Impact Assessment - Annual Report 2001

OUTPUT II: IMPACT OF PAST RESEARCH MONITORED
2.1 Impact of user participation in Natural Resource Management Research by: N. Johnson
In 1999, the CGIAR Systemwide Program on Participatory Research and Gender Analysis
(PRGA) initiated a study of the impact of user participation in natural resource management.
An economist from BP1 was contracted to work on the design and implementation of the
project. Following are 1) a summary of the study and findings, 2) a description of the
conceptual framework and hypotheses and 3) the summary and conclusions. Further
information is available in the publications listed below.
2.1.1 Executive Summary
This study assesses the impacts of incorporating user participation and gender analysis in
natural resource manage ment research. Four types of benefits and/or costs are considered: (1)
impact on the technology developed and its adoption, (2) strengthening of human and social
capital among participating individuals and communities, (3) establishment or strengthening
of feedback links to formal research, and (4) costs of research. A typology of participation at
different stages of the research process is used to develop type- and stage- based hypothesis
for each of these four impacts. The hypotheses are evaluated in the context of three
participatory NRM research/development projects. The three projects are: the design and
development of integrated crop management (ICM) sweetpotato technologies by the Centro
Internacional de la Papa (CIP) and partners in Indonesia (1994-97); participatory testing of
legume based soil fertility technologies by the International Center for Research in the Semi
Arid Tropics (ICRISAT) in Malawi (1997-2000); and World Neighbors’s (WN) use of farmer
experimentation to adapt and diffuse soil conservation practices in Honduras (1981-1989).
Fieldwork and analysis of the three case study projects was done between August 2000 and
February 2001. Both qualitative and quantitative data were used, including existing project
documentation; open-ended interviews with project staff, farmer participants, and other key
informants including community leaders and policy makers; and statistical and econometric
analysis of survey data. Staff of the three projects participated actively throughout the
process.
The main findings of the study are summarized below by type of impact:
Impacts on technology and adoption. In all cases, farmer input influenced the technology
development process. Farmer impact on the technologies developed by the projects was
greatest when farmer input came early in the research process (CIP) or when technology
testing was done in a collaborative (empowering) way that gave scope for significant farmer
contribution (CIP, WN). In all cases, user participation contributed to greater awareness of
the technologies among farmers. In two of the three cases (CIP and WN), user participation is
linked to increases in adoption of project technologies. In the CIP case, detailed

72

Impact Assessment - Annual Report 2001

production data show that exposure to the ICM technologies is associated with higher
per-hectare income from sweetpotato production.
All cases used some type of gender analysis. In two of the three cases (CIP and WN),
gender analysis revealed that women were not important stakeholders in the NRM
activities that the projects were promoting. Only in the ICRISAT case were women
specifically targeted as beneficiaries and deliberately incorporated into the project as
participants. Disaggregation of participant input by gender did not reveal significant
gender differences in overall ranking of the technologies tested, however there were some
important differences in perception of specific characteristics of the technologies that
could be useful in designing gender-sensitive diffusion strategies.
Human and social capital impacts. Large human capital impacts were observed among
participants in the two projects that used collaborative participation at the testing stage
(CIP and WN). Where technology testing was consultative (functional) (ICRISAT),
useful agronomic and economic research results were obtained, but increases in
participant capacity and skills were small. Significant human and social capital impacts at
the design stage of the research process were not observed, even where empowering
participation was used. In general, human capital impacts were more prevalent than
social capital impacts. This may be due to the fact that the technologies being developed
and diffused were all essentially plot- level and did not require significant collective
action for implementation.
Since increases in human capital were only observed among direct participants in the
projects, if women do not participate directly in project activities they will not obtain
these benefits.
Feedback to formal research. In all cases, feedback to formal research and development
institutions was observed. These impacts were stronger for IARCs and NGOs than for
NARS. While some of the feedback was technical in nature and influenced institutional
research priorities (CIP), most was methodological, such as information about barriers to
adoption (ICRISAT), which is likely to benefit future research and extension efforts. In
all cases, the projects stimulated some researchers in their own and/or other institutions to
adopt more participatory methods. Feedback to formal research occurred with both
consultative and collaborative participation.
Costs of research. Incorporation of user participation was generally associated with four
types of additional costs: communications/workshops; farmer participant costs;
researchers field work costs; and training of researchers. Only in the first case do the
costs necessarily imply an increase in overall project expenditure. Farmer costs were
observed to replace (and sometimes reduce) researcher/research assistant costs at the
design, testing, and diffusion stages. Spending time in the field is a critical part of
participatory projects, however researchers must also spend time in the field to get good
results in conventional on- farm trials. Some of the observed cost increases may be more
associated with quality than with participation. In all projects, researchers increased their
own capacity and skills, either via formal training or learning-by-doing. These are

73

Impact Assessment - Annual Report 2001

essentially start-up costs incurred because the methods for collection and analysis of data
from participatory research processes are often new to researchers. Over time, as more
researchers gain experience in participatory research methods, these costs should decline.
Neither conducting gender analysis nor intentionally incorporating women as project
participants (ICRISAT) increased project costs.
Available data only allowed us to make a rough cost effectiveness estimate for the WN
project. Cost per hectare of land under soil conservation practices for the WN project
was estimated to be US$208. Similar per hectare costs for comparable projects that did
not use the same empowering participatory methods were between US$845 and
US$6000. The difference is the high and sustained adoption levels achieved by WN.

2.1.2 METHODOLOGY OF CASE STUDIES
Introduction and Objectives
Scope of the Study
Scientists using participatory methods have observed the success of this approach in a
variety of situations, and have documented the results in a number of case studies (e.g.,
Hinchcliffe et al 1999). However, the impacts and costs of using participatory as opposed
to more conventional approaches are rarely systematically analyzed or reported. Until we
have a better understanding of the tradeoffs associated wit h using participatory methods,
scaling up or institutionalizing participatory approaches will be difficult within
agricultural and natural resource management (NRM) research institutions.
By analyzing three research/development projects that used participatory methods in
applied research on NRM, this study aims to improve our understanding of the costs and
benefits of using participatory research. Using a shared conceptual framework, the
incorporation of participatory methods in each case is evaluated in terms of its impact in
the following four areas (for more detailed conceptual framework for impact analysis, see
p.1-22 in Lilja et al 2000):
(1) Adoption and impact of the technologies developed,
(2) Strengthening of human and social capital among the participating individuals and
communities,
(3) Establishment of feedback links to formal research, and
(4) Costs of doing research.
What we seek to evaluate in this study is not the overall impact of a research project that
used participatory techniques, but rather the incremental effect on impacts and costs that
can be attributed to incorporating stakeholder participation. Therefore, in each case study
the impacts are assessed against an appropriate conventional research counterfactual.
Participatory research can be done in many ways, and different methods may have very
different implications for outcomes and impacts. Selecting the appropriate approach will

74

Impact Assessment - Annual Report 2001

require anticipating how participation will impact on both project structure and goals.
Based on a typology of participation, we developed a comprehensive set of hypotheses
that link the incorporation of different types of participation at different stages of the
innovation process to the four impact areas identified above. In each case study, the
relevant hypotheses are tested within the context of the specific project. Taken together,
the results will provide some empirical evidence upon which to evaluate both project
impacts and the usefulness of the typology as a tool for project design, implementation,
and evaluation.
A final objective of the study is to look at the implications of participatory research for
different stakeholder groups, particularly women and the poor. In each case, the methods
of gender analysis used in a project are identified, and the impacts desegregated by
gender.
This project does not seek to reach general conclusions about which type of participation
is appropriate in what area; nor will we have the last word on whether or not the targeting
of women is cost effective. Much more experience in both the implementation and
evaluation of participatory research methods in many contexts will be needed before
these kinds of questions can be answered in any definitive way. What we hope to provide
in this study is a set of examples with sufficient diversity that a broad range of
researchers, research managers, and others can find similarities to their own work. It is
through their representativity of processes rather than their combined statistical power
that these studies will be useful in beginning to understand and evaluate the impacts and
costs of using participatory methods in agricultural and NRM research.
Conceptual Framework for Impact Analysis of Participatory Research: Types of
Participation and Their Implications for Impact
Types of Participa tory Research
The expected impacts of incorporating stakeholder participation in research are
contingent upon the nature of the approach used. Lilja and Ashby (1999a) develop a
typology of participation based on who makes decisions that permits analysis at different
stages of the research process. However, the research process is understood as being
iterative rather than linear. The typology defines the two decision- makers as “scientists”
and “farmers.” A generic term “farmers” is used to describe any target group and the term
“scientists” for outside agencies, extension systems, or formal research agencies.
Underlying this typology is the assumption that differences in who makes a decision will
result in differences in what decision is made. This need not be the case; however, cases
where the assumption holds are the most appropriate for participatory research methods.
The following is extracted from their framework.

75

Impact Assessment - Annual Report 2001

Stages of innovation
The innovation process can be divided into three stages – design, testing, and diffusion.
•

•

•

In the design stage, problems or opportunities for research are identified and
prioritized, and potential solutions to priority problems are determined. The outcome
of the decisions made at this stage is an array of potential solutions. They can be any
of the following: a completely new solution is invented and needs to be tested; a new
application of an existing solution is identified as having potential, but needs to be
tested; or an existing solution can be used, but needs to be promoted.
The testing stage is when potential solutions chosen for testing are evaluated.
Decisions are made about who does the testing, and about where and how it is done.
This stage results in recommendations to intended users about the innovation or
technology for mass distribution.
The diffusion stage involves building the awareness of recommended solutions
among future users. It involves decisions about when, to whom, and in what way to
build awareness, supply new inputs, and teach new skills to future users. The outcome
of decisions made at this stage is full or partial adoption, or no adoption.

Farmer participation at different stages of innovation can have different impact on the
technology or innovation design, as well as on the potential adoption or acceptance
among the intended users. Farmer participation early in the design stage helps reduce the
likelihood that the technologies being developed are ultimately unacceptable to farmers.
Their participation in planning and setting goals may help steer the research in a more
focused fashion and more directly towards farmers’ priority needs. Commonly, farmer
participation steers research into completely unanticipated directions. Similarly, who
participates at different design stages may lead to different priorities being identified for
different beneficiaries.
Who makes the key decisions in the participatory process?
In characterizing the participation in an innovation process we are concerned with
organized communication between or among the groups. By organized communication
we mean a well-defined procedure (such as informal surveys, group interviews, transect
walks, and formal surveys). Organized communication is not an ad-hoc opportunistic
event. We also differentiate between one -way communication, which is always scientist
initiated and where farmers respond to scientists’ inquiries, and two-way
communication, which may be scientist- or farmer- initiated, and scientists make sure
that farmers understand their opinions and ideas or their proposals and objectives, and
vice versa.
“Who makes decisions” is one way of deciding the balance of power in a participatory
process. We define five different types of participatory approaches depending on who
makes the decision at various stages in the innovation process. A different type of
participation is possible at each of the three stages of innovation (this builds on the

76

Impact Assessment - Annual Report 2001

previously known work on “categories” of participation, for example see Biggs and
Farrington 1991.)
(1)
(2)

(3)

(4)

(5)

Conventional (non-participatory): Scientists make the decisions alone without
organized communication with farmers.
Consultative : Scientists make the decisions alone, but with organized
communication with farmers. Scientists know about farmers’ opinions, preferences,
and priorities through organized one-way communication with them. Scientists may
or may not let this information affect their decision. The decision is not made with
farmers nor is it delegated to them.
Collaborative : The decision is shared between farmers and scientists, and involves
organized communication among them. Scientists and farmers know about one
another’s opinions, preferences, and priorities through organized two-way
communication. The decisions are made jointly; neither scientists nor farmers make
them on their own. No party has a right to revoke the shared decision.
Collegial: Farmers make the decisions collectively in a group process or through
individual farmers who are involved in organized communication with scientists.
Farmers know about scientists’ opinions, preferences, proposals, and priorities
through organized two-way communication. Farmers may or may not let this
information affect their decision.
Farmer experimentation: Farmers make the decisions individually or in a group
without organized communication with scientists.

Why does it matter who makes the decisions in the participatory process? If outsiders or
scientists make all the key decisions without farmer participation in the early stage of an
innovation process, farmers cannot influence many features of the innovation that are
fixed by those decisions. The outcome of the participatory research is different when
scientists and farmers plan together in the early stage and share key decisions, hence
increasing the likelihood that the farmers’ top priority is addressed. Participatory research
has a very different outcome if farmers make all the planning decisions and only consult
scientists late in the process when problems arise.
Implications for Impact
The expected impacts of incorporating participatory research approaches at different
stages of the innovation process are described in the following sections. Again, we are
interested in the impact of stakeholder participation on economic benefits from
technology adoption; the impacts of human and social capital benefits from participation;
and feedback to research and the cost of research. The second and third impacts are
examples of process impacts that occur as a result of the participation itself rather than as
a result of the technologies developed via participatory research methods. In the case of
process impacts, the type of interaction between scientists and farmers directly affects the
kinds of impacts that occur. Therefore, the hypotheses related to these impacts vary by
type as well as by stage.

77

Impact Assessment - Annual Report 2001

Technology impacts
The economic benefits associated with technologies developed using participatory
research are dependent on many factors including the specific technologies,
agroecological environment, input supply, and farmer and household characteristics.
However, some general hypotheses about how stakeholder involvement at different
stages might influence the adoption are given below.
Design stage:
(H1) The proportion of the targeted beneficiary group that could potentially be reached
by the project increases because the priority topic chosen for research is more
relevant to the needs and priorities of targeted farmers.
Testing stage:
(H2) The number of potential adopters within the target group increases because the
specific technology 1 selected for recommendation is more appropriate given
farmers’ criteria and constraints.
Diffusion stage:
(H3) The probability increases that potential adopters for whom the technology is
appropriate will be aware of it, and that adopters will be willing and able to
adopt and recommend it to others.
Social and human capital impacts (among beneficiaries)
It is hypothesized that through the process of interacting with researchers, the human and
social capital of participating individuals and communities can be strengthened. These
impacts would only be anticipated as a result of empowering participation, meaning
collaborative or collegial.
Design stage:
(H4) Collaborative: Farmers/communities improve their ability to interact with
outsiders, to articulate and evaluate their opinions and priorities, and to negotiate
joint solutions with other stakeholders who may have different opinions.
(H5)

Collegial: Farmers/communities improve their ability to interact with outsiders,
particularly their ability to attract the interest and support of researchers for
farmers’ problems and priorities.

Testing stage:
(H6) Collaborative: Farmers/communities enhance their own testing and evaluation
skills with an increased knowledge of scientific methods of experimentation and

1

For the sake of grammatical simplicity, in the text we will refer to “a research topic identified” or “a
technology tested or recommended.” In reality however, participatory research processes often identify
more than one priority problem, possible solution or appropriate technology.

78

Impact Assessment - Annual Report 2001

evaluation, and improve their ability to negotiate joint recommendations with
other stakeholders who may have different opinions.
(H7)

Collegial: Farmers/communities enhance their own testing and evaluation skills
with an increased knowledge of scientific methods of experimentation and
evaluation, and improve their ability to convince researchers of the validity and
relevance of farmers’ results.

Diffusion stage:
(H8) Collaborative/collegial: Farmers/communities learn what is involved in mass
diffusion of technology, particularly the complexity of adoption decisions and the
importance of complementary inputs such as seed, credit, or information.
A final hypothesis relates to the fact that, in many cases, participatory projects involve
farmers working together with other farmers as well as with researchers.
(H9)

The increased communication among farmers may result in better information
and in information sharing among farmers and within the broader community,
strengthening community social capital.

Feedback to formal research impacts
The previous section looked at the process impacts of participation on the beneficiaries.
In this section, we look at the benefits for the formal research process, specifically on
researchers’ access to information about farmers. These impacts can occur with any type
of participatory research, either functional or empowering.
Design stage:
(H10) Consultative: Researchers learn about farmers’ priorities and solutions.
(H11) Collaborative: Researchers understand farmer priorities and solutions—
including any new shared priorities or solutions that farmers and researchers
identify as a result of working together—and incorporate them into their work.
(H12) Collegial: Researchers learn about farmers’ priority problems and solutions by
observing their decisions about problems, solutions, and innovations.
Testing stage:
(H13) Consultative: Researchers learn farmer criteria for evaluating technologies.
(H14) Collaborative: Researchers understand farmer criteria and methods for testing
and evaluation of technology—including any new shared criteria or methods that
farmers and researchers identify as a result of working together.
(H15) Collegial: Researchers learn about farmers’ testing and evaluation methods and
criteria by observing their actions.

79

Impact Assessment - Annual Report 2001

Diffusion stage:
(H16) Consultative: Researchers learn about the factors that affect farmers’ adoption
decisions and what these imply for the diffusion process.
(H17) Collaborative: Researchers learn about farmer-to-farmer diffusion practices and
about what kinds of information and skills both farmers and extension workers
need to support this spontaneous diffusion.
(H18) Collegial: Researchers may learn about spontaneous farmer-to-farmer diffusion
through observation of farmer activities.
Finally, a general hypothesis that would apply at all stages is that:
(H19) Researchers begin to understand that working with farmers may require new
types of skills such as facilitation and conflict resolution that were not as
important when research was carried out entirely on-station.
This would be expected to increase as participation moves from functional to
empowering.
Cost of research impacts
As with the impact on economic benefits, the impact of participation on research
organizations’ costs is largely an empirical question. Several general hypotheses are
possible, however.
(H20) Moving from conventional to consultative or collaborative forms of participation
generally increases formal research organizations’ costs at the particular stage
where it is incorporated; however, it may reduce cost at subsequent stages.
(H21) Collegial research reduces research costs to formal research organizations at the
stage where it is implemented because costs are transferred to farmers.
(H22) Participation without compensation increases farmers’ costs unless it relies
exclusively on those farmers (often a small and unrepresentative group) who
already experiment on their own with new technologies and practices.

