately.

Historically, national extension organizations of most

developing countries have been largely abused, both in terms of

mission and budget allocation decisions. This neglect has

resulted in a largely ineffective public bureaucracy of poorly

paid, poorly supported and poorly trained civil servants.

Further, governments have used extension organizations to perform

regulatory enforcement tasks (tax collection, loan repayment,

etc.) because of their immediate contact with large numbers of

the rural population, rather than to perform the tasks of tech-

nology extension. This reputation is clearly recognized by both

of its immediate constituencies--researchers and farmers. There

are many development professionals who recognize this situation

and the constraint it poses to the potential effectiveness of

FSR. One FSR practitioner has written recently, "It is urgent

for extension to modify its historical mission, organization,

staffing and training approach in order to exploit opportunities

created by FSR...If this does not happen it is unlikely that a

change in research, such as a shift to FSR...will be particularly

successful in its long-term impact on agricultural development"

(Johnson and Claar, 1984, in Flora, #8).

On the other hand, there is a group of development

professionals who claim that the on-farm work of FSR, especially

if involvement of large numbers of farmers is built into the

research plan, does the job of extension (Simmonds, 1983). This

claim is based on accumulating research results regarding the

eagerness of farmers to adopt improved technology appropriate to

their production system and the dynamic nature of the diffusion

process in rural communities. This view of the role of FSR in

the technology extension process augers, not for a strengthening

of extension agencies as they exist, but, for a radical

restructuring. The emphasis in terms of extension institution-

building under this view would be on the creation of highly

capable extension specialists with strong disciplinary and

commodity training with clearly identified responsibilities in

all of the activities of the FSR cycle, rather than the

maintenance of large numbers of field agents whose primary job is

to deliver the message from farm gate to farm gate (reference the

Training and Visit System). It is probably not reasonable to

develop national extension capabilities in all developing

countries along the same model--whichever is chosen as most

attractive conceptually. What is clear is that there is an

immediate need to recognize this weakness and to design a

strategy which will address the need for the integration of

research and extension functions.

Much of the existing literature on FSR is concerned with

methodological issues in Farming Systems Research but has been

relatively silent regarding methodological requirements for

extension purposes. However, this subject is receiving

increasing attention, much of which is directed to the need for

ex ante consideration of extension requirements (Chamala and

Keith, 1984; Bremer, 1983; Hart, 1983). The reasons for this

neglect of extension requirements seem to fall into one of two

types of attitudes on the part of FSR program managers: 1) once

a relevant technology is developed extension is comparatively

easy; or 2) we'll worry about it when we get to it (Chamala and

Keith, 1984; Bremer, 1983--respectively). But, as Hart points

out:

"...linking extension and research to form a farming

systems research and extension (FSR/E) process is not an

additive step. Technology transfer activities cannot be

simply added on to the technology generation activities;

they must be combined to form an integrated process. The

reason that this integration has occurred in very few FSR

projects is not just because of a lack of linkages between

research and extension institutions. Technology transfer

is qualitatively very different from technology generation

...FSR is a process that moves from general to particular

(area selection, identification of target farmers, specific

constraints, etc.), while extension moves from the specific

to the general--the goal is to transfer technology to all

farmers where it is appropriate." (1983, p. 1)

Hart claims that additional information is needed, and "a

systematic attempt to generate this information as part of the

technology generation process is the place to start" (1983,

p. 4). Bremer goes one step further in insisting that "we should

regard the form of the extension message as a determinant of the

research design, and not as an issue conceptually separate from

it" (1983, p. 17).

The consequences of not incorporating these specific

information needs of extension staff as ex ante design

considerations of the research process, as reported by Chamala

and Keith of their experiences in Australia, were that only in

one agro-ecological area did the extension staff have clearly

definable systems with a neat package to promote. In other areas

they only had vague principles for application under varying

agro-ecological conditions (1984). According to many FSR

practitioners and observers, much of the FSR technology

evaluation work is concerned with multilocational testing. They

claim this activity not only helps to define the specific

conditions under which technologies can be applied or that

extension agents learn the details of the technologies and their

application, but that these multi-locational testing programs are

part of an extension strategy (Shaner et al, 1982; Zandstra et

al, 1981). Given the clearly polar opinions in this area, we

think that specific attention needs to be given to an examination

of the question of methodological requirements for extension and

the ex ante integration of extension into the research process in

FSR projects and programs.

Finally, with regard to the issue of the role of FSR in

technology extension, we believe it is important to at least

surface a concern discussed by Bremer at length--the form of the

research output in relation to the development of extension

strategies compatible with that output. She observes that the

so-called "tech pack" approach to packaging research output was a

move on the part of the research communities of developing

countries to change the requirements made of an extension field

agent from a generalist required to interpret research results

for specific application to a conduit for delivering a pre-

packaged message, under the assumption that the extension field

agents at large were not prepared for or capable of such

interpretation. However, it has been widely recognized that

"tech packs" are subject to considerable modification resulting

in the possibility of only partial adoption of specific

components of the intended package (Hildebrand, 1979; Harrington,

1980; Kampen, 1979; Collinson, 1982).