Gender Analysis in Participatory Research
Technological, policy, or other changes often have different impacts on different
stakeholder groups. One group that is often differentially affected is women. The
systematic disaggregation of data and analysis by gender is referred to as gender analysis.
Gender analysis and the targeting of women can be carried out in both participatory and
non-participatory research. Three common methods of doing gender analysis in the
context of agricultural research and technology development are 2

2

Extracted from Lilja and Asbhy 1999b.

80

Impact Assessment - Annual Report 2001

(1)

Diagnostic gender analysis. Gender differences in the client group(s) for the
research are described, and different problems or preferences are diagnosed. This
information is not taken into account in priority setting, design of solutions for
testing, or their evaluation and adoption. Diagnostic gender analysis may conclude
that gender differences are not an important criterion for designing the research; or
it may identify gender differences as an obstacle to adoption of technical solutions
for men or women members of the client group.

(2)

Design-oriented gender analysis. In addition to describing gender differences in
the client group with respect to their problems and preferences, different research
and development (R&D) paths are designed that take into account gender-based
constraints, needs, and preferences. Design-oriented gender analysis may result in
men and women developing and adopting different technologies, which may
require different dissemination approaches.

(3)

Transfer-oriented gender analysis. In addition to describing gender differences in
the client group with respect to their problems and preferences, different adoption
and dissemination paths are designed to overcome access to, and adoption of, a
given technology known or assumed to be of similar importance to men and
women. Transfer-oriented gender analysis results in the same technologies being
disseminated to men and women in different ways.

If diagnostic gender analysis results in the conclusion that gender differences are
important, the project can choose to target women specifically. Targeting can occur in
either the development of the technology, or in the design of the dissemination strategy.
In participatory research, attention to gender can go beyond targeting women as
beneficiaries to deliberately incorporating women into the research process. Designoriented gender analysis would be consistent with the incorporation of women at the
design and testing stages. In transfer-oriented gender analysis, women could be
incorporated at the dissemination stage.
Different ways of targeting women as beneficiaries and/or participants have different
implications for impact. Specifically, if women are not participants, then they will be
excluded from the process impacts described in the previous section. Whether or not
women must be participants in order to be beneficiaries of technology impacts is an
empirical question that will be examined in each of the cases. The cost impacts of
including women will also be examined.

81

Impact Assessment - Annual Report 2001

Selection of Cases
Criteria for Selection
Three cases were selected for analysis of the costs and impacts of incorporating farmer
participation in NRM research. Several criteria were used to select the projects. The first
criterion was to identify projects that had documented impact or that had been operating
long enough to generate intermediate or final impacts. Additional criteria were to include
a range of geographical areas, types of NRM technologies, and implementing
organizations. The cases are part of an inventory of projects doing participatory research
on natural resource management (NRM inventory, www.PRGAprogram.org) The
projects selected are outlined below.
The Centro Internacional de la Papa (CIP) development of integrated crop
management (ICM) technologies and practices for farmer field school (FFS) for
sweet potato in Indonesia (1990s)
During 1995-97, CIP, with support from UPWARD3, and in collaboration with public and
private sector groups, implemented a project to develop a protocol for a sweet potato
ICM-FFS in Indonesia. Collaborators were Mitra Tani, a local nongovernmental
organization (NGO); the National Research Institute for Legume and Tuber Crops; and
Duta Wacana Christian University. Project activities were implemented in major sweet
potato growing areas in East and Central Java, where it is grown as an important cash
crop throughout the year, mostly in rotation with rice. The project strategy relied on
participatory approaches and methods at all stages: needs assessment and project design;
R&D of ICM technologies and practices; design of farmer learning protocols applying
the FFS approach; pilot-scale implementation of the sweet potato ICM-FFS; and
monitoring and evaluation. To institutionalize the sweet potato ICM-FFS model that was
developed, and allow for large-scale farmer learning and implementation, staff from the
National IPM Program (NIPMP) and 30 local NGOs underwent FFS facilitators’ training;
NIPMP staff in June 1997, NGO staff in April 1998. These local extension organizations
implemented and funded follow-up programs, and a second research project was initiated
to evaluate their activities during a 2-year period (1998-99). Mitra Tani carried out the
work, with methodological and financial support from CIP and UPWARD.
The International Crops Research Institute for the Semi-Arid Tropics (ICRISAT)
work on models for the participatory testing of soil fertility technologies in Southern
Africa (1990s)
The ICRISAT Mother–Baby (MB) trial model is an upstream participatory research
methodology designed to improve the flow of information between farmers and
researchers about technology performance and appropriateness under farmer conditions
(Snapp 1999b). The methodology was initially developed and implemented to test soil
fertility management technologies in Malawi that were legume-based, and was later
3 The Users’ Perspectives for Agricultural Research and Development (UPWARD) is a CIP-affiliated
network of Asian researchers conducting participatory R&D projects in root crop systems.

82

Impact Assessment - Annual Report 2001

expanded to Zimbabwe. The trial design consists of two types, mother and baby trials.
The mother trial is researcher-designed and conforms to scientific requirements for
publishable data and analysis. A baby trial consists of a single replicate of one or more
technologies from the mother trial. A single farmer manages each baby trial on his or her
own land. A typical implementation of the methodology would include a single mother
trial and numerous baby trials within a village. The MB trial methodology has three
goals. The first is to generate data on which to assess technology performance under
realistic farmer conditions. The second is to complement the agronomic trial data with
farmers’ assessments of the adoption potential of technologies. This information helps
researchers understand how the technologies fit into farmers’ broader farming and
livelihood strategies. The third goal is to encourage farmers to actively participate in the
trials, and is expected to stimulate farmer experimentation with, and adoption of, new
technologies and practices.
World Neighbors (WN) soil conservation work in Honduras (1980s and early 1990s)
This project, supported by WN, the Coordinating Association of Resources for
Development (ACORDE), and the Ministry of Natural Resources of the Government of
Honduras, promoted improved soil conservation practices in south central Honduras from
1981-1989. The project worked in 41 communities in three municipalities – Guinope, San
Lucas, and San Antonio de Flores – in the state of El Paraiso. The project’s approach
went beyond strictly increasing agricultural productivity through the adoption of soil
conservation practices, to improving economic, social, and ecological conditions via
agriculture (Plan del Programa). Although the project was primarily one of development,
it had a significant capacity building component, teaching farmers the principles of soil
conservation technologies, training them to experiment and adapt technologies, and
imparting knowledge about selection and improvement of genetic materials for green
manure. The project carried out these activities in the context of community groups, and
trained local farmers to take over extension jobs after several years. The purpose of these
activities was to build social as well as human capital, while strengthening organizational
capacity and the capacity and commitment to share knowledge within the community.
The project methodology was that described in Bunch (1982) that advocates a
combination of 80% practical training and 20% theory. Significant increases in adoption
were observed in the study areas during the course of the project. According to WN
reports, nearly 1400 farmers tripled their basic grains’ yields as a result of adopting soil
conservation practices. Subsequent follow- up studies indicate that further adaptation and
adoption continue (Bunch and Lopez 1999). Increases in productivity were also observed,
as were increases in farmer experimentation, and the exchange of information among
farmers. After several years working mainly on agriculture with men, the project added a
component for women. This focused on sanitation, home gardens, and food preparation.
Representativity of Cases and Types of Participation
To assess the representativity of the three cases, they can be compared with an inventory
of participatory NRM research cases which was complied by the PRGA
(www.prgaprogram.org). According to an analysis of the cases, the majority are from

83

Impact Assessment - Annual Report 2001

Africa (38%) followed by Asia (34%) and Latin America (22%) (see Johnson, Lilja and
Ashby, 2001 for a thorough analysis of the inventory cases). The most common resource
in those cases was soils, and this is also the major focus of the three cases in this impact
study. Of the three study cases, one (WN) was NGO supported; the others are
International Agricultural Research Center (IARC) cases. The important actor that is
missing here is the National Agricultural Research System (NARS). In each of the cases,
NARS were partners in the implementation, but did not undertake the work directly. This
is also consistent with the findings of the analysis of the inventory. Two of the three
study cases are managed by CGIAR centers. CG centers were also the dominant
organizational type in the inventory (37% of cases), however it is likely that that the
predominance of CGIAR work in this area is overstated due to the fact that CGIAR
organizations (PRGA, CIAT) compiled the inventory.
The type of organization appears to be associated with project design and
implementation. The NGO project is essentially an extension program that incorporates
farmer testing and experimentatio n as a dissemination mechanism. In the IARC projects,
in addition to working with farmers on a specific technology, methodologies were
developed for systematically implementing similar work with other farmers. In these
cases, the methodology itself is an output, not just the specific technologies developed in
given field sites.
Table II-1 represents an attempt to place the cases within the framework of the typology
presented in the previous section.
Table II-1.
Stage

Types of participation used in the three case studiesa .
Empowering

(non-participatory)

Functional
Consultative

Design
Testing

WN, ICRISAT
-

CIP
ICRISAT

CIP
CIP, WN

-

Diffusion

(ICRISAT)

(CIP)

CIP, WN

CIP, WN ICRISAT

a.

Conventional

Collaborative

Collegial

Farmer
Experimentation

WN = World Neighbors – Honduras, CIP = Centro Internacional de la Papa – Indonesia, and
ICRISAT = International Crops Research Institute for the Semi-Arid Tropics – Southern Africa.
Parentheses indicate an activity that is planned, but remains to be executed.

Overall, the table shows that at the design stage, participation tends to be slightly more
consultative, with farmers giving input, but researchers making decisions. Only in the
case of CIP did stakeholder participation substantially change researchers’ agendas,
budgets, and work plans. Where significant research expenditures will occur, it appears
that control still rests with the researcher, largely reflecting the sectoral nature of R&D
funding. Cases in the inventory also tended to be more consultative.
As mentioned earlier, classifying project activities as either testing or diffusion was
difficult. It appears that traditional, planned diffusion now begins in the testing stage,
where researchers hope that by involving farmers in trials, both participating and non
participating farmers will gain awareness of the technologies. This change may be related
84

Impact Assessment - Annual Report 2001

to the fact that the information- intensity of NRM practices requires farmers to learn
something about how the technologies work. It may not be possible to adopt a technology
off-the-shelf with no adaptation. Traditional diffusion may occur at the level of the
methodology for facilitating adaptation rather than the technology itself. This is a subject
that will be examined in more detail in the context of the case studies.
Use of Gender Analysis
All three of the projects undertook diagnostic gender analysis. In two of the cases, CIP
and WN, it was decided that women did not play a significant role in the principal
activity being addressed in the project. World Neighbors responded by implementing a
separate set of activities for women focused on nutrition and health. Women were not
excluded from the soil conservation activities, but self-selection of participants resulted in
few women being involved.
Some Issues in Empirical Analysis
Before presenting each case study in detail, this section discusses some of the common
empirical challenges associated with analyzing the impact of participatory NRM
research. Each case study had to address the issues in some way in the empirical analysis.
Controlling for Selection Bias
Selection bias is an issue in any analysis where the treatment groups (study communities)
were not randomly selected. When projects choose to work with specific individuals or
communities, they may be doing so for reasons that may also be associated with the
observed impacts. For example, interventions based on local collective action are often
implemented in communities that have high levels of social capital. Failure to account for
this could result in a project taking credit for social capital when in fact social capital
contributed to the project’s success. Further, even if the project did have a measurable
impact on such a community, it would be difficult to extrapolate about what the impact of
a similar project might be in an area where social capital is not so high. Knowing how
communities were selected is important so that appropriate control communities can be
identified. The ideal situation is to collect pre- and post-project data for many replications
of the project in different types of communities as well as in sites where no intervention
occurred. This allows us to look at changes associated with the project and to control for
the influence of other factors on observed outcomes. Without this, extrapolating beyond
the specific project site(s) is difficult.
In addition to researcher-selection bias, self- selection bias is also likely to occur in
participatory projects. Because participation is voluntary, people can choose to participate
or not. In any given situation, the people who choose to participate are likely to be
different from those who choose not to do so. For example, the type of individual who
wants to volunteer his or her time to be part of a local agricultural research group is likely
to be someone who already has an interest in experimentation, or someone who has a
high level of commitment to working for the good of the community. The consequences

85

Impact Assessment - Annual Report 2001

of self-selection bias are similar to those of researcher-selection bias. However, it is
harder to control for because the criteria for selection are largely unobservable and
uncorrelated with observable characteristics such as age, education, or income.
Extrapolation of impacts from self-selected participants to the broader population may
not be appropriate.
Identifying the Appropriate Counterfactual
As mentioned earlier, the goal of this analysis is to look at the impact of incorporating
stakeholder participation on the costs and impacts of developing and disseminating NRM
innovations. This implies that the counterfactual is the impact that would have occurred if
the project had used only conventional methods. In this study, the conventional research
and/or extension counterfactual is only used when comparing technology impacts and
research costs. We are making the assumption that non-participatory projects do not have
impacts on human or social capital or on the research process. These impacts—which we
refer to as process outcomes—result from the interaction of researchers and farmers; thus
they could not occur in a non-participatory project.
The extent to which this counterfactual can be achieved in each case study varies. In
some cases, participatory methods were either not part of the original project plan or
resulted in major changes in project activities. In these cases, we have enough
information to use the original project plan to construct a hypothetical counterfactual. In
other cases, we will need to identify an appropriate comparison with a project that
addressed a similar problem in a similar community us ing non-participatory methods.
This may be another project by the same or a different organization.
The Definition of Technology and Adoption
In every project farmers were encouraged to experiment and make changes to the
technologies. This was an important goal of the projects. However, it complicates impact
assessment because it is not always clear whether an innovation is really an adapted
version of a project technology or something entirely unrelated. This type of outcome is
common, and getting at the causality of these impacts may require more qualitative and
participatory methods of data collection than conventional impact studies that have
clearly defined technologies and definitions of adoption.

2.1.3 THE IMPACTS AND COSTS OF USER PARTICIPATION IN
NRM RESEARCH: A COMPARATIVE ANALYSIS OF THREE
CASES
In this section, all empirical impacts observed in the three case studies are discussed
separately, by stage of research process (see table II-34 for summary). Given the types of
participation in the cases, the main hypotheses that we can test relate to consultative
participation versus conventional research methods at the design and testing stages, and
collaborative participation at the dissemination stage. The remainder of this paper is
organized around the three stages of the research process-design, testing and diffusion..

86

Impact Assessment - Annual Report 2001

For each stage, relevant hypotheses for all four impacts are assessed. A summary of the
comparative analysis is presented in table II-34. The concluding section assesses the
usefulness of the typology of participation and the conceptual framework developed for
this analysis.
User Participation in Project Design
The CIP project was the only one that involved user participation in project design where
key problems and potential solutions are identified. In both ICRISAT and ACORDEWN, project staff made these decisions, drawing on past experiences and in consultation
with other R&D professionals.
Technology Impacts
The hypothesis regarding participation at the design stage is not type-specific, which
means that impact would be expected to increase as participation of farmers increased. It
was hypothesized that participation at the design stage would lead to an increase in the
proportion of the targeted beneficiary group that could potentially be reached by the
project because the priority topic chosen for research would be more relevant to the needs
and priorities of targeted farmers (H1). The findings in these case studies support this
hypothesis with regard to impacts on the size of the potential beneficiary group.
In the case of CIP, user input in the design stage led to significant changes in the focus
and activities of the project, most importantly the shift from IPM to ICM. Analysis of
both production data and opinions of FFS and non-FFS participants show that ICM
aspects of the work were indeed most relevant to farmers, and contributed to the benefits
that farmers obtained.
Where farmers did not participate at the design stage, topics initially selected for research
were less relevant to users’ immediate needs. In the ICRISAT case, evidence supports the
researchers’ conclusion that soil fertility was the problem; however, the technologies
chosen for testing (the legume intercrops and rotations) were found not consistent with
users’ preferences and cons traints. Researchers selected the technologies based on their
agronomic performance, although they knew that they had not been accepted in the past.
Lack of adoption was in part the motivation for seeking farmer input via the MB trial
methodology. Useful information for research was obtained as a result of the project;
however, adoption remains low.
The WN project was similar to the ICRISAT case in that it worked with existing
technologies whose adoption had been low. However, in ACORDE-WN’s case, the
project’s flexible implementation and intensive focus on farmer capacity building and
adaptation overcame some initial inconsistencies between user needs and project
technologies. The conclusion from these three studies is that if adoption is the goal in the
short run, then involving users in the early stages can help identify technologies that are
relevant and appropriate. If users are not involved at the beginning, it may mean that
farmers either spend more time later in adapting technologies or they simply do not adopt
them.
87

Impact Assessment - Annual Report 2001

Impacts with regard to women are mixed; however, there is no support for the
assumption that not including gender explicitly hindered a project’s ability to achieve its
stated NRM objective. Diagnostic gender analysis was carried out in the CIP case, and it
was found that women generally played a small and secondary role in sweet potato
production. Therefore they were not targeted in subsequent project activities. Some
women did ultimately participate in the SP ICM FFS as implemented by the NIPMP, and
the results of the impact assessment do not show gender differences in the economic
benefits of participation.
The ICRISAT case did not conduct gender diagnostic analysis; however, it was assumed
that cultural aspects of agriculture (e.g., men’s vs. women’s crops) would make this
project relevant to women. Women played a part in the project activities, and appeared to
have benefited economically to the same extent that men did.
In the WN case, no formal diagnostic gender analysis was done. However, again project
staff knew that women were not involved in agricultural production and therefore did not
include them in NRM project activities. In this case, we did not find examples of women
benefiting directly from the technologies or practices promoted by ACORDE-WN.
Including women in the design stage of the NRM component of this project would likely
have resulted in a very different project, perhaps not focusing on soil conservation or
even crop agriculture. Not including women does not appear to have diminished the
impacts with regard to the project-stated soil conservation objectives.
Although these studies support the hypothesis that farmer participation at the design stage
increases the size of the potential beneficiary group, no support was found for the implicit
assumption that the magnitude of benefits increases as participation moves from
functional to empowering. The CIP did a consultative assessment of constraints and
opportunities in sweet potato production. They also did data collection and interpretation,
and pilot field schools to test the methodology for sweet potato collaboratively. Given the
iterative nature of the project, the usefulness of consultative work done just at the
beginning is difficult to compare with the ongoing collaborative activities. However,
clearly some of the lessons from participation, such as the importance of non-pest issues
or the opportunity for improving efficiency of nutrient management, emerged in both
consultative and collaborative analysis. In fact, the consultative analysis may even have
identified more constraints. Farmer-researchers were initially reluctant to believe some of
the researchers’ conclusions from needs’ assessment, such as the importance of economic
analysis or the possibility that current nitrogen fertilizer use was excessive. Over time,
and as a result of experimentation activities, they were convinced. Collaborative
participation was important in refining the issues and in developing and testing
technologies, but regarding their contribution to identifying problems and priorities,
collaborative participation at the design stage did not appear to be better than
consultative, and may even have been less effective.

88

Impact Assessment - Annual Report 2001

Human and Social Capital Impacts
It was assumed that human and social capital impacts would only occur with empowering
participation, which we define here as collaborative or collegial. Using collaborative
participation at the design stage was expected to improve participants’ ability to interact
with outsiders, to articulate and evaluate their opinions and priorities, and to negotiate
joint solutions with other stakeholders who may have different opinions. Here we can
assess whether these impacts occurred in the collaborative aspects of the CIP case.
Because no women participated in the collaborative research activities, genderdifferentiation of benefits cannot be assessed. We can look at whether any impacts were
observed among the participants in the consultative aspects, and test the assumption that
no impacts occur with functional participation.
Because the principal participants in the CIP collaborative activities at the design stage—
the farmer-researchers—were also the participants in collaborative activities at
subsequent stages, separating the two effects is difficult. A way of doing so is to look at
progress over time. Project staff observed that farmer-researchers were initially hesitant
to advance opinions, and that the content of their input was often simplistic. As their
experience with experimentation and sharing information grew, they became much more
confident and more capable of managing complex concepts, activities, and interactions.
This suggests that their involvement in design-stage activities alone was not sufficient to
increase their capacity. However, comb ining design and testing stages led to significant
impacts. This finding does not support the hypothesis that collaborative participation at
the design stage has human and social capital benefits.
Regarding impact on the participants in the consultative activities, the individuals who
participated in these activities were not an explicit focus of the data collection. Some
were involved because they later participated in the FFS, but as a group the farmers who
collected field data or who answered questions in the consultative needs’ assessments
exercises were not systematically evaluated. We can state that in all our interactions with
farmers, the farmer-researchers were significantly different from others in terms of their
ability to express themselves, to answer and ask questions, and to explain complex issues.
They were also more likely to experiment, and to be sought out by others for advice.
Further, the village leaders and other key informants specifically mentioned changes in
the human capacity of farmer-researchers, but said nothing about these impacts with
regard to other farmers. These observations are consistent with the assumption that there
are no human and social capital impacts with functional participation at the design stage.
Feedback to Formal Research Impacts
Similar to the human and social capital benefits among participants, we would also
expect to find benefits to the knowledge and capacity of researchers as a result of their
interactions with farmers. With consultative participation, researchers would be expected
to learn about farmers’ existing problems and their priorities for solutions. In
collaborative participation, researchers and farmers interact so that although researchers
still ultimately learn about farmers’ priorities and constraints, these may have evolved as

89

Impact Assessment - Annual Report 2001

a result of the interactions with researchers. Further, in collaborative participation,
researchers incorporate farmers’ perspectives in their work because authority for
identifying final priorities, problems, and solutions is shared. Impact can be observed
among researchers at the project, program, or institutional levels.
The results from the CIP case support the general hypothesis that user participation
increases researchers’ knowledge of user priorities and constraints. The project
contributed to changes in CIP research priority decisions regarding sweet potato weevil in
Asia, the importance of scab, and the the need to screen germplasm for important
commercial characteristics like starch content. The ICM concept is also widely used
within the center.
Support for the hypothesis that the nature of feedback impacts differs by type of
participation is less strong. As reported earlier, the initial information generated from
consultative and collaborative participation at the design stage was very similar. As
mentioned above, researchers drew some conclusions from their consultative work that
did not emerge initially in the collaborative work. Over time, as the collaborative
interaction between researchers and farmers continued, new priority topics and problems
emerged, partly as a result of new knowledge and perspectives gained through
interaction, confirming that collaboration can influence farmers’ priorities and criteria.
Cost of Research Impacts
It was hypothesized that costs increase as participation moves from conventional to
consultative or collaborative participation, although they may be reduced later.
Participation is also hypothesized to increase farmers’ costs. In general the results of
these three cases support these hypotheses; however, the magnitudes of the differences
may not be as large as is often assumed. For example, ICRISAT invested in participatory
activities with researchers from different institutions to identify problems and select
technologies. It may have been possible to include farmers with little additional cost, and
their input could have resulted in changes in technologies and/or implementation
strategies. In the CIP case, the financial costs of the consultative and collaborative work
were not so different. The consultative was shorter in duration with more high-cost
researcher time. There is no evidence that conducting diagnostic gender analysis
increased costs. CIP’s collaborative design-stage work lasted longer—2 to 3 years, but
this also included many of the testing stage activities—and involved more farmer and
field assistant time. The CIP case suggests that even at the collaborative stage some
research costs, such as managing data collection, can be passed on to farmers, which
raises their costs, but lowers total project costs even when farmers are compensated. This
impact was previously hypothesized to occur only in collegial participation.

90

Impact Assessment - Annual Report 2001

User Participation at the Testing Stage
At the testing stage, where solutions are tested and evaluated and recommendations
made, all three projects used some form of user participation. The ICRISAT project used
consultative, and the CIP and ACORDE-WN projects used collaborative.
Technology Impacts
The size of the potential pool of potential adopters is determined at the design stage
through the selection of the problem to be addressed and the type of solutions to be
considered. At the testing stage, user involvement is hypothesized to contribute to
increased adoption by helping ensure that the technology or technologies ultimately
selected for dissemination are appropriate for the largest number of people either in
general or in specific beneficiary groups. The cases provide some support for this
hypothesis.
The ICRISAT case, where participation was consultative, provides on suggestive,
preliminary evidence about possible future impacts. Where farmers carried out trials
according to researcher-designed protocols, farmers’ perceptions and rankings of the
technologies differed from those of researchers. This information could be useful for
future technology development or for designing dissemination strategies. Because women
participated in the trials, preferences could be disaggregated by gender. Some differences
were found, but were not statistically significant. Because testing was consultative, we
cannot know whether farmers had any ideas about how the technologies might be adapted
to make them more appropriate for their purposes, nor whether gender differences would
have mattered here.
Where testing-stage participation was collaborative, farmer input led to changes in the
technologies themselves, which were linked to higher levels of adoption. Farmerresearchers in the CIP project made a significant contribution to testing, evaluation, and
adaptation. They selected which technologies to test, designed and implemented trials,
and evaluated results. In many cases, such as fertilizer use, they may have obtained the
same results and interpreted them in the same way as researchers would have done. In
others, such as cultural practices or the testing of the FFS methodology, they provided
insights that researchers would not have had.
In the case of the ACORDE-WN project, the technologies selected by the project were
presented to farmers as a basis on which to begin a self-sustaining process of innovation.
High levels of adoption and especially adaptation, both during and after the project,
support the hypothesis that farmer involvement in testing and modification is key to
achieving high levels of adoption with NRM technologies such as soil conservation
practices.

91

Impact Assessment - Annual Report 2001

Human and Social Capital Impacts
Since human and social capital impacts were assumed to occur only with empowering
participation the relevant hypothesis to test related to the collaborative participation used
in the CIP and ACORDE-WN cases. The hypothesis was that farmers/communities
would enhance their own testing and evaluation skills with an increased knowledge of
scientific methods of experimentation and evaluation, and improve their ability to
negotiate joint recommendations with other stakeholders who may have different
opinions (H6). The null hypothesis of no impact would be expected for the ICRISAT
project. The experience of the three cases is consistent with the hypothesis.
In the cases of CIP and ACORDE-WN, where farmers and researchers carried out
experiments collaboratively over a period of several years, human capital benefits to the
farmers were significant. As mentioned above, they increased their knowledge and
understanding of agroecology and of experiment methods. In the case of the CIP farmerresearchers and the ACORDE-WN farmer- leaders, they also increased their selfconfidence and ability to interact with outsiders such as researchers and extension agents.
The farmer-researchers and farmer-leaders continue to experiment, and to be sources of
information in their communities. No such impacts were found among participants in the
ICRISAT case, whose interaction with researchers was significantly less intense. The
baby-trial farmers enjoyed participation and benefited directly from the opportunity to
work with the field assistant and ask him questions. However, we found no evidence of
substantive changes in their understanding of soil fertility management issues or
experiment methods. It is important to note that these results reflect the impacts of the
MB trail method as observed in Chisepo. ICRISAT’s subsequent impact assessment
work found that when implemented in a more collaborative manner, the MB method did
generate human and social capital impacts.
Although human capital benefits were visible and significant, social capital benefits were
less so. In the CIP case, many benefits were widely socialized in the sense that the
farmer-researchers shared what they learned with their neighbors. However, this seemed
to be due more to already existing habits of and networks for information sharing rather
than to new patterns of behavior stimulated by the project. In the ICRISAT case,
information sharing increased in the context of the project because participants were
specifically instructed to share what they were doing with neighbors—something they
were not always doing spontaneously. We found little evidence that this would continue
after the project ends. No other group activities were formed in connection with the
project activities. In both cases, the lack of externalities or other aspects of the resources
and technologies that might require collective action surely contributed to the lack of
social capital impacts.
The ACORDE-WN project had impact on social capital variables such as information
sharing and community activities. The technologies promoted in this project were also
essentially at plot level and therefore did not require collective action for effective
implementation. However, the project devoted a lot of time to building individual and
group capacity in addition to promoting soil conservation technologies. Despite these

92

Impact Assessment - Annual Report 2001

efforts, in terms of overall community solidarity, evidence on impacts was mixed,
reflecting that these types of impacts are complex and difficult to assess, and that
stakeholder differentiation is highly important.
Feedback to Formal Research Impacts
Consultative participation at the testing stage is expected to increase researcher
knowledge about farmers’ criteria for evaluating technologies. With collaborative
participation, the potential benefits go beyond researchers learning about farmers’ criteria
to include the establishment of new shared criteria and the incorporation of the criteria
into their research activities.
Researchers and extension agents in all cases benefited from their interactions with
farmers. In the ICRISAT project, researchers not only learned of new criteria—such as
the ability of technologies to suppress weed growth—but also learned that farmers give
less value to other criteria, such as yield potential, than researchers do. Some evidence
showed that women and men may have different criteria, and further project work will
focus exclusively on women. Researchers from other institutions have adopted the
particular participatory methodology used, the MB trial method, for getting basic farmer
input from many farmers.
The CIP approach was much smaller scale, working more intensively with fewer farmers.
As mentioned in the previous sections, farmers learned a lot from researchers and were
clearly influenced by researcher knowledge and methods. The benefits to researchers
from the testing stage activities of this project were not so much related to specific
criteria or aspects of technologies, but rather to more conceptual issues concerning the
different purposes for which experiments are conducted. Within the project, there were
experiments carried out by farmer-researchers to assess new technologies, and
experiments with the FFS for learning and demonstration purposes. The project had not
initially recognized these as different, with different implications for how they should be
presented to participants, what skills were needed from both participants and researchers
for implementation, and how the results should be interpreted. The project is influencing
CIP’s research program at a broader level. An NGO involved in the project has also made
radical program changes and now incorporates elements of research in all its activities.
Impact on the NARS involved was limited.
In the ACORDE-WN project, intensive interactions occurred between farmers and
extension agents concerning how to test, evaluate, and adapt technologies. Because this
was an extension-oriented project and the purpose of experimentation was to facilitate
adaptation and adoption, systematic data on the experiments and their results were not
recorded, analyzed, or published. This limited the extent to which others learned from
specific technical findings of the trials. However, a great deal of research and extension
attention has been attracted to the methodological aspects of the project, especially at
international R&D organizations, mainly because of its high adoption rates. In what
attention there has been to farmer adaptation, such as in the case of live barriers, the

93

Impact Assessment - Annual Report 2001

purpose has been to call attention to the participatory methodology, not the adaptations
themselves.
Costs of Research Impacts
It is hypothesized that moving from conventional to consultative or collaborative forms
of participation increases research costs, at least in the short term. Long-run cost
effectiveness should, however, increase. Farmers’ costs also increase unless the project
works only with the subset of farmers who already engage in experimentation.
In the case of the ICRISAT MB trials, the main costs associated with supporting the baby
trials were enumerator salaries, training for project staff in participatory methods, and
research time learning and conducting analysis of the type of data generated by the trial
method. In the latter two categories, these are one-time costs associated with the PR
methods being new in the Center. Future PR projects using the same staff will not have to
bear these costs to the same extent. Therefore, the main recurring cost is enumerator time.
How this cost compared to non-participatory projects depends on the nature of those
trials. In some cases, on-station trials or carefully managed trials where field assistants
regularly visit test plots could be more costly than MB trials (especially on a per trial
basis). On the other hand, less intensive methods, such as one we were told of where
researchers sent trial kits by bus to extension agents to plant, monitor, and send back
results, would clearly be less costly to the research program. Controlling for quality, the
actual implementation costs of the MB trials likely were not significantly higher than a
non-participatory method.
Farmers were not compensated except for the seed they received from the trial; yet many
farmers, both male and female, wanted to be involved initially and stayed involved in
subsequent years. This suggests that the trials were not a financial burden for the
participants, although some of this may be because the field assistant helped with much
of the work. The field assistant said it was easier to work with women because they were
more likely to follow the protocol.
In the CIP case, the regular and intensive interaction among researchers and farmerresearchers via workshops added to costs. However, the costs of the actual testing done
by the farmer-researchers were not especially high. The farmers were compensated for
their time, but they were paid relatively little compared to a researcher or a research
assistant. In the early years of the project, farmers were less skilled, but as their
experience and capacity grew—largely because of the intensive interaction with
researchers—the quality of their experiments improved. As was the case with ICRISAT,
use of PR methods is costly at first because new skills must be learned. Once these skills
are in place, then the costs decline and quality improves.
The ACORDE-WN project dedicated many more human resources to the communities
where they worked, but worked with farmer-extension workers rather than researchers or
professional extension agents. This probably lowered costs and increased impact.

94

Impact Assessment - Annual Report 2001

However, comparing this project methodology to conventional research trials is difficult
because trial data were not kept or analyzed.
User Participation at the Diffusion Stage
The CIP and ACORDE-WN projects used collaborative participation in the diffusion
stage, which involves identifying target beneficiaries and designing an extension strategy
and/or methodology involving complementary inputs. The ICRISAT project did not have
an explicit design component for participatory dissemination. What has been done to
date, including production of an extension brochure, is conventional.
Technology Impacts
It was hypothesized that by incorporating users in the design of the dissemination
activities, adoption would increase by making sure that those for whom the problem and
technology were relevant and appropriate had adequate information, skills, and other
inputs necessary to adopt. The experiences of the CIP and ACORDE-WN projects
support this hypothesis. In the CIP case, for example, the flexibility of the FFS
methodology and the focus on experimentation and skill building made it possible for
NGOs to implement it for crops other than sweet potato. In the ACORDE-WN project,
the willingness of staff to adjust both technologies and project activities to suit farmer
preferences increased farmer interest and participation.
In both the CIP and the ACORDE-WN cases, the key complementary inputs were
knowledge and skills. Evidence from the study of spontaneous diffusion in the ICRISAT
case suggests that seed availability may be a constraint there. However, little attention
was given in that project to farmer knowledge and capacity for adaptation.
None of the projects designed gender-differentiated strategies for diffusion of their NRM
technologies. Evidence from studies of spontaneous diffusion in the ICRISAT and CIP
cases suggest that women and men receive and diffuse information through different
sources. This suggests that if a main part of a project’s diffusion strategy is by farmer-tofarmer dissemination, attention to gender may be important in achieving impact.
Human and Social Capital Impacts
Incorporating users into the design of diffusion strategy is expected to increase their
understanding of what is involved in mass diffusion of technologies, including the
complexity of decision making and the importance of complementary inputs. The results
of the cases with regard to this impact are mixed. In the case of WN, the farmers who
were selected to work as promoters and later extension workers learned a great deal about
the diffusion of technologies, with many going on to work as extension agents,
consultants, and even founders of agricultural service centers. In the CIP case, an attempt
to train farmer-researchers to become FFS trainers was not successful. The farmers felt
uncomfortable with the process, and much preferred their roles as researchers and
resource people. The reason for the different outcomes may be that the CIP farmer-

95

Impact Assessment - Annual Report 2001

researchers were selected for one purpose (research) and then trained for another
(training/facilitation). In the ACORDE-WN case, the farmers were part of a lengthy
selection process that included both experimentation and extension activities.
Feedback to Formal Research Impacts
The hypothesis for feedback to research from collaborative participation was that
researchers learn about farmer-to- farmer diffusion practices and about what kinds of
information and skills both farmers and extension workers need to support spontaneous
diffusion (H17). Both the CIP and ACORDE-WN case experiences support this
hypothesis, with each project making changed to its proposed dissemination strategy
based on farmer input.
The impacts regarding the knowledge and skills needed for successful systematic
dissemination were much stronger than the ones relating to farmer-to-farmer
dissemination. This is especially evident in the ACORDE-WN case, where diffusion
beyond the project communities has been limited eve n 15 years after the project ended.
As mentioned earlier, the main “complementary input” needed is information and skills,
which it may not be possible to leave to “market forces” in the same way as inputs such
as credit or seed production.
Cost of Research Impacts
Costs were hypothesized to increase as participation increased, and this was the case in
the projects at least in the short run. However, cost effectiveness appears to have
increased because of participation. In the CIP case, running trial FFSs to get farmer
feedback and train farmer-researchers was costly, but farmer input led the project to
abandon plans to develop its own implementation capacity and work through others, thus
reducing project costs and likely enhancing cost effectiveness. It also contributed to the
important lesson about farmer-researchers’ unsuitability for training/facilitation.
In the ACORDE-WN case, training farmer-extension agents incurred some costs,
although much of the early learning was done as part of regular project activities. These
farmer-extension agents came to replace outside project staff, which reduced costs and
enhanced impacts. Comparisons of cost effectiveness figures show that the project was
much more cost effective than other projects promoting similar technologies. In both
cases, it must be pointed out that the farmers involved in the design of diffusion strategies
had also been involved in other earlier testing stage activities of the projects. This
experience clearly increased their effectiveness to contribute in the diffusion stage
activities.