Bremer then raises the question of reasonable alternatives

to the tech pack. She notes the increasing popularity in many

countries, primarily resulting from support of the World Bank, of

the Training and Visit (T&V) System. The "advantage" of the T&V

System is that it "requires even less thought on the part of the

extension field agent than the tech pack method. In this system

technologies are divided up into information bits suitable for

delivery during a single visit to the farmer" (1983). However,

the operating-cost requirements of such a system must be

considered prohibitive for most developing countries in a

recurrent-cost framework when judged in comparison with nearly

any other approach. Bremer identifies three basic "packaging"

alternatives (1983):

o modified tech pack -- Recommendations packaged as tech

packs for a specific subset of farmers, defined by a

common farming system, which would have a specified

domain of applicability; the Puebla Project in Mexico

is offered in example.

o technology catalogue -- Recommendations packaged as a

catalogue of technologies, each of which would have been

identified through the FSR process and tested for

agronomic and economic effectiveness under small farm

conditions; the farmer would then be presented with a

range of choices and selection would be based on his

judgment of relevancy.

o modular technologies -- Recommendations packaged in

modules for adding new enterprises to an existing system

or otherwise making relatively major modification to the

farmer's system.

The obvious requirement of each of these is a very capable

extension field agent. The important point being made here is

that the form of the research output and the strategy for

extension needs to be matched before extension staff are expected

to actively disseminate technologies. This requirement has

implications for extension planning at the research design and

testing stages of the FSR cycle.

4. What is the role of FSR in providing linkages between the

national research and extension organizations) and other

agricultural planning, development and service organizations?

The role of FSR in providing linkages between the

agricultural research and extension organizations) and other

organizations which serve the agriculture sector is probably the

most controversial of the issues surrounding FSR. FSR activities

entail the collection and analysis of data on a wide number of

both technical and human (e.g., farm management, social,

cultural, economic, political, and other institutional)

environmental factors which are determinants of the farming

systems extant in any country. The heart of the controversy lies

in the fact that some FSR practitioners view this data collection

and analysis as pertinent only to the tasks of identifying the

nature of environmental constraints and determining opportunities

for modification within the constraints which management faces.

Others believe that if this information were made available to

those in a position to modify the nature of the environmental

constraints (e.g., policymakers, bankers, etc.), then the
"opportunities set" would widen, giving researchers and farmers a

greater probability of achievement of commonly held objectivess.

For the proponents of the former argument, which advocates

the use of information on socioeconomic constraints to management

(farmer) options in the identification of technical opportunities

for innovation, we do not believe they would deny the utility of

their data and analyses to others (non-agricultural research and

extension agencies). Rather, we believe it is a question of

mission, objectives, priorities, and responsibilities. For

Hildebrand, the time frame for the technology development mission

is short--work on socioeconomic constraints which require a

longer time horizon are relegated to the non-solution set of

possible alternative interventions:

"The treatments also should fit within the capabilities of

farmers to carry out in the short run. That is, they should

not depend on a non-existing infrastructure such as a credit

program or the development of new markets, etc. The object

is to produce an improved technology which can be put into

practice immediately by the farmers so they can begin

benefitting..." (1976, p. 7)

For Zandstra, the objective of management and thus the researcher

is to increase the efficiency of a given subsystem of farming

enterprises. He posits that Y=f(M,E) where yield is a function

of management and environment (1981, p. 7). Environmental

variables are considered as variables by researchers only to the

extent that management is to control the environment. To

evaluate the functional relationship the researcher seeks to

determine how to vary the enterprises within the subsystem to get

the best returns for different production environments. Because

the functional relationship covers a wide variety of subsystems,

the researcher must eventually formulate a statement about the

effect of different management practices on the performance of

different combinations of enterprises within a subsystem in

a given environment (Zandstra et al, 1981). Morris, a ranking

agronomist in the IRRI Cropping Systems Program, claims no

agronomic bias in the fact that the FSR process must deal with

questions in order of priority and "where conditions of agronomic

adaptation cannot be met, questions of managerial feasibility and

economic viability are academic. Even if managerial feasibility

is the stumbling block, technical solutions may be more

acceptable to government officials than programs that require

massive budget outlays for credit, subsidies and infrastructure

development" (1984, p. 14). These strongly voiced opinions go

far in reminding us that the clear expectation of everyone for

FSR is technology development, dissemination and adoption by

small farmers. Further, modification of the technical production

environment is making progress toward that objective and within

the purview of researchers.

The other side of this argument acknowledges the importance

of the environmental or exogenous human variables in determining

the type of farming systems extant in any country, as wells

the important potential for improvement in the efficiency of the

whole farm system if those constraining exogenous human

variables could be favorably modified.

Figure 5 presents a schematic framework for Farming Systems

Research which has been adapted from Norman (1980, p. 12). The

adaptation involves the addition of the site/target area

selection stage and the identification of recommendation

domains. It is clear from this schematic that Norman believes

that external institutions are a determinant force in the make-

up of the current farming system. It is also clear that he

believes that information collected and analyzed in the various

FSR activities has both a "feedback" and a "feed-forward"

utility. Specifically, the FSR information/analytical results

must flow to external institutions as well as to commodity-

oriented applied researchers. Linkages with extension services,

delivery system agencies, management of development projects and

policy-making agencies is very important in determining the

effectiveness of existing support systems (External Institutions

in Figure 5) and anticipated changes in the future. "In most

countries micro-level information for policy analysis is

scarce...detailed information generated through the FSR approach

could be very important for identifying changes in policies that

would complement the introduction of improved practices" (Gilbert

et al, 1980).