96

Impact Assessment - Annual Report 2001

Table II-34.
Stage

Design

Summary of the main impacts of user participation by stage and type
Main impacts a
Technology and its
adoption

Human and
social capital

Feedback to
research

Costs of research

Highly important if
goal was adoption
and/or if subsequent
farmer adaptation
was unlikely.

Low, even in
empowering
participation.

Important impacts
within and beyond
the projects. Limited
impact on NARS.

Cost increase compared to
conventional, but
empowering was not more
costly than functional.

Empowering
participation not
necessarily better
than conventional.

Diagnostic gender analysis
did not increase costs of
consultative participation.

Empowering
participation not
necessarily better
than conventional.
Lack of gender
analysis was not a
problem for
achieving project
initial NRM goals,
but none of the
projects specifically
targeted women as
beneficiaries of
NRM work.

Continued.

97

Impact Assessment - Annual Report 2001

Table II-34.
results).

Summary of the main impacts of participation by stage and type (preliminary

Main impacts a

Stage

Testing

Technology and its
adoption

Human and
social capital

Feedback to
research

Costs of research

Important observed
or potential impacts
in all cases.

Very high human
capital impacts in
collaborative, low
impact in
consultative.

Impacts observed
within and beyond
the projects.

Recurring costs of
participatory trials not
significantly different from
conventional on-farm
trials. Costs increased with
collaborative aspects such
as workshops, rather than
with actual trial costs.

Collaborative is
better than
consultative in
terms of achieving
impact.

Lower impact of
testing activities
on social capital,
although may be
due to nature of
resource or
technology.
Not including
women as
participants in
collaborative
testing deprived
them of human
capital benefits.

Significant impacts
observed.

Additional costs regarding
training and data analysis.
However, these are onetime costs that occurred
because PR methods are
new.

Collaborative not
necessarily better
than consultative

Including women in
consultative testing did not
increase costs.

High impacts
observed on a
subset of nonrepresentative
participants.

High impact
regarding recognition
of importance of
skills and knowledge
as key
complementary
inputs issues, less on
farmer-to-farmer
dissemination.

Short-run costs increased
slightly, but overall cost
effectiveness also
increased.

No strong support
for importance of
gender
differentiation.

Diffusion

Impacts observed
from farmer input
to the methodology.

Gender
differentiation may
be important.
a.

NARS = National Agricultural Research Systems; NRM = natural resource management, and PR =
participatory research.

Donors List – PRGA (Ford Foundation, BMZ)
Staff List – Nancy Johnson and James Garcia
Collaborators : Nina Lilja, Margaret McKee and Jacqueline Ashby (PRGA);
Elske van de Fliert, CIP; Sieglinde Snapp, Joseph Rusike, Steve Twomlow, ICRISAT;
Roland Bunch and Gabino Lopez, COSECHA

98

Impact Assessment - Annual Report 2001

References Cited in Report
Biggs S.D, Farrington J. 1991. Agricultural research and the rural poor: A review of
social science analysis. International Development Research Center (IDRC), Ottawa,
Canada. 139 p.
Braun A.R, van de Fliert E. 1997. The farmer field school approach to IPM and ICM in
Indonesia: user participation. In: Local research and development: institutionalizing
innovations in root crop agriculture. Users’ Perspectives with Agricultural Research
and Development (UPWARD), Manila, Philippines.
p 44-64.
Braun A.R, Thiele G, Fernández M. 2000. Farmer field schools and local agricultural
research committees: complementary platforms for integrated decision making in
sustainable agriculture. AgREN Newsletter, Overseas Development Institute, July.
von Braun J. 1988. Effects of technological change in agriculture on food consumption
and nutrition: rice in a West African setting. World Develop 16:1083-98.
Bruce J. 1989. Homes divided. World Develop 17: 979-92.
Bunch R. 1982. Two ears of corn: a guide to people-centered agricultural development.
World Neighbors, Oklahoma City.
Bunch R. 1988. World Neighbours program profile. Guinope Integrated Agricultural
Development Project, progress report from 31 Jly 1987 to 30 Jne 1988. World
Neighbours. 4 p.
Bunch R. 1990. Low input soil restoration in Honduras: the Cantarranas farmer-tofarmer extension programme. International Institute for Environment and
Development (IIED) Gatekeeper series No. 23. IIED, UK. 12 p
Bunch R, Lopez G. 1995. Soil recuperation in Central America: sustaining innovation
after intervention. International Institute for Environment and Development (IIED)
Gatekeeper series No. 55. IIED, UK. 18 p.
Bunch R, Lopez G. 1999. Soil recuperation in Central America: how innovation was
sustained after project intervention. In: Hinchcliffe F, Thompson J, Pretty JN, Guijt I,
Shah P, eds. Fertile ground: the impacts of participatory watershed management.
Intermediate Technol Publ, London. p 32-41.
CIMMYT (Centro Internacional de Mejoramiento de Maíz y Trigo). 2000. Farmers’
voices are heard here. 4 p brochure, contact m.banzinger@cgiar.org.

99

Impact Assessment - Annual Report 2001

Dvorak K.A. 1996. Catalogo de prácticas y proyectos de conservación de suelos en
Centro America. Internal Report, Centro Internacional de Agricultura Tropical
(CIAT), Tegucigalpa, Honduras.
EAP (Escuela Agrícola Panamericana). 1995. Resultados del sondeo sobre las
comunidades dentro de la cuenca del río Yeguare. Reporte del Departamento de
Desarrollo Rural, Sección de Extensión, EAP, Zamorano, Honduras. 10 p.
Escalón J.F. 2001. Percepciones del trabajo de campo. Estudio de impacto del proyecto
integrado de desarrollo de Vecinos Mundiales. Report to Participatory Research and
Gender Analysis (PRGA) systemwide program. Available from N Lilja, PRGA. 8 p.
van de Fliert E. 1995. Progress report to Users’ Perspectives with Agricultural Research
and Development (UPWARD) II (for period covering Feb-Oct 1995) on project
identification of integrated pest management (IPM) and IPM training needs for sweet
potato in East and Central Java, submitted by Centro Internacional de la Papa (CIP)Yayasan Mitra Tani-Malang Research Institute for Food Crops (MARIF).
van de Fliert E. 1996a. Progress report to Users’ Perspectives with Agricultural
Research and Development (UPWARD) III (for period covering Nov 1995 – Apr
1996) on project identification of integrated pest management (IPM) and IPM training
needs for sweet potato in East and Central Java, submitted by Centro Internacional de
la Papa (CIP)-Yayasan Mitra Tani- Malang Research Institute for Food Crops
(MARIF).
van de Fliert E. 1996b. Progress report to Users’ Perspectives with Agricultural
Research and Development (UPWARD) IV (for period covering May-Oct 1996) on
project identification of integrated pest management (IPM) and IPM training needs
for sweet potato in East and Central Java, submitted by Centro Internacional de la
Papa (CIP)-Yayasan Mitra Tani- Malang Research Institute for Food Crops (MARIF).
van de Fliert E. 1997a. Progress report to Users’ Perspectives with Agricultural
Research and Development (UPWARD) V (for period covering Nov 1996 – May
1997) on project identification of integrated pest management (IPM) and IPM training
needs for sweet potato in East and Central Java, submitted by Centro Internacional de
la Papa (CIP)-Yayasan Mitra Tani- Malang Research Institute for Food Crops
(MARIF).
van de Fliert E. 1997b. Progress report to Users’ Perspectives with Agricultural
Research and Development (UPWARD) Terminal (for period covering Nov 1994 –
Oct 1997) on project identificatio n of integrated pest management (IPM) and IPM
training needs for sweet potato in East and Central Java, submitted by Centro
Internacional de la Papa (CIP)-Yayasan Mitra Tani- Malang Research Institute for
Food Crops (MARIF).

100

Impact Assessment - Annual Report 2001

van de Fliert E, Asmunati R. 1995. Progress report to Users’ Perspectives with
Agricultural Research and Development (UPWARD) I (for period covering 1994-95)
on project identification of integrated pest management (IPM) and IPM training needs
for sweet potato in East and Central Java, submitted by Centro Internacional de la
Papa (CIP)-Yayasan Mitra Tani- Malang Research Institute for Food Crops (MARIF).
24 p.
van de Fliert E, Braun AR. 2001. Conceptualizing integrative farmer participatory
research for sustainable agriculture: from opportunities to impact. Agric Human
Values, forthcoming.
van de Fliert E, Braun AR, Asmunati R, Wiyanto, Widodo Y. 1996. One step back,
two steps forward: sweet potato integrated crop management development in
Indonesia. Paper presented at the Users’ Perspectives with Agricultural Research and
Development (UPWARD) 5th Annual Conference, Institutionalizing Innovations in
Root Crops and Research and Development, 8-12 Dec, Papanga, Philippines.
van de Fliert E, Johnson N, Asmunati R, Wiyanto. 2001. Beyond higher yields: the
impact of sweet potato ICM farmer field schools in Indonesia. Centro Internacional
de la Papa (CIP) Program report. CIP, Peru.
Folbre N. 1986. Hearts and spades: paradigms of household economics. World Develop
14: 245-55.
Freeman H. 2000. A farmer participatory research methodology comparison: case
studies in Malawi and Zimbabwe. Mimeo available from author, International Crops
Research Institute for the Semi-Arid Tropics (ICRISAT), PO Box 39063, Nairobi,
Kenya. 39 p.
Gego H.A.M. 1999. Report on visit to Flores, 4-19 Sept 1999. Internal document
available from Centro Internacional de la Papa (CIP)- Users’ Perspectives with
Agricultural Research and Development (UPWARD), Indonesia.
Guyer J. 1988. Household budgets and women’s incomes. African Studies Center
Working Paper No. 28. Boston University, Boston, MA.
Haddad L, Hoddinott J, Alderman H. 1994. Intrahousehold resource allocation: an
introduction. Policy Research Working Paper No. 1255. World Bank, WA.
Heinrich G.M, Rusike J, Twomlow S. 2001. Farmer participatory research on
integrated soil water and nutrient management at ICRISAT-Bulawayo: approaches
used, lessons learned and directions for the future. Paper presented at Integrated
Natural Resource Management Workshop, 28-31 Aug 2001, Centro Internacional de
Agricultura Tropical (CIAT), Cali, Colombia.

101

Impact Assessment - Annual Report 2001

Hellin J. 1999. Soil and water conservation in Honduras: addressing whose reality?
AgREN Newsletter, July, Overseas Development Institute (ODI), UK.
Hellin J, Larrea S. 1998. Ecological and socioeconomic reasons for adoption and
adaptation of live barriers in Guinope, Honduras. Advances in Geo-Ecology 31:13831388.
Hinchcliffe F, Thompson J, Pretty J.N, Guijt I, Shah P, eds . 1999. Fertile ground: the
impacts of participatory watershed management. Intermediate Technol Publ, London.
ICRISAT (International Crops Research Institute for the Semi-Arid Tropics). 1999.
Methodology to develop practical soil fertility technologies through farmer/researcher
partnerships. Proposal submitted to the Rockefeller Foundation. ICRISAT, India.
ICRISAT-MAI (International Crops Research Institute for the Semi-Arid Tropics-Ministry
of Agriculture and Irrigation). 2000. Cost effective soil fertility management options
for smallholder farmers in Malawi. ICRISAT and MAI, Malawi. 24 p.
Izquierdo V.A. 1994. Diagnostico participativo agrosocioeconomico de la aldea de
Lavanderos, municipio de Guinope, departmento del Paraíso. Tesis del Escuela
Agrícola Panamericana (EAP)- Zamorano, Honduras.
Johnson, N., N. Lilja and J. Ashby, 2001, “User participation and gender analysis in
participatory natural resource management: An analysis of the PRGA project
inventory,” CGIAR Program on Participatory Research and Gender Analysis for
Technology Development and Institutional Innovation, Working Paper No. 10, Cali:
CIAT.
Kanyama-Phiri G, Snapp S, Kamanga B, Wellard K. 2001. Towards integrated soil
fertility management in Malawi: incorporating participatory approaches in
agricultural research. Exper Agric, forthcoming.
Krishna A, Uphoff N, Esman M. 1997. Reasons for hope: instructive experiences in
rural development. Kumarian Press, USA. 324 p.
Larrea S. 1997. Experiences y lecciones de agricultores innovadores sobre el desarollo
rural: caso de Guinope Honduras. MSc Thesis. Escuela Agrícola Panamericana (EAP)Zamorano, Honduras. 94 p.
Lilja N, Ashby J.A. 1999a. Types of participatory research based on locus of decision
making. Participatory Research and Gender Analysis (PRGA) Working Document
No. 6. PRGA-Centro Internacional de Agricultura Tropical (CIAT), Cali, Colombia.
Lilja N, Ashby J.A. 1999b. Types of gender anlaysis in natural resource maangement
and plant breeding. Participatory Research and Gender Analysis (PRGA) Working

102

Impact Assessment - Annual Report 2001

Document No. 8. PRGA-Centro Internacional de Agr icultura Tropical (CIAT), Cali,
Colombia.
Lilja N, Ashby J.A, Sperling L (eds.). 2000. Assessing the Impact of Participatory
Research and Gender Analysis. PRGA-Centro Internacional de Agricultura Tropical
(CIAT), Cali, Colombia. 287 p.
Pretty J. 1995. Regenerating agriculture: policies and practices for sustainability and
self-reliance. Earthscan, UK. 320 p.
Rusike J. 2001. Economic analysis of agronomic trials (mimeo). Available from author
at International Crops Research Institute for the Semi-Arid Tropics (ICRISAT),
Bulawayo, Zimbabwe.
Rusike J, Twomlow S, Freeman H, Heinrich GM. 2001. Impact indicators for
comparing participatory research approaches to promote soil fertility in semi-arid
southern Africa. Paper presented at Integrated Natural Resource Management
Workshop, 28-31 Aug 2001, Centro Internacional de Agricultura Tropical (CIAT),
Cali, Colombia.
Scarborough VS, Killough DA, Johnson N, Farrington J. 1997. Farmer led extension:
concepts and practices. Overseas Development Institute (ODI) Technology
Publications, UK. 214 p.
Selener D, Chenier J, Zelaya R. 1997. Farmer to farmer extension: practical lessons
from Central America. International Institute for Rural Reconstruction (IIRR),
Ecuador.
Sherwood S, Larrea S. 2001. Looking back to see ahead; farmer lessons and
recommendations after 15 years of innovation and leadership in Guinope, Honduras.
Agric Human Values 18:195:208.
Smith K. 1994. The human farm, a tale of changing lives and changing lands. Kumarian
Press, USA. 144 p.
Snapp S.S. 1998. Soil nutrient status of smallholder farmers in Malawi. Commun Soil
Sci Plant Anal 29:2571-2588.
Snapp S.S, ed. 1999a. Improving soil management options for women farmers in Malawi
and Zimbabwe. Procs Ecoregional Workshop on Launching DFID-supported Project
“Will women farmers invest in improving their soil fertility management?
Participatory experimentation in a risky environment,” 17-19 May 1999, Bulawayo,
Zimbabwe. International Crops Research Institute for the Semi-Arid Tropics
(ICRISAT), Zimbabwe.

103

Impact Assessment - Annual Report 2001

Snapp S.S. 1999b. Mother and baby trials: a novel trial design being tried out in Malawi.
Target Newsletter of the Southern African Soil Fertility Network, Vol. 17,
No. 8.
Snapp S.S. 1999c. Report to Rockefeller Foundation on the project “Methodology to
Develop Practical Soil Fertility Technologies through Farmer/Researcher Partnerships”,
27 Oct 1999. International Crops Research Institute for the Semi-Arid Tropics
(ICRISAT), India.
Snapp S.S. 2000a. Draft of final report to Rockefeller Foundation on the project
“Methodology to Develop Practical Soil Fertility Technologies through
Farmer/Researcher Partnerships”, 5 Oct 2000. International Crops Research Institute
for the Semi-Arid Tropics (ICRISAT), India.
Snapp S.S. 2000b. Malawi Mother/Baby Trial Case Study; participatory approaches to
improving farmer-researcher-extension communication and soil productivity
technologies. Paper presented at the Systemwide Program on Participatory Research
and Gender Analysis for Technology Development and Institutiona l Innovation of the
CGIAR (PRGA) Seminar, Nov 2000, Nairobi, Kenya.
Snapp S.S, Phiri RH, Moyo A. 1998. Soil fertility experimentation by farmers and
researchers for drought-prone regions of Zimbabwe and Malawi. Procs Risk
Management Stakeholder Workshop, 1-3 Oct 1997, Kadoma, Zimbabwe. Centro
Internacional de Mejoramiento de Maíz y Trigo (CIMMYT) and International Crops
Research Institute for the Semi-Arid Tropics (ICRISAT), Zimbabwe.
Snapp S.S., Rohrbach D.D., Simtowe F, Freeman H.A. 2001. Sustainable soil
management options for Malawi: can smallholder farmers grow more legumes? Agric
Ecosys Environ, forthcoming.
Thiele G, van de Fliert E, Campilan D. 2001. Whatever happened to participatory
research at the International Potato Center? Agric Human Values, forthcoming.