Figure 5. Schematic Framework for Farming Systems Research

FARMING SYSTEM
RESEARCH STAGES

1. Site/Target Area
Selection

2. Description or
Diagnosis of Present
Farming System

3. Design of Improved
Systems

4. Testing of Improved
Systems

5. Extension of Improved
Farm System

CURRENT FARMING ---
SYSTEM

Identification of
Recommendation
Domains

(hypothesis formulation)

- EXTERNAL
INSTITUTIONS

4'

I

I- ------- BODY OF
Experiment Station Trials ------ KNOWLEDGE
1 i

i I
j i---.--------- ---------------- ----

Trials at Farm Level -----------------

Farmers' Testing --------------------- ------------

S---------------------- :
I <
IF

MODIFIED FARMING SYSTEM -- ------------ --------------
--------------- I

Adapted from D. W. Norman, "The Farming Systems Approach: Relevancy for the Small Farmer," 1980.

Harrington, reporting on a major 1980 CIMMYT-sponsored

conference for FSR social scientists, also recognized these dual

feedback/feed-forward linkages and responsibilities of the FSR

process.

"Policy makers were also noted as potential users of FSR

results. This is, again, a case where the use of research

results is potentially significant, but where linkages are

currently not well developed. FSR results should be of

interest to policy makers because they can often be used to

measure the costs (in terms of foregone agricultural

production) of current policies and to predict the impact of

alternative policies on farmer decision-making. This is not

to say that FSR results can be used to decide on correct

policy. Rather, research results provide necessary (but not

sufficient) information upon which to base such decisions.

Examples of such use of FSR results are, however, rare."

(1980, pp. 6-7)

Gilbert et al note that the objective of traditional

agricultural research has been narrowly focused on improving the

productivity of a single commodity by reducing resources used in

the production of that commodity (1980). Within the specifica-

tion of this problem-solving framework (input use reduction

yields increased productivity, ceteris paribus) the technical

factors are the only factors eligible for treatment as variables.

However, the objective of FSR is:

"with the productivity of the entire farming system and,

as a result, it will examine non-technical changes that

are exogenous--factors such as improving marketing

arrangements for inputs and outputs. The flexibility

inherent in the FSR approach also assists in linking

the micro and macro perspectives in designing strategies

more effective for specific rural areas or groups of

farmers. National policies, such as pricing and trade

policies, that affect agricultural producers may be

explicitly considered when diagnosing existing farming

systems and designing improvements. Changes in such

policies may be the most critical ingredient in efforts

to improve the lives of small farmers." (1980, p. 19)

Collinson (1982), speaking from CIMMYT's perspective on FSR

in East Africa, has stated that some device for coordinating the

efforts of policymakers, planners, researchers and extensionists

is long overdue. He believes that recommendation domains have

the potential of helping select research priorities, promoting

communications between farmers and the bureaucracy and among

departments in the bureaucracy. In addition, the diagnostic

sequence can be used to identify needed changes in policies and

institutions to support local development programs.

In summary, then, it does not appear that there is an

argument regarding whether or not to collect and analyze

information on the "environmental" or "exogenous human" factors

determining the farming system. The argument is really based on

a major methodological difference between consideration of

elements of the non-physical environment as variables with

possibilities for modification or as parameters which must be

assumed as unalterable throughout the technology development

and dissemination phases of the FSR process. Where FSR assumes

the potential for directed change in the environment there must

be effective linkages to "outside" organizations in a position

to effect change.

5. Should FSR take more than a field crop orientation in

setting the research agenda?

Whether or not a commodity-based or cropping system

orientation in the conduct of adaptive research, to the exclusion

of work in the other commodities and subsystems of the existing

farm systems, qualifies as Farming Systems Research is really a

non-issue. Cropping Systems Research (CSR), as this type of

research is referred to, is clearly not FSR. CSR does not share

objective, core characteristics (in total) or the entire range of

activities with FSR. In this regard, CSR must be considered "a

horse of another color." Regardless, there is considerable CSR

being implemented under the "guise" of FSR and there is much

discussion among FSR practitioners regarding the issue.

To begin this discussion, it is necessary to have a common

definitional base.

"Farming systems research addresses itself to each of

the farm's enterprises, and to the interrelationships

among them and between the farm and its environment.

The research uses information about the farm's various

production and consumption systems--the animal production

system, the cropping system, the secondary production

activities such as mat weaving, processed food, etc., that

add value to primary products--and about the farm's

environment--biophysical, institutional, social, economic--

to identify ways to increase the efficiency with which

the farm uses its resources.

"Cropping systems research, on the other hand, is a

subset of farming systems research that is confined to

the farm's crop-production enterprise. The various

crop-production activities are considered modifiable,

taking into account the relationships between the

crop production enterprise and other production-

consumption activities and the physical, biological,

and socioeconomic environment. The objective is to

increase the benefits derived by crop production with

available physical, biological, and socioeconomic

resources." (Zandstra et al, 1981, p. 7)

There are several aspects of Farming Systems Research which

make it more difficult to practice than to conceptualize and

discuss. FSR by definition implies understanding of the whole

farm system before diagnosis and prioritization of research

activities can be accomplished.

"Any systems understanding starts from qualitative

description of components and their interactions, goes

on to quantitative enumeration of states and flows and,

only when the latter are tolerably well defined,

attempts modeling or synthesis." (Simmonds, 1983, p. 24)

In regard to FSR, practicality dictates that a qualitative

description of components and their interactions and some

(limited) degree of quantitative enumeration is as far as the FSR

team can go. The limitations are imposed by reason of time,

personnel and financial considerations which go beyond the

research resource investment capacities of developing countries.

This conclusion can be supported by examining the experience of

AID in its support for Agriculture Sector Modeling (ASM) projects

during the 1970s.

Such detailed understanding goes beyond the level of

sophistication required by practically oriented field adaptive

research teams.

"The grandly 'holistic' approach of FSR Sensu Stricto would

be satisfied by no less than a description of the entire

system. More realistically, and certainly for the purposes

of On-Farm Research/Farming Systems Perspective (OFR/FSP), a

subsystem is selected (depending on the interests of the

investigator) in awareness of the interactions with other

subsystems...In real life, therefore, systems isolated for

study are always subsystems arbitrarily defined for the

purpose in view. They are never 'holistic' in any serious

sense of that rather over-used word. In practice what is

wanted is sufficient understanding to attain the necessary

level of FSP and no more...For OFR/FSP, a partial,

non-holistic, subsystem knowledge will suffice or, anyway,

has to suffice in practice." (Simmonds, 1983, pp. 25-6)

In addition to the element of practicality in limiting the

sophistication of understanding of the whole farm system, there

is a lack of methodologies for integrating non-crop commodities

or subsystems into the FSR process. According to Norman, who is

one of a few who have attempted FSR in the holistic sense of the

Sensu Stricto model,

"Presently there is no standard methodology for undertaking

FSR. Indeed, the term farming systems research is somewhat

of a misnomer. To date most FSR has been confined to crop

production processes. Yet even here methodologies for

undertaking such work need to be improved. Apart from pleas

for desirability (Boer and Welsch, 1977) the FSR approach

has rarely been applied to livestock unless these impinge

directly on crop processes. There is a need to develop a

more holistic systems approach which goes beyond

agricultural and livestock production and includes the

marketing process and off-farm enterprises (Gilbert et al,

1980)." (1980, p. 22)

Simmonds (1983) and Zimmermann (1984) have also noted the

critical importance of perennial subsystems in tropical farming

systems, as well as the lack of analytical methods for

incorporation into the FSR process.

There is one additional consideration relevant to an

understanding of why FSR in theory has become CSR in practice.

The particular operational mandate of the organization conducting

FSR activities dictates the scope of its research focus. All of

the IARCs, except ILRAD, have organizational mandates to work on

specific commodities, in specific resource regimes or regions.

The 1978 CGIAR TAC Stripe Review of FSR not only acknowledged

that the original mandates of the IARCs embodied the FSR

philosophy and included clear scope for the development of FSR

(as several centers have done), but strongly recommended that all

IARCs, except ILRAD, develop clearly identifiable FSR programs or

an orientation of their existing commodity research towards FSR

(TAC, 1978). However, in the same document the TAC review team

limited the scope of the FSR programs to a focus on those

commodities and resource regimes included in their respective

Center mandates. Hence, the FSR programs of the IARCs are

largely CSR, with some program components focused on New Farming

Systems Development (e.g., ICRISAT watershed development

program).

Thus, there are three major reasons for less than the whole

farming system focus in FSR: limitation by reason of

practicality, lack of applicable methodologies, and restrictive

organizational mandate. The existence of sound reasons for FSR

differing in theory and in practice does not, however, negate the

validity of concerns expressed over this issue. Trees, grasses,

animals, fish and off-farm employment figure prominently in the

household-consumption needs and resource-allocation decisions

made by small farmers in the developing world. To the extent

that these production and income-generation activities of the

farm family are not fully incorporated into FSR diagnosis,

research prioritization, agenda and experimentation, then FSR

will never be more than on-farm Cropping Systems Research with a

systems perspective.

6. How should AID consider the question of FSR cost in relation

to concepts of benefit, effectiveness and efficiency?

Consideration of the cost of Farming Systems Research as an

issue for project design and evaluation of AID FSR projects has

received substantial attention in the recent past. These

concerns have been expressed by both field practitioners and AID

personnel. In general, Agency personnel are asking questions

regarding cost in relation to the concepts of benefit and

effectiveness whereas practitioners are primarily concerned with

cost in relation to the concept of efficiency. This division of

concern probably properly reflects the relative project

implementation/management responsibilities of these two parties.

Estimates of the benefit/cost relationship can provide AID

decision-makers with useful information about the anticipated

"investment" impact on the target population and is used ex ante,

in the project design and approval processes. On the other hand,

considerations of cost-effectiveness and efficiency are properly

those of a mid-term or end-of-project performance assessment.

For many years the question of whether benefit/cost analysis

is used appropriately by the Agency and others in the design and

evaluation of agricultural research projects has been the subject

of continuing debate. We do not intend to discuss that debate in

the context of this report. We do wish to raise two points in

relation to the benefit/cost issue in general. First, as we

pointed out, benefit/cost estimation is typically considered a

tool of ex ante analysis for decision-making among alternative

investments proposed to address the same objective. If FSR

programs are complementary to existing R&E programs, there is a

question regarding the validity of estimation of relative

benefit/cost ratios (Gilbert et al, 1980). However, if the

Agency has decided to use the measurement of benefits and costs

to judge the efficacy of its FSR projects (rightly or wrongly),

then immediate attention must be paid to the development of a

methodology for doing so. As Ferguson has noted, "While it is

rarely possible to measure directly the cost and benefit ratio of

research programs, it is possible to evaluate progress against

goals and objectives if these have been clearly stated" (1983,

p. 5). The immediacy of the need for methodological development

derives from the need for specification of information

requirements and collection methods. It is not reasonable to

think that an evaluation team could approach FSR projects which

are 3-5-7-10 years into implementation and find the information

they require to be available unless the information required for

analysis is specified in advance of project implementation, the

methods for collection are standardized, and the requirement for

collection is made an integral and continuing process and a

responsibility of the implementing agency during project

implementation. To date such a standardized methodology does not

exist, therefore uniform collection of information is not

occurring while the FSR portfolio continues to mature.