104

Impact Assessment - Annual Report 2001

OUTPUT II: IMPACT OF PAST RESEARCH MONITORED

2.2 Evaluación ex-post: El cambio técnico en las sabanas de Colombia. - by:
L.Rivas
Meta 2001. Evaluar el impacto del cambio técnico logrado en el período 1967-1997
en el cultivo de arroz en Colombia.
Resultados sobresalientes
? Las estimaciones de los beneficios sociales señalan que si bien a nivel del conjunto del
país, el cambio técnico evaluado muestra excedentes económicos positivos tanto para los
productores como para los consumidores, al desagregar tales beneficios por región y por
sistema de producción, se puede apreciar que no todas las regiones del país, ni todos los
sistemas de producción, resultaron ganadores en el proceso de cambio técnico durante el
período evaluado.
? En las regiones arroceras donde los procesos de adopción tecnológica fueron más
dinámicos e intensos como el área Central del país y los Llanos Orientales, se generaron
grandes excedentes tecnológicos.
? En el Bajo Cauca, la Costa Norte, los Santanderes y la región Sur – Occidental, en
donde la reducción de los costos de producción fue inferior al 20%, el efecto de la caída
de los precios del arroz implicó una pérdida neta para esos productores.
? La totalidad de los beneficios de la modernización del cultivo de arroz en Colombia se
concentraron en los sistemas de riego y secano mecanizado. El secano manual al no
adoptar nuevas tecnologías resultó ser un perdedor neto.
? Los resultados sugieren que en los procesos de diseño de nuevas tecnologías
agropecuarias para un país o región, es importante que las opciones tecnológicas a
implementar tengan un de amplio espectro en cuanto a regiones geográficas, sistemas de
producción y tipo de productores. Un cambio técnico muy exitoso como el evaluado, no
necesariamente generó beneficios para todos los actores.
? Lo anterior implica que para buscar la equidad social a través del desarrollo técnico se
requiere tener un amplio abanico de posibilidades para todos los grupos de productores
involucrados. En particular para los productores de menores recursos.

Informe de avance
Introducción: Dentro de las actividades de evaluación del impacto del Convenio MADR
– CIAT se incluye la evaluación del impacto ya logrado por la adopción tecnológica en el
área de actuación del mencionado Convenio, la Amazonia y Orinoquia colombianas
(A&O).

105

Impact Assessment - Annual Report 2001

El cultivo de arroz en el país y en la región de interés es de gran importancia
socioeconómica por la magnitud del área sembrada, el nivel de producción alcanzado y
por constituirse en un alimento básico en muchas regiones de Colombia. Este cultivo es
líder en lo referente a adopción de nuevas tecnologías de producción, observándose un
rápido crecimiento de su productividad en el transcurso de las últimas décadas.
La gran dinámica del cultivo observada a nivel de país como un todo, también se aprecia
en la región objetivo, especialmente en la Orinoquia. Hacia 1981 el cultivo de arroz en
A&O ocupaba 69.3 mil hectáreas, 36% del área en cultivos en esa región, y generaba 298
mil toneladas de arroz paddy. Hacia fines de la década del 90 el área plantada se había
más que duplicado, situándose alrededor de 155 mil hectáreas para una producción total
de casi 730 mil toneladas (Rivas, 2000). El avance de los rendimientos en producción de
arroz en A&O, es un de los principales indicadores del progreso técnico en esa región de
Colombia. En efecto en el período 1967-1997 los rendimientos crecieron desde 1720 a
4273 kg/ha. (Rivas 2000 B).
La adopción tecnológica en arroz ha estado íntimamente relacionada con cambios
sustanciales en los sistemas de producción imperantes y en su distribución regional.
Hacia la primera mitad del siglo pasado, la producción arrocera del país se concentraba
en la Costa Norte (28%) y el Magdalena Medio (35%), en tanto que la producción de los
Llanos Orientales solo representaba el 6% del total del país (Scobie & Posada,1977). La
información del Ministerio de Agricultura para 1997 muestra una reducción sustancial de
la participación de la Costa Norte en la oferta nacional de arroz, 9.3%, y un avance muy
significativo de los Llanos Orientales, 31.2% (Cuadro 1).
Los estudios de Scobie & Posada (1977) y de Fedearroz (1997) documentan claramente
como el predominio de los sistemas de producción de secano que prevalecieron en el
escenario de la producción nacional de arroz hacia la primera mitad del siglo pasado, ha
declinado paulatinamente. En 1954 ellos representaban el 42% de la producción total, en
1975 esa proporción había caído a 9% y en 1997 solo llegó a 2%. Entre 1981 y 1997 el
área cultivada bajo el sistema de secano manual cayó de 200 a 28 mil hectáreas.
(Fedearroz, 1997).
El desarrollo tecnológico que aportó variedades de alto rendimiento que se utilizan
intensivamente en los sistemas de producción bajo riego y en condiciones de secano
mecanizado, implicó sustanciales avances en la producción, la productividad, y el
consumo en Colombia. La adopción tecnológica en el cultivo de arroz ha sido
permanente y muy dinámica y los indicadores del progreso técnico son el empleo de
insumos modernos como la semilla certificada y uso de equipos cada vez mas sofisticado
como tractores y cosechadoras. (Fedearroz, 1997)
Con base en lo anterior y dada la heterogeneidad de los sistemas de producción a través
de las regiones arroceras del país, se plantea que el impacto económico del cambio
técnico resulta de distinta magnitud e intensidad si se evalúa a nivel de región y de
sistema de producción.

106

Impact Assessment - Annual Report 2001

Impacto del Cambio Técnico en producción de arroz en Colombia.
Para evaluar el impacto ex - post de las nuevas tecnologías de producción de arroz en
Colombia se utilizó el Modelo DREAM (Wood y Baixt, 1998), el cual permite evaluar el
impacto tecnológico desagregando por región y sistema de producción.
Para tal evaluación se consideraron 6 regiones productoras y dos sistemas de producción.
Las regiones consideradas fueron: Central, Llanos Orientales y Amazonia, Bajo Cauca,
Costa Norte, Santanderes y Sur Occidente. Los sistemas de producción evaluados fueron
riego y secano mecanizado considerados como un solo sistema y el secano manual.

Cuadro 1. Distribución geográfica de la producción de arroz en Colombia
1967-1997
1967
1997
Producción
Producción
%
%
‘000 tm
‘000 tm
Central 1/
205.1
31.0
688.0
37.6
Llanos Orientales y Amazonia 2/
119.1
18.0
571.0
31.2
Bajo Cauca 3/
172.0
26.0
253.0
13.8
Costa Norte 4/
92.6
14.0
170.0
9.3
Santanderes 5/
33.1
5.0
104.0
5.7
Sur Occidente 6/
39.7
6.0
44.0
2.4
Total
661.5
100.0
1830.0
100.0
1/ Central: Huila, Tolima, Caldas, Cundinamarca y Boyacá. 2/ Meta, Casanare, Arauca, Caquetá,
Amazonas, Putumayo y Vichada. 3/ Antioquia, Bolívar, Córdoba y Sucre. 4/ Cesar, Guajira y Magdalena.
5/ Santander y Norte de Santander. 6/ Cauca, Valle, Nariño y Chocó.
Fuente: Cifras del Ministerio de Agricultura de Colombia, agrupadas según la regionalización de la
producción aportada por Fedearroz (1997).
Región

En la región Central, la más importante en cuanto a volumen de producción, ella se
concentra en los departamentos del Tolima y Huila, en los cuales se observan altos
niveles de desarrollo tecnológico y en donde predomina el sistema de riego. En los
Llanos Orientales, la segunda región en importancia, el grueso de la producción se genera
en los departamentos del Meta, Casanare y Arauca. En el primero de ellos utilizan
predominantemente el riego y el secano mecanizado, en el Casanare es más frecuente el
sistema de riego y en Arauca el secano mecanizado (Fedearroz, 1997).
El bajo Cauca, que incluye los departamentos de Antioquia, Bolívar, Córdoba y Sucre,
presenta niveles tecnológicos relativamente bajos, debido al predominio del sistema de
secano manual. En la Costa Norte del país una alta proporción de su producción se
genera en condiciones de riego. Los Santanderes y el Sur Occidente son áreas arroceras
relativamente marginales, que en conjunto no superan el 10% de la producción nacional.

107

Impact Assessment - Annual Report 2001

Supuestos técnicos y Económicos
En el Cuadro 2 se incluyen los supuestos económicos y técnicos utilizados para evaluar el
impacto económico de la adopción tecnológica en el cultivo de arroz en Colombia en el
período 1967-1997. Esta información se refiere a las condiciones de equilibrio inicial del
mercado de arroz en las diferentes áreas arroceras del país en términos de producción y
precios. También las elasticidades y el crecimiento autónomo tanto de la oferta como de
la demanda y los factores de desplazamiento de la función de oferta originadas en el uso
de las nuevas tecnologías. Estos supuestos son resultantes de la revisión de numerosos
estudios sobre la industria arrocera del país, entre otros: Scobie &Posada, 1977; Montes
et al, 1980; Gutiérrez y Herford,1974; Sanint L.R., 1992 y los Anuarios Estadísticos del
Ministerio de Agricultura de Colombia de varios años.

Resultados
Los resultados de la evaluación se expresan en términos de los beneficios sociales
recibidos por consumidores y productores y que se originan en el crecimiento de la
productividad y en la reducción de los precios reales en los mercados de arroz.
Las estimaciones señalan que si bien a nivel del conjunto del país, el cambio técnico
evaluado muestra excedentes económicos positivos tanto para los productores como para
los consumidores, al desagregar tales beneficios por región y por sistema de producción,
se puede apreciar que no todas las regiones del país, ni todos los sistemas de producción,
resultaron ganadores en el proceso de cambio técnico durante el período evaluado.
En las regiones arroceras donde los procesos de adopción tecnológica fueron más
dinámicos e intensos como el área Central del país y los Llanos Orientales, se generaron
grandes excedentes tecnológicos. En ellas los costos de producción unitarios se redujeron
en una proporción superior al 80%. Por el contrario, en las áreas más marginales y en
donde la magnitud e intensidad del camb io técnico fue significativamente menor, se
observan excedentes negativos para los productores. En el Bajo Cauca, la Costa Norte,
los Santanderes y la región Sur – Occidental, en donde la reducción de los costos fue
inferior al 20%, el efecto de la reducción de los precios del arroz implicó una pérdida
neta para esos productores. Por ejemplo en Bogotá, el principal centro consumidor del
país, el precio real al consumidor de arroz se redujo en casi una tercera parte en el
período de análisis (Figura 1).

108

Impact Assessment - Annual Report 2001

Figura 1. Precios reales de arroz al consumidor en Bogotá: 1967-1997
$ de 1990/kg

400.0
350.0
300.0
250.0

y = -2.0394x + 272.86

200.0

2

R = 0.2754

150.0
100.0
50.0
0.0
1967

1970

1973

1976

1979

1982

1985

1988

1991

1994

1997

Precios de arroz de primera, deflactados por el IPC 1990=100
Fuente: cálculos basados en cifras del Banco de la República y el DANE.

Al evaluar el impacto del cambio técnico sobre los diferentes sistemas de producción se
consideraron dos tipos: 1) Riego y secano mecanizado, considerados como un solo
sistema y 2) Secano manual. Ellos presentan un acentuado contraste a lo largo de un
período de casi tres décadas en términos de producción, áreas cultivadas y productividad.
El secano manual tiende a desaparecer al contraerse drásticamente tanto sus áreas
sembradas como el volumen de producción aportado por este sistema, el cual ha sido
relegado cada vez más hacia área más marginales y de menor aptitud para la agricultura.
A comienzos de la década de los 60 el 56% del área arrocera de Colombia utilizaba este
sistema de producción. Hacia fines de los 90 solo el 2% del área cultivada con éste cereal
utilizaba el secano manual (Cuadro 3 )
Cuadro 3

Distribución del área y la producción de arroz en

Colombia, según sistema de producción: 1961-1997
Distribución del área (%)
Distribución de la Producción (%)
Secano
Riego & secano
Secano
Riego & secano
manual
mecanizado
manual
mecanizado
1961
55.7
44.3
42.3
57.7
1997
6.8
93.2
2.2
97.8
Fuente: Cálculos basados en cifras de Fedearroz., 1997.
Año

109

Rendimiento (tm/ha)
Secano
Riego & secano
Manual
mecanizado
1.5
2.6
1.4
4.5

Impact Assessment - Annual Report 2001

Cuadro 2 Coeficientes Técnicos e Impacto del cambio tecnológico en la producción de arroz en Colombia,
según regiones
1.1

Valor presente de los beneficios sociales: 1967-1997

Región 1/

Equilibrio Inicial 2/

Crecimiento autónomo
(%)

Elasticidades precio

Desplazamiento de la
oferta (%)

Valor presente de los
beneficios tecnológicos
us$ millones (i=10%)

Cantidad

Precio

‘000 tm

‘000$/tm

Oferta

Región Central
Llanos Orientales
Bajo Cauca
Costa Norte
Santanderes
Región Sur - occidental

205.1
119.1
172.0
92.6
33.1
39.6

584.6
485.3
641.2
594.2
540.7
583.3

0.80
0.80
0.80
0.80
0.80
0.80

Total Colombia

661.5

592.7

0.80

Demanda

Oferta

Demanda

3.0
4.0
1.0
1.5
2.9
2.5
-0.50

2.6

2.7

Horizontal

Vertical

3/

4/

73.0
68.0
12.2
11.7
10.8
0.6

91.3
85.4
15.2
14.7
13.5
0.8

176.3

220.9

Productor

Consumidor

556.2
231.9
-84.3
-53.4
-25.3
-44.1
580.7
873.5
Total : 1454.2

1/ Las regiones consideradas se definen así: Región Central: Huila, Tolima, Caldas, Cundinamarca y Boyacá. Llanos Orientales: Meta, Casanare, Arauca,
Caquetá, Putumayo y Vichada. Bajo Cauca: Antioquia, Bolívar, Córdoba y Sucre. Costa Norte : Cesar, Guajira y Magdalena. Santanderes: Santander y Norte
de Santander. Sur Occidente : Cauca, Valle, Nariño y Chocó. 2/ Cantidades y precios de equilibrio del mercado de arroz en el año inicial de evaluación. Los
precios se expresan en pesos de 1997. 3/ Aumento porcentual de la producción debido al cambio técnico. 4/ Reducción porcentual de los costos unitarios con
respecto a la tecnología tradicional..