The second benefit/cost-related issue is one concerning the

time frame for research in terms of impact considerations.

Although there is no consensus on this, development professionals

with significant experience with adaptive research in the

developing-country context claim that 15 to 25 years is not an

unreasonably long period to allow for the natural technology

diffusion and adoption process to occur and for incidence of

benefit to be measurable in terms of indices which get at the

objective (e.g., increased productivity, increased production and

income of adopters, income effects on non-adopters, societal

welfare, etc.).

In terms of implementation evaluation the real issues with

regard to cost considerations are questions of effectiveness and

efficiency (McDermott, #7; Bernsten, 1984, in Flora #8). FSR is

being supported by AID in the hopes that it can help make the

existing system of agricultural research and extension more

effective in generating appropriate technologies which can and

will be adopted by large numbers of farmers, resulting in

increased production and income in the agriculture sector and

increased welfare of society at large. Ferguson notes that it

must be clearly recognized that it is impossible to measure the

effectiveness of agricultural research in a time frame of less

than 10 to 15 years because of the difficulty of sorting out

other factors affecting adoption from the impact of public FSR/D

and extension services (1983, p. 6). McDermott points out that

the issue of effectiveness of research (FSR) leads to an

examination of management and organization systems structures and

processes:

"One (consideration) is that in a donor project to build

indigenous capacity, operational efficiency is not

necessarily the first test. Building capacity (to conduct

FSR) involves learning, both on the part of nationals and

expatriates and donors, and adjustment of both design and

operation based on that learning. Given this, the first

test is effectiveness. Once it is learned how to be

effective, efficiency must be achieved. Based on

effectiveness and efficiency, the program must be expanded

so that it can have a significant impact on the economy.

Craving for economy-wide impact can cause the process to be

short-circuited with the result that none of the three are

achieved." (#7, pp. 2-3)

Bernsten points out that evaluation provides a basis for

assessing effectiveness of a project, the results of which can be

used by policymakers in deciding on future funding, program

administrators in deciding on mid-course corrections and field

staff in planning and implementing the research activities

(1984, in Flora #8). Again, a case can be made for development

of an evaluation methodology. A standardized methodology can and

should be developed for assessing the effectiveness of AID's FSR

projects. Is FSR effective in reorienting national agriculture

research and extension to the task of appropriate technology

development and dissemination? If not, why not? Is management

and/or organization the problem? If so, how? What can be done

to correct these deficiencies, to improve effectiveness?

It is important in assessing the effectiveness of FSR to

examine both organizational structure and management processes.

Because FSR is so new and because AID has such a large number of

FSR projects in similar states of maturation (requiring

evaluation during the same period) we feel that there is a very

important opportunity for the Agency to learn some important

lessons regarding the design and implementation of FSR projects.

If this could be done within a relatively short period of time,

then the Agency could actually use this information to base

future FSR design and funding considerations. There are those

who would agree that such process evaluations are country- and

project-specific and that application of a standardized

methodology for evaluation is not, therefore, appropriate. We do

not share this opinion. The existence of a conceptual framework

for considering the place of FSR in relation to other research

and in relation to extension gives us this opportunity.

Although refinement will be necessary, this conceptual framework

provides functional and organizational requirements for

"operationalization" of a national technology innovation process.

The essential question is, then, has management been effective in

organizing and allocating resources so that each of the necessary

technology innovation functions is being performed in an

integrative and collaborative mode? If not, why not?

One additional observation on the issue of cost-

effectiveness: It is a measurement tool which can give

management/administration an idea of the relative performance of

alternative approaches to achieving the same objective. The

importance of this observation is that in order to use this tool

for evaluation there really needs to be an alternative. In the

case of FSR we believe that there is no clearly recognized

alternative at this time. For Hildebrand, cost-effectivenss is

not the real issue. He has pointed out that the only

"alternative" to FSR which could be considered as an alternative

would be the existing national agricultural research and

extension system, but it is not an alternative (personal

conversation, 1984). FSR is complementary and in addition to the

existing discipline and commodity research programs (Norman,

1980). Thus the costs of integrating FSR into the existing

system are in addition to those of the existing system. Figure 6

presents Hildebrand's conceptualization of these cost

relationships. Hence, the issue is reduced somewhat to a major

concern for the effectiveness of FSR in relation to AID's

objectives for it. FSR effectiveness can be measured in terms of

how well it addresses the functional and operational gap referred

to in section 2. That is, has FSR been effective in serving the

needs of upstream applied research, farmers and policymakers

(Gilbert et al, 1980)?

Figure 6.

Relationship of Costs: FSR in Comparison to Existing R&E

Farming Systems
Research Program

National Agricultural
Research Program

Sunk Cost

Land
Building
Equipment
Laboratories
Other commodities

Recurrent Cost

Salaries
Supplies
Equipment
Communications
Logistical support
Training

New resources required
to initiate FSR program

National Agricultural
Extension Program

Sunk Cost
Land
Building
Equipment
Other commodities

Recurrent Cost
Salaries
Supplies
Equipment
Communications
Logistical support
Training

///// ///////////A\\\\\\\\\\\\\\\\

Additional resources required to
augment the existing capacities of
Research and Extension to incorporate
FSR into existing programs

Adapted from personal communication with Peter Hildebrand, October 10, 1984.

The question of efficiency is critical in consideration of

the costs of FSR. Efficiency should be viewed at two levels,

system and activity/methodology. Proponents of FSR claim that

incorporation of FSR into the research and extension process and

institutionalization of FSR into the existing system will improve

the effectiveness and efficiency of the entire R&E process and

system. This is based on the proposition that such incorporation

provides feedback information to disciplinary and commodity

research which will assist in reorienting those programs toward

real problems identified for applied research effort through the

FSR diagnostic and experiment stages and incorporates the

requirements for technology dissemination in all of the FSR

stages. To the extent that this feedback and incorporation do

not occur, improvements in overall system efficiency cannot be

expected and this becomes an issue for evaluation.

One of the most common criticisms of FSR is that it is

viewed as costly to implement (Gilbert et al, 1980; Harrington,

1980). We,believe that this is not necessarily so. At the

outset, questions of expense should be considered in project

design. What are these cost elements? Typical inputs to

agricultural research projects fall into the categories of

expatriate and national salaries and support, infrastructure

development, support commodities and equipment, operating costs

and training. In comparing FSR with traditional research/

institution-building projects, an examination of budgeted inputs

by line item reveals that FSR does not (should not?) involve

infrastructure development. Main ministerial office buildings,

research station complexes, laboratories, etc., will have all

been provided for in the process of building a national capacity

for applied (traditional) research. There will be some need for

reallocation of space and facilities as new staff are brought on

board--but no new physical construction should be required. It

is true that FSR teams will need to live in the area in which

they work, but local housing should be available.

What about the question of salaries? We believe that this

is where the majority of planned expenditure will be found and is

an area that AID should examine closely in project design and

evaluation. Local salaries, as an item of current operating cost

and planned recurrent cost budgets of host-country institutions,

will need to be examined in terms of actual requirements

for additional personnel. There will be need for administrative

and management personnel, FSR program leaders/research

specialists, extension specialists, and locally recruited

research assistants/technicians. Because of the

special skill training and professional orientation requirements

for personnel in each of the areas, there will probably need to

be new (additional) personnel brought into the organizationss.

However, new personnel should probably be held to the

level required to perform the necessary administrative/management

and field research functions. This will be especially important

at the beginning of the FSR institutionalization process as FSR

is new and unproven and institutional jealousy and territoriality

move into full swing. The issue of expatriate salaries could be

an area for real reduction in planned expenditure. Currently,

AID budgets between $120,000 and $150,000 per year per person for

expatriate salary and support. Given the current five- to seven-

year life-of-project period, and calls for extension beyond that,

these figures will escalate rather rapidly. The role of

expatriate should be examined seriously by AID regarding FSR.

Should AID be providing expatriates as research management

advisers or as field research team members? In either case, do

these expatriates need to be resident over the entire term of the

project or can initial long-term assignments be phased out over

the first two or three years into short-term advisory services in

program and research management?

Training needs have been discussed elsewhere. There seems

to be a consensus among experienced FSR field practitioners that

the bulk of training should be short-term, on-location and

problem-oriented. In contrast to expenditure requirements of

formal degree training in U.S. or third-country universities,

the costs of short-term, in-country training will be much less.

Additionally, the IARCs have been identified by many FSR

practitioners as probably the best source of training assistance.

Given the two-way flow of benefits derived from IARC

participation in national FSR programs, the costs to AID for this

IARC assistance should be negotiable (downward).

With regard to the financial requirements of FSR for support

commodities and other local operating costs, the FSR modus

operandi demands significant expenditure. However, this has

never been a hidden cost element in planning for FSR project

implementation. There are higher costs associated with taking

research off-station and to farmers' fields throughout the

country (Harrington, 1980). About this there should be no doubt.

However, there are presumably higher returns from the

expenditure. It is a fairly well established "theorem" in the

management of development projects that where money has been

obligated it will be expended regardless of the level. Perhaps

in the design phase for FSR projects, complete specification of

supporting commodities could be held to a minimum, with the

contingency line item held, perhaps artificially, higher than

usual. (Contigency is usually estimated as a flat five to 15

percent of the line item subtotal excluding inflation. Perhaps

the contingency line item could include an amount, in excess of

that percentage applied to all other line items, which has been

earmarked for commodities and local logistical support.) Such an

approach would place the burden of judgment on project

administrators and managers, rather than on project design

personnel.

There is another area where gains in efficiency can lead to

reduced project budget requirements. The experience of

technical program management is a critical factor in being able

to capture these potential efficiency gains. The question

centers on the ability of technical managers to plan the

research process according to the minimum amount of effort

required to satisfy the condition of reliability of data and

validity of results (Harrington, 1980), in the shortest amount

of time (Gilbert et al, 1980; Hildebrand, 1976). This issue is

present in all of the stages of the FSR cycle. In the selection

of target areas (recommendation domain identification), how much

secondary data need to be collected, organized and analyzed and

what are the most likely (and ready) sources of this information?