110

Impact Assessment - Annual Report 2001

La estimación de los beneficios sociales al productor, según sistema de producción,
revela que todos los beneficios se concentraron en el sistema de riego & secano
mecanizado, en el cual el proceso de adopción de nuevas tecnologías fue intenso y de
carácter masivo, lo cual posibilito la expansión de sus áreas cultivadas, en desmedro del
secano manual para el cual la oferta de nuevas tecnologías prácticamente fue inexistente.
La figura 2 muestra el valor presente de los beneficios tecnológicos capturados por cada
sistema de producción.
Figura 2. Valor presente de los beneficios del cambio técnico al productor de arroz
en Colombia, según sistema de producción:1967-1997 (i=10%)

1000
800
600
400
200
0
-200
-400
Riego & secano mecanizado

Secano manual

Excedente total

Los resultados obtenidos sugieren que en los procesos de diseño de nuevas tecnologías
agropecuarias para un país o región, es importante que las opciones tecnológicas a
implementar tengan un de amplio espectro en cuanto a regiones geográficas. sistemas de
producción y tipo de productores. Un cambio técnico muy exitoso, como el de arroz en
Colombia, no necesariamente generó beneficios para todos los actores. Esto implica que
para buscar la equidad social a través del desarrollo técnico, se requiere tener un amplio
abanico de posibilidades para todos los grupos de productores involucrados.
Actualmente los productores de arroz de secano corresponden al grupo pequeños
agricultores de áreas marginales que viven en condiciones de subsistencia. La falta de
desarrollo técnico para ese sector ha implicado pérdida de competitividad y una
reducción dramática de su participación en el mercado del cereal. Si se considera que
ese segmento de los productores de arroz es huérfano debido a la falta de nuevas
alternativas tecnológicas, se plantean varias posibilidades. a)Se debe efectuar un esfuerzo
importante para dotar a ese sector con nuevas tecnologías que mejoren su competitividad
b) Se deben encontrar otras opciones de producción diferentes al arroz para estos
productores y enfatizar la investigación de arroz en los sistemas bajo riego. La respuesta
no es fácil, no existe consenso entre los investigadores y planificadores sobre las posibles
111

Impact Assessment - Annual Report 2001

líneas de acción. Algunos argumentan que las posibilidades del arroz de secano son muy
limitadas y que en consecuencia todo el énfasis debe colocarse en los sistemas de riego y
de secano favorecido, en tanto que otros investigadores consideran que el potencial de los
sistemas de secano ha sido subestimado.
Independiente de lo anterior, lo más práctico parece ser, que el portafolio de investigación
sea lo suficientemente amplio en términos de cultivos y sistemas de producción, para
permitir un proceso de modernización de la agricultura en el cual todos los actores y en
particular los de menores recursos, en los diferentes sistemas y regiones geográficas,
tengan la posibilidad de involucrarse en el proceso y capturar beneficios del mismo.
Donante: Ministerio de Agricultura de Colombia. Convenio de Cooperación Técnica y
Científica MADR - CIAT
Referencias
Federación Nacional de Arroceros – Fedearroz (1997). El arroz en Colombia; 19801997, Bogotá, Colombia.
Gutiérrez Alemán Néstor y Reed Hertford (1974). Una Evaluación de las
Intervenciones del Gobierno en el Mercado de Arroz en Colombia, Centro
Internacional de Agricultura Tropical, CIAT, Folleto Técnico No. 4, Cali,
Colombia.
Ministerio de Agricultura y Desarrollo Rural, Anuario Estadístico, varios años,
Bogotá, Colombia.
Montes Gabriel, Ricardo Candelo y Ana Milena Muñoz de Gaviria (1980). La
Economía del Arroz en Colombia. Departamento Nacional de Planeación, Unidad
de Estudios Agrarios, en: Revista de Planeación y Desarrollo, Vol. XIII, No1, Enero
– Abril.
Rivas Libardo (2000). Desarrollo Tecnológico y Crecimiento en la Orinoquia Amazonia
colombianas: Pasado, presente y futuro. Tendencias socioeconómicas y Evaluación
ex – ante del cambio técnico. Proyecto de Evaluación de Impacto, CIAT, Octubre.
Rivas Libardo (2000b). Evaluación del Impacto ex –post de la adopción de nuevas
tecnologías : Forrajes y arroz en la Amazonia y Orinoquia de Colombia. Proyecto
de Evaluación de Impacto, CIAT, Junio.
Scobie Grant M. and Rafael Posada T. (1977). The Impact of high – yielding rice
varieties in Latin America, with special emphasis in Colombia, CIAT, April.
Sanint Luís R. (1992). New Rice technologies for Latin America: Social benefits, past
reminiscences and issues for the future. Trends in CIAT Commodities 1992. CIAT,
Working Document No 111, June.

112

Impact Assessment - Annual Report 2001

2.3 Economic Evaluation of Agricultural R&D for the Savanna Ecosystems of
Latin America: The Case of Maize in Meta Department, Colombia – by L.
Mosquera
Introduction
In recent years, with an ever increasing demand for food from a growing population,
there have been pressures to expand the agricultural frontier as a means of enhancing
food production. Two geographically large areas for potential expansion are the forest
and savanna regions in the humid tropics. However expansion into forest areas has many
negative environmental consequences – reduced biodiversity and wildlife habitat reduced
carbon storage, and often, significant disruption of the hydrological cycle for downstream
water uses. Thus, expansion of agriculture into the savanna areas has long been an
agricultural development goal. However, despite the great potential for these very large
tracts of land, particularly in Latin America and Africa, it has proved extremely difficult
to generate new technologies that can provide environmentally and economically
sustainable production because of the poor quality of most tropical savanna soils.
Tropical savanna soil has been classified as highly acid with high levels of aluminium
toxicity and, generally speaking, infertile (Sanchez and Salinas, 1983).
For several years the Instituto Colombiano Agripecuario (ICA), the International Centre
for Tropical Agriculture (CIAT), and the International Centre for Wheat and Maize in
Mexico (CIMMYT), have been working on livestock-based production systems for
savanna soils. This has included breeding grasses tolerant to high aluminium saturation
and the development of crop varieties that are more tolerant to special conditions
encountered in the savannas. The latter approach has been tried in four regions since
1980: Eastern Africa, South and Southeast Asia, the Andes and Central America. And
CIMMYT has maintained a presence in the Andean region since late 1977 for cooperative research work in Colombia, Ecuador, Peru, and Bolivia in the development of
new maize varieties fo r acid soil.
Genetically, maize is more susceptible than rice to aluminium toxicity, and this
complicates the crop improvement – breeding – process. Late in the 1980 a co-operative
program started (ICA, CIMMYT and CIAT) to develop the variety SIKUANI V-110,
which is tolerant roughly to 55% aluminium saturation, and provides yields of about 3.0
ton/ha under those conditions and about 7.5 ton/ha in normal fertile soils.
Maize has a comparative advantage over other species as a component in savanna
cropping systems for the following reasons:
1) It does not need high levels of mechanisation for sowing and harvest
2) It is less demanding in labour
3) It does not need industrial processing for direct human or animal consumption
4) Cultural practices may be performed manually or with reduced mechanisation.
Improved maize technologies could, thus, potentially lead to the improvement of the
Eastern Savanna (Llanos Orientales) of Colombia for crop and animal production, reduce

113

Impact Assessment - Annual Report 2001

expansion pressure into Amazonian forest areas, and help in the long-run recovery of
these soils. With new maize varieties the inherent acidity, and infertility of soils in the
Llanos Orientales becomes less problematic for rural households.
The variety SIKUANI V-110 was released in 1994 by the Corporación Colombiana de
Investigación Agropecuaria, Regional 8 (CORPOICA), based in Villavicencio – but
including places such as La Libertad, Carimagua and Meta, where the most of the
research was carried out.
The research undertaken in the Llanos Orientale s focuses on sustainable production
system in maize, to meet increased human food and animal feed demands, to develop
some comparative advantage for farmers in that region, and to generate technologies
compatible with the fragility of this particular agro-ecosystem.
The origin of the SIKUANI V-110 variety lies in the 192 materials evaluated under
aluminium toxicity conditions (with 45 to 75% of saturation). The best material
populations were chosen from these trials for subsequent breeding and further
improvement and only after thirteen year of tests and trials in such acid soils was the
variety considered ready for release.
The evaluation trials were conducted in Colombia, Thailand, Indonesia, Peru, Venezuela
and Brazil, and the representative yield was around 3.0 ton/ha. The results presented by
researchers show that this variety has potential in other regions in the world with similar
soils to those of the Colombian savannas (saturation in aluminium 55%, acid, and low
availability of phosphorus).
However, these results were only achieved through the combined efforts of research
institutes (ICA, CORPOICA, CIMMYT AND CIAT) over 17 years.
Background and justification
Colombia is the third largest maize producer in South America, after Brazil and
Argentina. Maize plays an important role in the diet of the Colombian population. Great
efforts are therefore made to maintain and increase production steadily, both by better
cultural methods and by the use of more suitable varieties and hybrids. All of these
productivity increasing activities require investments in agricultural research – research
targeted to improving yields, reducing crops losses, minimising the need for cultivation in
fragile soils, improving the processing and nutritional value of the grain, and so on.
However, it is important to ensure that limited public research and development (R&D)
dollars are targeted to the mix of commodities, geographical areas, production
technologies and, often, social groups that can best maximise social welfare.
Greater demands for accountability, shrinking research resources, and the growing
complexities of research goals are focusing the attention of R&D managers on the need
for improved R&D evaluation and decision making methods. Furthermore, the
increasingly competitive nature of R&D funding is accelerating the search for areas of

114

Impact Assessment - Annual Report 2001

comparative advantage or complementarily in research – in order to select appropriate
R&D partners and strategically important research themes.
Methods for evaluating the direct production effects of technology, such as those
involving increased genetic potential or improved efficiency in the use of inputs such as
seed, labour, machinery, and fertiliser are fairly well established and increasingly applied.
However, when the research must include productivity changes arising from the
interaction between new technologies, production systems, and natural resources – such
as poor savanna soils, these methods are not effective. There is an urgent need to
systematically extend the R&D eva luation framework in order to encompass this natural
resource quality issue.
As previously discussed, a major development issue in the extensive humid tropical
savannas of Latin America is the difficulty of managing the highly weathered soils in a
sustainable way. Even modest attempts at increasing livestock and crop productivity must
be undertaken with great care if a rapid decline in (already low) soil fertility is to be
avoided. Furthermore, new technologies should not only prevent soil productivity losses
in a cost-effective manner but, ideally, should also help to increase the inherent
productive capacity of soil over the long run. As has been demonstrated in the Brazilian
Cerrados, there are potentially large payoffs from bringing savannas into more intensive
agricultural production. In Colombia, however, the savanna (or Llanos) area has been
subjected to much land speculation associated primarily with the laundering of drug
money, and land prices have been artificially inflated relative to their agric ultural
opportunity cost. At the same time, continued guerrilla operations and high interest rates
have done little to foster positive attitudes to long-term land-enhancing investments. It is
important, therefore, that realistic analyses are made of the potential economic
attractiveness of new agricultural technologies targeted to such areas.
This study is concerned with the development of improved methods to assist in the ex
ante economic evaluation of the likely outcomes of research and extension investments
into new maize technologies target to the Llanos conditions.
Objectives of the Research
The principal objective of this thesis is to apply new methods of economic evaluation to
agricultural production systems that incorporate technologies designed to simultaneously
enhance crop productivity and soil resource conditions. If such an approach can be
validated it can serve not only to assess the likely impact of technologies currently under
development for the Colombian Llanos, but also as a general means of testing a range of
strategic research policy and technology design options. The approach could help to
design research that has greater research benefits, and alter the distribution of benefits in
ways seen more socially acceptable (for example, by increasing the share of benefits
going to poor rural households).
Specific Objectives
1. To describe the existing production systems in the Meta department with regard to the
production of maize.

115

Impact Assessment - Annual Report 2001

2. To analyse the trends in prices, production costs and technology adoption
(specifically in technical and traditional maize production systems) in Meta
department.
3. Make an assessment of the size and distribution of economic benefits by municipios,
by producers and by consumers.
Next to the fact analysis regarding to these specific objectives the following questions
were formulated in order to address points 2 and 3:
1. What are the likely socio-economic benefits of maize research for the Llanos (Meta)?
What are the impacts on modern and traditional maize growing producers and maize
consumers? How the impacts vary across municipios?
2. Assuming no technical change and falling world commodity prices can Llanos crop
producers remain competitive? How much would the sector shrink in the next 25
years if no new technologies were introduced?
The Statement of the Problem
Accelerated loss of soil productivity under cultivation
The development problem faced is that while the savannas are extremely extensive their
agricultural use is limited, primarily because of poor soil productivity. Not only are the
savanna soils of relatively low inherent fertility under natural conditions, but even those
low levels degrade relatively rapidly under cultivation (typically the soils are
unproductive after 3-5 years of cultivation). Furthermore, the soils are relatively poorly
drained, and in the rainy season many areas are difficult to access. These biophysical
limitations have, in turn, provided little incentive for systematic, sustained investment in
infrastructure in the savanna. Most economic exploitation has been associated with lowproductivity extensive livestock operations. Investment in agricultural research targeted
to the savannas has been made in the expectation that more intensive, sustainable
production systems can be developed. There is even a hope that some of these production
systems could bring about significant long-term increases in the productive capacity of
savanna soils (e.g., by building up organic matter, and improving the soil’s physical and
biochemical properties).
The Llanos Orientales are located in a humid tropic ecosystem, with an annual
precipitation in excess of 2100 mm. and an average temperature of 26o C. Those
environmental conditions can induce rapid soil degradation in the kind of agroecosystems under development for intensive agriculture in the region. Soil degradation
drastically reduces the profitability of production systems with low crop yields, low
response to fertiliser application, low seed germination, fungus infections, soil erosion
and more labor costs, among other things.
The rotation systems recommended in the region, most incorporating green manure, are
maize-soybean, rice-soybean, rice-cotton, maize-cotton and also rotations with sorghum

116

Impact Assessment - Annual Report 2001

during the second semester. These types of rotations requir e careful management. It has
been reported that using legumes (as a green manure) in association with maize crops
increases soil organic matter, the micro-organism and the availability of macro and
micronutrients, decreases aluminium (Al), increases pH and helps to aggregate the soil in
ways that improve water holding capacity, ventilation and electric conductivity, and the
regulation of soil temperature (Lal, 1995, mentioned by Luna and Rodas 1996).
Designing of Scenarios
Simulating the process generation, adoption and economic impact of the technology, it’s
necessary to know the context in which the technology is developed and adopted. The
likely way how R&D impacts to the society is related with a lengthy range of factors –
many of them not related with R&D products themselves. Among these factors could be
included the macro approaches, such as basic economic and the demography structure of
the regions, as well fiscal policies, agricultural and commercial – domestic and
international. This includes a number of factors that influence the capacity and
availability of the producer to adopt new technologies or practices. These factors are
often directly related with expectations about the input markets and agricultural outputs.
The ex ante analysis of the research economic impacts, has been implemented in the
software DREAM (Wood & Baitx, 1998), which is a temporal simulation of the
generation and adoption of new technology as well its subsequent impact on markets and
social welfare. Besides this, the DREAM model can explicitly take into account other
factors, such as taxes or subsidies to production and consumption, and also production
and consumption growth which arise from sources other than technological change.
Those factors, as well as the producers and consumers response to price changes
(elasticities), are parameters of the model over which the users have control.
To develop scenarios it is necessary to consider the strategic framework in which the
R&D investment is carried out. This includes the institutional objectives or goals to be
reached by the R&D, i.e. increasing agricultural productivity and rural incomes, food
security improvement, or minimisation of negative impacts on the environment.
Other information to formulate scenarios is related to the scale and purpose of the
evaluation. There are two dimensions in the scale – the first one is connected with the
geographical target and the another one is concern with the split between research
components, i.e. national programs, projects and even sub-projects. These factors are
critical in determining the level of detail and complexity required for the analytical
representation.
As concrete example of the factors that influence in the design of scenarios:
• Agricultural price policies – often expressed as taxes and subsidies related to
production and consumption. This factor can change in the time.

117

Impact Assessment - Annual Report 2001

•

Development factors that promote changes in the agricultural product supply – like
possible investments in rural infrastructure, credit systems or better peasant
organisation models.

•

Factors that stimulate demand – population and real income growth, changes in the
consumer’s taste and new uses for agricultural commodities.

•

Heterogeneity of production and consumption in relation to research objectives. It is
possible to estimate what would be the technological impacts in specific social groups
– small producers, hillside producers, rural and urban poor consumers etc. It is
necessary to have appropriate groups for the disaggregation of regional data. The
framework allows for different definitions for each group, but the analysis treats the
groups simultaneously giving the possibility to observe not only the impact on the
group, or but also how it would change the equilibrium among groups according to
the different patterns of research investment.

In this case a number of regions of market regions on municipalities were defined; and it
is possible to evaluate the simultaneous effects of the new technologies impact adopted in
one or all regions. For the analysis it is necessary to assume that the total production
across all regions is equal to the consumption in all regions, but it is not necessary that
production and consumption are in equilibrium in each region. The model computes the
annual prices for each region considering the conditions of all market simultaneously
(endogenous).
The results of the evaluation have several elements and many ways to display them. The
detailed results are important for the scientists who explore alternative strategies of
implementation for a given research theme.
Data
The information on which this study is based comes from primary and secondary sources.
Primary data come from several sets of “long-term” (3-5 year) experimental field data
and on- farm trials involving different crop, crop rotations, and management practices in
the Llanos. The experiments monitored changes in crop yields over time as well as
recording the simultaneous changes in a range of soil properties. These benchmark data
help to establish the economic impacts of new production systems in an experimental
setting. The on-farm trials and municipal production data provide additional information
on the extent to which producers are actually adopting new technologies given the overall
sector trends and the resource and market constraints that producers face. These data have
been collected over the last 10 years by CIAT in collaboration with the Colombian
national research agency CORPOICA (formerly, until 1993, known as ICA), the
International Fertiliser Development Centre (IFDC) and others.
Secondary information for the Meta province was collected in Bogotá and Villevicencio
includes; production, area, yield, and prices, as well as census data on population and
farm size. As far as possible these data were collected at the municipal level (of which

118

Impact Assessment - Annual Report 2001

there are 24 in Meta). Additional information included detail on wholesaler prices at
Villevicencio and Bogotá the major markets for agricultural trade to and from Meta
Department. In the case of maize, the region is a net importer to meet the needs of both
human and animal consumption
In terms of the market data, we have used the municipio as the basic economic unit of
analysis. Maize production data was collected at the municipio. Only two demand
(consumption) regions were defined, one to represent the current and projected human
consumption in the Meta department, and the other to represent the current and projected
demand for animal feed. Another region, external to the Meta was defined as a
production source for the net import of maize currently required by the department. A
preliminary list of the specific production and consumption regions defined for the
simulation is shown in table 1.
Maize production data is available for two maize production systems. The new
technologies being tested in the Carimagua experiments are “technical” (improved) maize
varieties. But many producers use long established, open-pollinated maize varieties, and
these are described as “traditional” maize production systems. Both technical and
traditional maize production is included in the economic analysis to ensure that the total
maize market is taken into account.

Table 1. Production and consumption regions defined for the maize simulation
Regions
Villavicencio

Vista Hermosa
Others – Technical
Fuente de Oro
Cumaral
San Juan de Arama
Granada
El Castillo
Puerto Concordia
Puerto Rico

La Macarena
Uribe
Puerto Rico
Others – Traditional
Mapiripan
Mesetas
Lejanías
External production source
Meta – Human consumption
Meta – Animal consumption

The R&D evaluation package, DREAM4 , was used to simulate the likely magnitude and
distribution of the economic benefits of technical change, after defining the scenarios for
estimating the potential economic impact of new production technologies in the Llanos of
Meta, department.
After defining the scenarios for estimating the potential economic impact of new
production technologies in the Llanos of Meta, the R&D evaluation package, DREAM,
4

Dynamic Research Evaluation for Management (DREAM), is a tool to support the economic impact
evaluation of agricultural research .

119

Impact Assessment - Annual Report 2001

was used to simulate the likely magnitude and distribution of the economic benefits of
technical change.
Conceptual Economic Frame: Representing Technical Change
If we first consider that a new technology is fully adopted, its likely impact will depend
on two main considerations:
1. The initial market cond itions for the specific agricultural commodity that the
technology was designed for.
2. The subsequent change in supply and demand conditions for the given commodity
over time. Specifically the difference between the market evolution with and without
the new technology.
In reality the decision to adopt (or not adopt) the new technology is taken by different
producers at different times so the likely rates and levels of adoption in different regions
are important elements in estimating the stream of economic benefits arising from
technical change. The full technology investment and use cycle incorporated in the
economic model includes:
1. A time lag for the development and testing of new technologies.
2. A measure of the uncertainty associated with the R&D process (probability of R&D
success)
3. The potential impact of technology, K (arising from, say, yield increasing or input
reducing technologies) that is expressed as a reduction in the unit cost of production.
4. A time lag following the release of a new technology until the ceiling level of
adoption is likely to be reached.
5. A ceiling level of adoption (in area or quantity terms)
6. A period of dis-adoption if, say, the effects of a single “wave” of new technology are
being modelled, and we know the new technology will itself be replaced within a few
years.
Basic Economic Model
In its simplest form, the DREAM approach can be represented for a single commodity
market as depicted in Figure 1: S0 represents the supply function before a researchinduced technical change, and D0 represents the demand function for a homogeneous
product. The initial price and quantity are P0 and Q0 ; after the supply shift they are P1 and
Q1. Suppose research generates yield increasing or input saving technologies. These
effects can be expressed as a per unit reduction in production costs, K, that are modelled
as a parallel shift down in the supply function to S1 . This research- induced supply shift
leads to an increase in production and consumption to Q1 (∆Q = Q1 - Q0 ), and the market
price falls to P1 (by ∆P = P0 - P1 ). Consumers are better off because R&D enable them to
consume more of the commodity at a lower price.

120

Impact Assessment - Annual Report 2001

The total (annual) benefit from the research- induced supply shift is equal to the area
beneath the demand curve and between the two supply curves (∆TS = area Io abI1 ).
This area can be viewed as the sum of two parts: (a) the cost saving on the original
quantity (the area between the two supply curves to the left of (Qo – area Io acI1 ) and (b)
the economic surplus due to the increment to production and consumption (the triangular
area abc total value of the increment to consumption – area Qo abQ1 – less the total cost
of the increment to production – area (Qo cbQ1 ) --. Alternatively, is possible to portion the
total benefits into the benefit to consumers in the form of the change in consumer surplus
(∆CS= Po abP1 ) and benefits to producer surplus (∆PS = area P1 bI1 minus area Po aIo ).
Although they receive a lower price per unit, producers who adopt the new technology
are better off, because their unit costs have fallen by an amount, K per unit, that is more
than the fall in price. Total benefits are obtained as the sum of producer and consumer
benefits. 5 Thus the size of the market, as indexed by the initial quantity Q0 , as well as the
size of the research- induced savings in the per unit cost of production, K, are critical
factors in estimating the economic benefits from R&D. Better estimates of K mean better
estimates of the benefits from research, and a better basis on which to allocate scarce
research resources.

5

The consumer surplus measure of the consumer benefit is equal to area P0 abP1 , i.e., rectangle P0 aeP1 (=
Q0 ×∆P) plus triangle abe. The producer surplus measure of the producer gain is equal to area P1 bcd in
figure 1, i.e., rectangle P1 ecd (= Q0 × [R - ∆P]) plus triangle bce.

121

Impact Assessment - Annual Report 2001

Figure 1: Surplus distribution on the basic model of research benefits
Price

S0
Shift

(K)

a

P0
P1

e

d
I0

c

I1

0

S1
b

D

Gross Benefit
Per year

Q0

Q1

Quantity

Source: Alston, Norton and Pardey(1995)

Given the site-specific nature of much agricultural R&D, knowledge of the agroecological factors that shape the various biophysical responses to a new technology -- be
it a new seed variety or a new crop management practice -- can substantially improve the
estimate of K used to calculate the benefits from research. A weighted sum across the
zone K’s, using, say, production shares as weights in the aggregation, is likely to yield a
more informative and accurate estimate of the overall K compared with an approach that
leaves this aggregation process implicit or undefined. Used in this way, multimarket
models, in a sense of more markets for one product, have the capacity to capture both
spatial variation in the effects of R&D that primarily have an agro-ecological dependence
and the spatial variation in market factors tha t span agro-ecological domains.
Agro-ecological and biophysical data are useful for evaluating the potential results of
R&D in changing productivity and natural resource conservation, but realising this
potential also depends on other factors such as rural infrastructure investment,

122

Impact Assessment - Annual Report 2001

communication systems, roads, education and health services, as well as the national
structure of the agricultural sector and distortionary government policies.
This framework also has the capacity to represent multiple regio ns (markets) and
interacting with each other through trade. Figure 2 shows the case of two regions. Under
the assumption of free trade it is not necessary to have internal equilibrium conditions in
each region.
Figure 2 Import Innovator Region and its National Impact and through Interregional Trade
(a) Region A. Production,
Consumption & Trade
(Importer region)

Region A Quantity

(b) Excess supply demand
& Trade

Trade Quantity

(c) Region B (or ROW)
Production, consumption
& Trade

Quantity
ROW

If the innovator region, Region A, is an importer, consumers world-wide benefit from the
research- induced, price decrease (domestic consumers gain Po ijP1 in graph (a), and ROW
producers loss (area Po fhP1 in graph (c). The innovator region gains. In contrast with the
export innovator region the ROW looses because the loss to ROW producers (area Po fhP1
in graph (c) exceeds the benefit to ROW consumers (area Po klP1 ). The net ROW loss is
shown as the area Po egP1 in graph (b) (which equals area Po fhP1 minus area Po klp1 =area
kfhl in graph (c) in Fig. 2.
The impact of research in a small-region importer of a commodity as in the case of Meta
department in Colombia is illustrated in Figure 3.

123

Impact Assessment - Annual Report 2001

Figure 3 Small Open Economy: Importer Case.

The initial equilibrium is defined by consumption, Co (which is iqual to Qo + QT 0 ) and
production, Qo at the world price Po , with a trade quantity, QTo (representing imports),
equal to the magnitude of the difference between consumption and production. Research
causes supply to shift from So to S1 and production to increase to Q1 . As a result, imports
decrease to QT1 . (and Q1 + QT1 = C1; and C1 =Co ). Because the region does not affect the
world price, the economic surplus change (iqual to area Io abI1 ) is all the producer surplus
while consumer surplus does not change (area abc).
One advantage of the small- region assumption is that, even when the government could
intervene in the commodity market, all the research benefits continue to accrue
domestically and there is no need to consider the ROW (unless there is leakage of
research results and consequent price spillovers feeding back from the ROW). The price
Po , is a constant in the analysis and defines the opportunity cost of resources used in the
production and consumption.
The global equilibrium price in the market, Po , is determined where the export production
surplus in the region A corresponds to the deficit of region B, (Qt o ).
If region A adopts new technology, the supply curves shifts and there will be more
production available for internal consumption and trade. The new supply curve is shown
as SA1 and, if the demand conditions in regions A and B do not change, a new
equilibrium will be reached with a lower price, P1 with a greater quantity of trade, QT1.
This lower price will be reflected in the internal prices of regions A and B, with
consequences to production and consumption in the national markets.

124

Impact Assessment - Annual Report 2001

Figure 4 Export Innovator Region and its National Impact and through Intern-regional Trade

(a) Region A production,
Consumption & trade
(exporter region)

Region A Quantity

(b) Excess supply, demand
& trade

(c) Region B (or ROW)
production, consumption &
Trade (importer region)

Traded Quantity

ROW Quantity

In general, the impacts of the new technology in these cases are:
•

For the innovator region: Lower price, more production, and more consumption.

•

For trade : Lower price, more commerce (higher levels of exports and imports).

•

For non innovator regions : (or lesser innovators): Lower price, less production and
more consumption.

Technology adoption

125

Impact Assessment - Annual Report 2001

Figure 5. Time lags and the pattern of adoption.
% Adoption
of New
Technology
100
Ceiling Level of Adoption

AMAX

Adoption

0

5

10

Disadoption

15

20

25

30

Year
Research

Development
Lag

Adoption Process

Source: Adapted from Alston, Norton and Pardey. 1995.

The presentation of the conceptual economic framework identified two situations with
respect to adoption and use of technologies: with and without the new technology (the
points and lines annotated as 0 and 1 in the figures 1, 2 and 4). It is important to know the
time lags associated with the research and adoption process. Figures 5 shows a typical
adoption curve of new technologies over time, which requires three parameters to define
this simplified representation:
A time lag following the release of a new technology until the ceiling level of adoption is
likely to be reached, a ceiling level of adoption (in area or quantity terms), and the shape
(or functional form) of the adoption curve.
The methodology requires the specification of these three parameters to approximate an
adoption curve for each region. The shape of this curve is sigmoidal up to the ceiling
level of adoption. This is supported in the literature. However, there is controversy with
respect to the disadoption concept. From one point of view, each technology has a cycle
of life until the user abandons it totally and, as from that moment, user benefits stop.
From another point of view, the producer will change technology only if the new one
gives a better profitability relative to the existing one. Therefore, the benefits (or profits)
of the old technology are the benchmark for the new technology and the ‘old technology’
part of the total profit always continues.
These concepts are implemented in the economic framework by making a simulation
from an initial period (the base period), through a specified time (the simulation period).
The simulation goes through four stages:

126

Impact Assessment - Annual Report 2001

R&D period: It is the investment period in R&D until the new technology is released.
During that period there are no benefits.
Adoption Period: (From the releasing till achieving the adoption ceiling). In this period
the benefits grow each year, in each region, in proportion with the adoption level. It is
possible that costs are still being incurred in adoption and diffusion of the technology,
Maximum adoption level of the technology: In this period adoption is maintained at the
adoption “ceiling” level. During these years the annuals benefits are maximum (and
equal) each year. The faster is the technological renovation cycle associated with a
research topic, the shorter is this period, and
Disadoption Period (optional): This period can arise from new and better technology
becoming available or, sometimes, the loss in effectiveness of the technology.
Disadoption does not exclude the possibility of receiving economic benefits (if, for
example, the old technology is somehow included in the new technology).
As a result of adoption of a new technology in a region, it is possible to estimate the shift
in the supply curve as:
Ki, t = K i. Ai, t
Where Ki,t is the unit cost reduction of production in region i for year t: Ki is the potential
unit cost reduction in region i if the technology is fully adopted; and Ai,t is the adoption
level in the region i for year t.
Estimation of the social benefits of research
The simulation over time estimates change s in prices and in quantities produced and
consumed as a consequence of the adoption of new technology. According to the
concepts of economic surplus it is possible to transform these changes into
approximations of changes in social welfare (Alston, Norton and Pardey, 1995). The
annual benefits (changes in the economic surplus) for the t year, can be expressed as
changes in consumer surplus, ∆PSt, changes in producer surplus ∆PSt, and total change in
surplus ∆TSt, and can be calculated as followed:
∆CSt = Po Qo Zt (1 + 0.5 Zt n)
∆PSt = Po Qo (K t - Zt) (1 + 0.5 Zt n)
∆TSt = ∆CSt + ∆PSt = Po Qo Kt (1 + 0.5 Zt n)

Po and Qo are the initial price and quantity, Kt, is the proportionate vertical shift down in
the supply curve. In year t we can define Zt = Kt e / (e + n), where n is the demand
elasticity, and e is the supply elasticity. Z represents the reduction in price, relative to its
initial (i.e. pre-research) value, due to supply shift. Or, put another way, it is the

127

Impact Assessment - Annual Report 2001

proportion of the market price reduction –(P0 – P1 ) relative to the absolute value of the
unit cost reduction Kt . P0.
Algebra for the research benefit calculation in a closed economy.
The basic model
The relative reduction in price is defined as Zt = Kt e / (e+n) = -(P1 – Po )/P o where Po and
Qo are equilibrium price and quantity before supply shift, e is the supply elasticity, and n
the absolute value of the price elasticity of demand. The equation for Z is obtained by
solving linear supply-and- demand equations for prices as a function of slope and
intercept parameters, treating a research-induced supply shift as an intercept change, and
converting to elasticities:
Supply:

Qs = α + β(P +κ) = (α + βκ) + βP

Demand:

Qd = γ - δP

Where κ is the absolute shift down of supply due to a cost saving induced by research. In
figure 1, κ =( Po – d), and the supply shift relative to initial equilibrium price is K = κ/Po
= (P o – d)/P o.
Equilibrium price change
Setting Qs = QD = Q yields the equilibrium price P = (γ - α - βκ)/(β + α). When κ = 0, Po
= (γ - α)/(β + δ); When κ = ΚPo , P1 = (γ - α - βΚPo )/(β + δ). The research induced
change in price is (P1 – Po ) = - βΚ Po /(β + δ) and the relative change in price is given by
-(P1 – Po )/P o = βΚ/(β + δ). Converting the slopes to elasticities (multiplying through the
numerator and denominator by Po /Qo ) yields:
Zt = Kt e / (e+n) = -(P1 – Po )/P o .
Consumer surplus
In figure 1, the consumer surplus change is given by ∆CS = Po abP1 = rectangle Po aeP1 +
triangle abe = (P0 – P1 )Qo + 0.5(Po – P1 )(Q1 - Qo ) or ∆CS = (Po – P1 )Qo[1+0.5(Q1 –
Qo )/Qo ]. Using the definition above that Z = -(P1 – Po )/P o so that (Q1 – Qo )/Qo = Zη
yields:
∆CS = Po Qo Z(1+0.5 Zη).
Producer surplus
The producer surplus change is ∆PS = P1 bI1 – Po aIo = P1 bcd + dcI1 – Po aIo = P1 bcd
given that dcI1 = Po aIo under the assumptions of a parallel supply shift and linear supply
and demand. ∆PS = P1 bcd = rectangle P1 ecd + triangle bce = (P 1 – d)Qo + 0.5(P 1 – d)(Q1

128

Impact Assessment - Annual Report 2001

– Qo ). Thus, ∆PS = (P 1 – d)Qo [1 + 0.5(Q1 – Qo )/Qo ]. Then it is possible to define (P1 – d)
= (P o – d) – (Po – P1 ) = ΚPo – ZPo and (Q1 – Qo ) = Zη. Thus,
∆PS = (K – Z)P o Qo (1 + 0.5Zη).
Total surplus
Note also that ∆TS = ∆PS + ∆CS = Po abcP1 = ( Po acd + abc), which in this instance
equals Io abI1 =( Io acI1 + abc) given that Po acd = Io acI1.
Using the above equations, a stream of annual benefits can be calculated over the whole
simulation period (see figure 2.6). In accordance with the preceding equations it is also
possible to distribute the total benefits between the producers and consumers. This kind
of information is extremely useful for strategic analysis and the design of research
programs or projects, especially since the dynamic nature of benefit and cost streams are
clearly visible.
Annual benefit and cost streams can be converted into present values, that is deflated
benefits and costs, by choosing a base year and a representative real (commercial minus
inflation) discount rate. For instance, the present values of gross benefits for the
consumer VCS, producer VPS, and in total, VTS, are calculated as follows:
VCS = Σ t ∆ CSt / (1 + r)t
VPS = Σ t ∆ PSt / (1 + r)t
VTS = VPS + VCS
Where the r is the real discount and t is the year (between 1 and T, the total period of the
economic simulation). With information about the R&D costs in each year, Ct, is possible
to calculate the net presents values, NPV, that could give better estimates of the true
economic attractiveness of the R&D, where:
NPV = Σt (∆TSt -Ct) / (1 + r)t
Another important measure is the internal rate of return, IRR, the discount rate at which
net present value of benefits is equal to the net present value of costs (e.g., the difference
between the net present values is zero:
0 = Σ t (∆TSt - Ct) / (1 + IRR)t
With the DREAM model it is possible to calculate the benefit streams (Figure 6), the cost
and benefit present values, NPV’s and IRR at different levels of aggregation – for
example at municipal and at department level) to obtain a clearer picture of what are the
incentives to undertake R&D targeted to each level. It is also possible to see the extent to
which different producer groups will likely win or lose from the introduction of the new
technology.

129

Impact Assessment - Annual Report 2001

Figure 6. Cost and benefit streams of R&D over time.

Source: Adapted from Alston, Norton and Parday. 1995

Research and Development
Since it was not possible to obtain all the relevant cost elements of the research
undertaken in generating the new maize and soybean rotation system, the emphasis was
focused solely on the estimation of the gross benefits of research, and how those benefits
might be distributed; by producers versus consumers, by municipio, and over time. 6
6

The cost information proved to difficult to collect for several reasons; there were several research
institutions involved; CIAT, CIMMYT, CORPOICA, and ICA, and several projects within each of those
institutions that contributed to the research. Furthermore, the accounting systems and record keeping made
it extremely difficult to assess how many scientists had been working, and with what level of effort, on
generating the specific technologies we are evaluating here. Thus, our benefit estimation will be in terms of
GROSS benefits only.