In the diagnostic stage, what level of detail do we need in order

to understand the farming system we are studying? Do we use a

"sondeo" approach or do we go straight into large-scale, formal

surveys? Is the area/population so large that formal surveys

must be conducted using strict statistical sampling procedures?

Is the farming system so complex that repeat visits and farm

record keeping need to be employed in order that we can be sure

our diagnosis of the problems) is reasonable? Do we need to

undertake aerial photography and ground verification, or will a

brief reconnaissance of the area by car suffice? In the design/

planning stage, what is the nature of the "environmental"

constraint set and how many special studies need to be conducted

to understand these constraints in relation to the whole farming

system? How many technical components will be simultaneously

included in the initial research and how many researcher-

managed and superimposed trials? What size plots? How many

treatments and replications? How many check plots? How many

animals? How many farmer cooperators? How many farmer-managed

tests? Is there a need for formal modeling? In the on-farm

research and analysis stage questions primarily center around

analyzing experimental results--which technique will yield a

level of reliability acceptable by professional/disciplinary

standards? What analytical methods are available to integrate

and synthesize findings across enterprises and subsystems? How

can we measure and how much data do we need to collect to measure

farmer acceptability? In the extension stage, questions revolve

around methods of conducting multi-locational tests and pilot

production programs. How many tests, farms, seasons, etc.? How

strict are the boundaries of recommendation domains? Can FSR

recommendations be more widely applied with relatively little

management modification?

Perhaps we have "gone overboard" on this point, but we wish

to stress the importance of this issue. In every FSR decision

taken regarding the actual research activities and methodologies

to be used for data collection, experimentation and analysis,

there will be an underlying decision taken to spend or save

salaries, time, logistical support costs, administrative and

management support costs. The experience of technical management

is critical to cost-efficiency gains in FSR.

One final point should be made regarding the cost-efficiency

of FSR. To the extent that so-called "shelf technologies" exist

in regional, national or international research stations/centers,

and to the extent the FSR team decides (consciously or

subconsciously) to ignore these essentially "ready" technologies

(perhaps needing integration work done) then great strides toward

inefficiency will have been taken.

7. In the design of FSR projects are time and objective

properly aligned?

The crux of this issue is whether or not AID has set its

objectives) for FSR projects at an artificially high level in

relation to the time allowed for objective achievement. Many FSR

practitioners and Agency personnel have voiced concern over this

issue because of the implications for the sustainability of FSR.

Until recently Agency administrators have not been able or

willing to commit themselves to a life-of-project period in

excess of five years. Recently, where strong arguments have

prevailed this period has been extended to seven years. Given

the assumption that the Agency clearly has improved technology

adoption by small farmers as their objective for FSR projects,

during interviews conducted in the course of this study with both

experienced FSR practitioners and Agency personnel, we asked the

question, "Given the experimental nature of FSR, beginning with

Day 1 of project implementation when would you expect to see

results in the form of new technologies designed, tested and

adopted by farmers?" In most cases, the answer was hedged, in

deference to a perceived difference in research requirements for

field crops, small ruminants and large ruminants. Further, no

one was willing to commit him/herself to an inclusion of the

period required for diffusion/adoption--so, the responses cover

only the time required to develop releasable technologies. The

shortest period of time estimated was three years for relatively

simple agronomic modification in field crop production; the

longest period was 25 years for significant improvement in large

ruminant production enterprises.

If we accept McDermott's assertion of the existence of a

learning curve in institution-building projects--FSR projects

must be considered in large part as institution-building--the

relevancy of this issue lies in the correlation between good

project design and accurate assessment of human capabilities and

the potential for institutional evolution. Perhaps we don't have

the methods for assessing, with a reasonable degree of

probability, human capacity for change and the necessary and

sufficient conditions to catalyze institutional and

organizational evolution. If this is true, then it seems once

again we have a need for methodological development. If this is

not true and the analytical methods exist, are they being

uniformly applied in the design of FSR projects? If such

analysis is not being conducted as part of the design process,

then it is reasonable to believe that undue optimism might be

exercised in the process of conceptually and temporily framing an

FSR project. In such a case we would be looking at enthusiastic

end-of-project-status statements such as "500,000 farmers

producing 10 million tons more food grain, creating 8,000 more

jobs in rural areas and reducing the projected trade deficit by

$120 million by the end of Year 5." Even where estimates of

output are more reasonable the fact remains that actual farmer

adoption of new technologies resulting in production and income

increases is AID's objective for FSR project performance

(Ferguson, 1983). With as little as we know about the natural

process of diffusion in rural developing-country communities, the

question remains, in such a relatively short time frame is the

"technology development and farmer adoption and positive upward

macroeconomic change" objective for FSR project performance

reasonable?

Bremer has suggested that sustainability of FSR programs

(during the initial process of institutionalization) requires

more modest program goals (project objectives) (1983). We

suggest this is especially important when goals are set for

institution-building in research and extension with a relatively

new and relatively unproven research strategy. "Any new system

should be implemented in such a manner that it will be given

adequate and necessary support over a period of time sufficient

to demonstrate its advantages" (Waugh, 1984, ch. 5, p. 2). One

senior AID program administrator has suggested that national

research and extension systems are built and supported to satisfy

some of a developing country's domestic economic needs. As such

they (R&E systems) are creatures of a political, social and

economic process which we don't really understand. We don't know

how to build R&E systems that work, but we keep trying. If we

accept this position then more modest, perhaps non-quantitative

project objectives which reflect progress along McDermott's

learning curve should be adopted. As one Agency official stated,

"...no I would not expect to be able to count farmers

using/employing a new technology by the end of a project but I

would expect to see significant institutional change."

The issue is important in terms of the implications of

repeated negative performance evaluation on the sustainability of

FSR. If program goals and project objectives are set at

unrealistically optimistic performance levels, then it will be

the odd program or project which lives up to expectation. This

situation will lead to repeated negative performance evaluations

which do not really reflect project accomplishments fairly. In

turn, the hard-core non-believers will be vindicated in their

belief that FSR is not reasonable research, program

administrators will face increasing difficulty in justifying full

budget/personnel requests, research and extension talent will

work in the FSR mode as long as donor funding is attractive, and

prospects for the continuation and vigor of the FSR program will

decline.

It appears, then, that if this issue is of real concern to

FSR field practitioners and Agency employees, the immediate

options would be to lengthen the time frame for project

implementation or modify project objectives to be more in line

with more accurate assessments of human and institutional

capacities for change in the process of technology innovation.

8. Are there circumstances when FSR should not be considered as

an AID program option in national agricultural development

strategy implementation?

Judging from the recent landslide support which FSR has

received, this issue has probably not received much attention.

However, we believe the question should be raised. Those who

have written on this subject have confined their concerns to one

of three areas: the state of the existing farming system

relative to "equilibrium;" proper identification of smallholders

with the capacity and intention to increase on-farm production;

and the state of the existing national R&E system relative to the

degree of change necessary for FSR project implementation.

On the first point Bremer couches her concern in terms of

the negative implications of inappropriate applications of FSR

on continued methodological development and cost-effectiveness.

"...FSR is not cost-effective or appropriate for all types of

farming systems...This judgment has already been made de

facto by the decision not to apply the FSR approach on

irrigated, monocrop small farm systems, even where such

systems exist side by side with diverse, rainfed multicrop

systems (as in the Philippines)." (1983, p. 20)

Bremer proposes the following hypothesis:

"FSR is most appropriate to small farm situations where

the farming system is in transition from a stable, low-

level equilibrium to a stable, high-level equilibrium.

It is least suited to systems in either type of

equilibrium.

"Farmers in a low-level equilibrium situation have honed

their system to the point where it is by and large as

productive as it can be under their highly constrained

circumstances. Comparatively few opportunities to change

their system for the better exist to be discovered by

agricultural research, until changes take place off-farm

(expansion of marketing systems, for example) to open up

their system.

"Farmers in a high-level equilibrium situation, by contrast,

are more likely to be relatively sophisticated consumers

of extension and researcher advice, quite able to apply

specific recommendations to their individual farming

systems without assistance.

"Farmers in transition (either because of new possibilities

created by technological change or, more commonly, negative

changes in the economy and ecology around them) are less

likely to be able to find their way to a new sustainable

system by themselves. They are therefore more in need of a

holistic approach to assist them in making technological

adjustments in their farming system, since they themselves

may not understand the changes going on around them.

"As a practical matter, more and more small farmers find

themselves, knowingly or unknowingly, in a transitional

situation. Even farmers pursuing a "traditional" system,

such as the slash-and-burn "kaingineros" of the Philippine

highlands or the small farmers of West Africa, are in fact

facing rapidly changing conditions caused by population

growth, new employment opportunities outside agriculture,

and deteriorating environments. Their traditional

technologies are no longer optimal or even viable under

these altered circumstances. These systems should be the

first priority for future FSR programs." (1983, pp. 20-21)

Garrett goes a step further by suggesting that all small

farmers are not alike and therefore need to be identified

according to household objectives, as part of the FSR target

area selection procedures. Essentially, she sees the

smallholder group as being stratified according to orientation

of production enterprise: 1) those who produce primarily for the

market; 2) those who produce primarily for home consumption; and

3) those whose on-farm production enterprises receive residual

resource allocation, after household consumption needs are met

via off-farm enterprise (Garrett, 1984). FSR programs which

include the latter group as a target for agricultural production

improvement will not likely meet with successful objective

MILLISECOND	CLASS.METHOD	MESSAGE
0	sobekcm_page_globals.constructor
0	sobekcm_page_globals.constructor	Application State validated or built
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.constructor	Navigation Object created from URI query string
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.display_item	Retrieving item or group information
0	sobekcm_page_globals.get_entire_collection_hierarchy	Retrieving hierarchy information
0	sobekcm_assistant.get_entire_collection_hierarchy
0	cached_data_manager.retrieve_item_aggregation
0	cached_data_manager.retrieve_item_aggregation	Found item aggregation on local cache
0	item_aggregation_builder.get_item_aggregation	Found 'all' item aggregation in cache
0	system.web.ui.page.page_load (ufdc.page_load)
0	sobekcm_page_globals.constructor.on_page_load
0	html_echo_mainwriter.add_style_references	Adding style references to HTML
0	html_echo_mainwriter.add_text_to_page	Reading the text from the file and echoing back to the output stream
96	html_echo_mainwriter.add_text_to_page	Finished reading and writing the file