130

Impact Assessment - Annual Report 2001

The economic evaluation is of a technology already developed, although the specific
maize-soybean rotation proposed had not yet been adopted. Thus, there is already data,
used in this study from experimental and on- farm trials about the effectiveness of the
rotation. The economic simulations were made to estimate the likely economic returns of
that technology within the Meta department according to given adoption and market
scenarios (See table 2).
Market data
The most important variables taken into account from a market perspective are:
Production, consumption and prices
The information (data) was gathered of the level of municipality for the traditional and
improved maize production systems (See table 2).
ü Production: Data on area, production and yield of technical and traditional maize
was obtained for the period 1995 to 1998, from the Agricultural, Livestock, and Rural
Development Secretary of the Meta Department. Other aggregated production data
were obtained for the last 10 years from the Ministry of Agriculture and Rural
Development.
ü Producer prices: Data on producer prices by crop semester for the periods 1997 (2nd
semester) and 1998 (1srt semester), also from Agricultural, Livestock, and Rural
Development Secretary from Meta Department. All these prices were converted into
constant prices of 1997, using the cereal consumer price index (CPI) of the capital
city of Meta, Villavicencio.
ü Consumption: To estimate consumption it was necessary to obtain data about
population and per-capita consumption of maize in Meta for the year 1997 for each
municipality. Animal consumption of maize was also estimated based on annua l
consumption per head, on the census of slaughtered cattle for the province.
Consumer prices: Detail information was obtained on wholesaler prices at Villevicencio
and Bogotá the major markets for agricultural trade to and from Meta Department. These
prices were taken in constant values with 1997 as base year using the cereal consumer
index prices for each town.
Elasticities
Price elasticity supply and demand were obtained for the commodity selected, these
values were calculated by IFPRI based on a dema nd and supply survey of maize in
Colombia. The elasticities values have little impact on the R&D total benefits; however,
these elasticities affect significantly the likely distribution of the benefits among
producers and consumers.

131

Impact Assessment - Annual Report 2001

Item

Demand

Supply

Maize

-0.4

0.5

Source: CIAT “Trends in CIAT commodities”. 1992.

Demand and supply exogenous (non R&D) rates of growth
Estimations of demand and supply growth rates for Meta Department were based on the
population growth projection of the Departamento Administrativo National de
Estadísticas of Colombia (DANE), United Nations Program for Development (UNPD)
and Agricultural Department of the United States, assuming there is no consumption percapita changes with respect to the base period (1995 – 1997). The animal consumption
estimation was projected taking into account information about livestock slaughtered
from DANE data. The supply growth rates (for growth not attributed to R&D) were
estimated on base of the harvest area and yield trends during the period 1998 – 1997.
Discount Rate
Taking into account the value of money across time it is necessary to discount from
future values using a real interest rate to calculate the R&D cost and benefit present
values. The real discount rate was estimated subtracting the inflation rate from the market
minimum interest rate.
Data Calibration
The data collected from different sources were integrated and harmonised checking up
the consistency and reliability of the different sources of information.
The adoption ceiling for the maize technology was estimated from average production for
each region weighted, using the information on production type from the table 3. For the
regions defined as traditional producers, the adoption ceiling is zero, because the new
technology evaluated in this thesis have been focused on improved maize. We have
arbitrarily assumed that 50% of the technological maize producers will adopt the new
technology, and have made some sensibility tests regarding changes in prices, quantities
and benefits if this assumption is changed.

Basic Research Factors
The table and 3 show some of the key technical factors of the new technology. Before
undertaking the analysis it is important to consider the following issues:
•

Deciding which market model would be most appropriate for the commodity and
technology combination being assessed.

132

Impact Assessment - Annual Report 2001

•

Deciding on the appropriate levels of aggregation of production and consumption.

•

Deciding how to represent the imbalance between production and consumption of
maize in the Meta department, that is, how to represent the external trade to maintain
the equilibrium among total supply and demand.

Markets and Trade
To support the definition of the technology scenarios to be evaluated it is useful to have a
global understanding of the maize market structure.
The maize market is modeled a multiple trading regions. Production is separated into
markets for each municipality. The total production in simply the aggregate production
from all municipalities. The model requires that total production and total consumption
across all regions are equal. It is not necessary the production and consumption are in
equilibrium in each individual region. The model estimates the annual market-clearing
prices considering the conditions of all markets simultaneously (that is, price is
endogenous). In setting up the production regions (the municipios), municipalities with a
small participation were grouped together as “others producer”. (See Table 3).
In this case of consumption just two regions were defined - one for human consumption
and one for animal consumption. Since the combined consumption is greater than the
combined production, a separate external (production) region was also defined that serves
as a source of the maize imports to the region.

RESULTS
The Maize Economy of Colombia and the Meta Department
Production trends
Maize production systems in Colombia, as in the rest of Latin America, are very variable,
reflecting not only the wide rage of production environments in which maize is grown but
also differences in farmer knowledge and access to resources. Maize is an important food
staple that is produced and consumed by a large part of rural population. Maize is often
grown in association with beans, cassava, and other food crops destined for home
consumption, and many farmers use. It is also of increasing importance for animal feed,
particularly for the fast growing pig and poultry sector. Colombia is the fourth- largest
maize-producing country in Latin America after Brazil, Mexico, and Argentina. Annual
production of maize in Colombia currently is around one million tons.
Maize in Colombia is grown in a wide range of agro-climatically diverse environments
by producers who differ in their resource endowment, technical knowledge, and crop
management practices.

133

Impact Assessment - Annual Report 2001

Figure 7: Total area, and production of Maize,
Colombia. 1976-1996
Area (000 ha)

Production (000 MT)

19
96

19
94

19
92

19
90

19
88

19
86

19
84

19
82

19
80

19
78

19
76

1400
1200
1000
800
600
400
200
0

At one extreme, small scale semi-subsistence producers grow maize primarily as a food
crop, some producers plant local varieties with low yield as a food crop; most of these
producers plant local varieties with low yield potential, high yield stability, and moderate
resistance to local pest and diseases. At the other hand extreme, relative large-scale
commercial producers grow maize primarily as a cash crop; most of this producers plant
improved maize material, especially hybrids. Between this two extremes lie many
intermediate type of producers.
The trend in area sown to maize for the period from 1990 to 1998 demonstrates a loss in
cultivated hectares equivalent to 306,037, which is equivalent to the 38,90% of maize
area with the greatest reduction being in the year 1998 (See Figure 7).
Maize production in Colombia in the last eight years has decreased by 261,371 tons
metric - in 1990 maize stood at 1,155,000 metric tons metric, but in 1998 production was
893,629 metric tons; this reduction is equivalent to 22,6%.
In average in the last thirty years, the yield of technical maize in Colombia has increased
from 2.43 ton/ha to 3 ton/ha, that is, an increase of 0.57 ton/ha or 23,46% in roughly
thirty years (Figure 8). The lowest value was in 1976 with a yield of 1,89 ton/ha, and the
highest in 1997 with 3,01 ton/ha (National Federation of Cereal Producers of Colombia –
FENALCE, Hoja del Cerealista. 1999).

134

Impact Assessment - Annual Report 2001

Figure 8: Area and production of technical Maize in
Colombia. 1987-1997
Area (000 ha)

Production (000 MT)

350
300
250
200
150
100
50
0
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997

Traditional maize production in Colombia, in spite of the reduction in the amount of
sowing hectares, falling from 560,300 hectares in 1970 to 489,520 hectares in 1997,
increased due to yield improvements (Figure 9 and 10). In 1970 production stood at
631,300 metric tons but by 1997 had reached 723,652 metric tons (FENALCE). Yields
increased from 1,13 ton/ha to 1.41
Figure 9: Area and production of traditional Maize in
Colombia. 1987-1997
1200
1000
800

Area (000 ha)

600
400

Production (000 MT)

19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97

200
0

ton/ha over the same period, an increase of 0.35 ton/ha, equivalent to 30,97%
(FENALCE: La Hoja del Cerealista, 1999).

135

Impact Assessment - Annual Report 2001

Figure 10: Yield of technical and traditional Maize in
Colombia. 1987-1997.
Technical (t/ha)

Traditional (t/ha)

4.00
3.00
2.00
1.00
0.00
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997

Consumption trends
Maize consumption patterns in Colombia have changed considerably with the migration
of a significant part of the population from rural areas to cities.
Traditionally, maize was been a traditional food crop grown mainly by rural households
for direct consumption or for feeding animals. As direct human consumption of maize
has increased over the year, feed use is currently less than maize for human consumption.
The per capita production (Figure 11) has decreased during the period 1991-1998 by 34%
which is equivalent to decreasing in production of 12,09 Kg/per-capita, while the
consumption increased tremendously during the same period in 106%, equivalent to
increase in consumption 37,95 Kg/per-capita.

136

Impact Assessment - Annual Report 2001

Figure 11: Percapita production and consumption of
Maize in Colombia. 1989-1998
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00

Production Kg/dweller

19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98

Consumption
Kg/dweller

Meta Department
Meta Department and Arauca Intendence have the highest yields and production of
maize. This region has 37,867 hectares that represents 11,08% of the Colombian total and
corresponds to 9,76% of the traditional system and 1,32% of the improved (technical)
system (Figure 12). The average yield for this region in the traditional sector is 1,140
kg/Ha and, 2,500 kg/Ha in the improved maize cultivation. For both traditional and
technical systems there are two periods of sowing; from mid February until end of March
and during the month of September to harvest in July-August and December-January
respectively. The technology used in the different cultivation systems, for the maize
production, corresponds to monoculture and multiple cropping with cassava and plantain.
The climate and soil conditions provide unfavourable conditions for the development of
maize cultivation. Technical assistance is another highly constraining factor in this region
and is linked to limited credit access.
Figure 12 : Total area, production and yield of
technical and traditional Maize in Meta, Colombia.
1984-1997
80
60

Area (000 ha)

40

Production (000 MT)

20

Yield (tn/ha)

19
96

19
94

19
92

19
90

19
88

19
86

19
84

0

137

Impact Assessment - Annual Report 2001

Comparing area planted in the years 1997-1998 for Meta Department, there has been a
reduction of 4,110 hectares which, in percentage terms, is equivalent to - 43% with
respect to the previous year.
The maize sowing trend for this Department for the period 1990-1998 shows loss in
cultivated areas equivalent to 10,240 hectares in technical maize, which is equivalent to –
65%, and while for the traditional commodity there was a slight increment of 1,215
hectares, it means 15% more area were cultivated in 1998.
The maize production in Meta in the last six years has fallen by 26,570 metric tons of
technical ma ize, since in 1993 the production was 42670 metric tons and in 1998 the
production was 16100 metric tons – a reduction of some 62%.
Description of the Results
The Tables 2 and 3 show in detail the final data integration. Each table show the values
used as inputs for the simulation run. The first table presents the summary of the market
data, and the second table summarises the research and adoption data. The number of
rows depends upon the level of disaggregation of the different production and
consumption regions (markets). If there is no sub-national disaggregation, a single region
can be defined that contains both production and consumption characteristics.
Table 2 shows the following row pattern:
Two groups of municipios (with and without significant adoption of technical maize
production practices). Only production information is defined for each muncipio,
consumption data is presented at the department level (for both human and animal
consumption). Each municipio has data on quantity of maize produced, farmgate price,
price elasticity, and supply growth rates (not caused by research).
Transmission factors are not used in this analysis (they are assumed to be 1.0). But the
model does remember the initial structural differences in price levels between regions.
These are called the price “wedges”. The municipio of Mesetas was arbitrarily chosen as
the base region from which to measure price wedges. When new technology is adopted,
prices change and the transmission elasticity determines how much of the price change in
any region is transmitted to the other regions. This is a crude way of dealing with
transport and other transactions costs between regions. Here, however, a perfect free trade
with no transactions costs was assumed (transmission elasticity = 1.0). If the transmission
elasticity is set to 0.0 the simulation assumes each region acts as a closed economy.
There are two rows of data to define two demand regions, one for human food demand,
and the other for feed demand. These regions do not have data about production,
technology and adoption.
One row represents production outside of the Meta department, which is not a target of
research, and which supplies the maize needed in the Meta, since the total regional

138

Impact Assessment - Annual Report 2001

production is less than the total demand. This external region has information about
production but not about technology and adoption. This external region supplies the
production deficit, and this importation represents 59% (29,467.4 metric tons of maize
per year) of Meta total consumption.
In Table 3, column 6 is the calculation of the net shift, Kmax , of the supply curve as a
percentage of the initial price, and as a absolute value. It is the product of the potential K,
the probability of success (here 100% because the technology has already been
developed), the max level of adoption, and the producer price.
In the case of municipios with technical adoption the average cost reduction per unit is
$13.34/tonne, and the potential impact varies from $8.77/tonne (Mesetas) up to $17/tonne
(Other – technical).
It is assumed that the R&D time is zero years, since the technology is already available
(but has not been adopted). So an adoption time of 6 years is assumed, and it is assumed
that the new technology will ultimately be adopted by all technical producers in each
muncipio.

139

Impact Assessment - Annual Report 2001

Table 2: Summary of Input Data – Market Data
Study:

Technical Change in Colombian Savannas

Scenario:

Maiz/Soybean Rotation – Meta Departments

Commodity: Maize

Regions:

20

Horizontal Multimarket - Spillover: No

Period:

20 years

Base year:

Discount:

10.0%

Benefit units:1000COL$

1997
Quantity units: Tonnes

SUMMARY OF INITIAL MARKET CONDITIONS
Elasticity Transmission Exog. Growth
Tax/Sudsidy
Region
Production Consumption Price
Sup.
Dem. Wedge Elast
Sup. Dem.
Sup. Dem.
------------------------------------------------------------------------------------------------------------(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
Tonnes
Tonnes
1000C$
1000C$
%/yr
%/yr
%/yr
%/yr
------------------------------------------------------------------------------------------------------------Municipios with traditional maize systems
Mapiripan
Puerto Concordia
Puerto Rico
La Macarena
Uribe
Other - traditional

1157.2
524.3
261.5
305.0
2000.0
1114.2

0.0
0.0
0.0
0.0
0.0
0.0

250.00
240.00
248.00
245.00
240.00
243.96

0.500
0.500
0.500
0.500
0.500
0.500

0.000
0.000
0.000
0.000
0.000
0.000

9.30
-0.70
7.30
4.30
-0.70
3.26

1.00
1.00
1.00
1.00
1.00
1.00

0.00
0.00
0.00
0.00
0.00
0.00

0.00
0.00
0.00
0.00
0.00
0.00

0.00
0.00
0.00
0.00
0.00
0.00

0.00
0.00
0.00
0.00
0.00
0.00

Municipios with significant technical maize systems
Mesetas
Villavicencio
Vistahermosa
Fuente de Oro
Cumaral
San Juan de Arama
Granada
Lejanias
El Castillo
Other – technical

2241.6
1588.7
730.0
2485.0
378.0
704.7
4376.2
1188.0
702.1
573.5

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

240.70
245.00
240.00
245.00
240.70
240.00
214.00
245.00
240.00
240.70

0.500
0.500
0.500
0.500
0.500
0.500
0.500
0.500
0.500
0.500

0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.00
4.30
-0.70
4.30
0.00
-0.70
-26.70
4.30
-0.70
0.00

1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

0.0
0.0

7725.8
42071.9

240.00
245.00

0.000
0.000

-0.400
-0.400

-0.70
4.30

1.00
1.00

0.00
0.00

1.03
2.01

0.00
0.00

0.00
0.00

0.500

0.000

0.00

1.00

2.80

0.00

0.00

0.00

Internal Meta demand
Meta Demand – Animal
Meta Demand – Human

External supply region providing net imports to Meta
Imports – External

29467.4

0.0

240.70

------------------------------------------------------------------------------------------------------------TOTAL
49797
49797

Note:
1.
2.

Significant growth in both human and animal consumption is projected based on human and livestock population trends.
Transmission price wedge is calculated by the program. It is the difference between the price in the “base” municipio (arbitrarily chosen as Mesetas) and all
other regions.

140

Impact Assessment - Annual Report 2001

Tabla 3: Summary of Input Data – Research and Adoption Data

SUMMARY OF R&D & ADOPTION DATA
K
Prob.of Max.
K Shift
K
----- Time Lags ------ Adopt
Region
Pot. Success Adopt
Price
Max Type Var
R&D Adopt AtMax Aband Form
--------------------------------------------------------------------------------------------------(1)
(2)
(3)
(4)
(5)
(6)
(7) (8)
(9) (10) (11) (12) (13)
-%-%-%- 1000C$ 1000C$
%/yr
-yrs- -yrs- -yrs- -yrsMunicipios with traditional maize systems
Mapiripan
Puerto Concordia
Puerto Rico
La Macarena
Uribe
Other - traditional

0.00
0.00
0.00
0.00
0.00
0.00

0.0
0.0
0.0
0.0
0.0
0.0

0.0
0.0
0.0
0.0
0.0
0.0

250.00
240.00
248.00
245.00
240.00
243.96

0.00
0.00
0.00
0.00
0.00
0.00

S
S
S
S
S
S

0.00
0.00
0.00
0.00
0.00
0.00

0.0
0.0
0.0
0.0
0.0
0.0

0.0
0.0
0.0
0.0
0.0
0.0

99.0
99.0
99.0
99.0
99.0
99.0

0.0
0.0
0.0
0.0
0.0
0.0

X
X
X
X
X
X

Municipios with significant technical maize systems
Mesetas
Villavicencio
Vistahermosa
Fuente de Oro
Cumaral
San Juan de Arama
Granada
Lejanias
El Castillo
Other - Technical

9.00 100.0
9.00 100.0
9.00 100.0
9.00 100.0
9.00 100.0
9.00 100.0
9.00 100.0
9.00 100.0
9.00 100.0
9.00 100.0

40.5
64.7
47.5
77.0
71.7
43.1
80.0
51.5
65.7
80.0

240.70
245.00
240.00
245.00
240.70
240.00
214.00
245.00
240.00
240.70

8.77
14.26
10.26
16.97
15.53
9.30
15.40
11.35
14.19
17.33

S
S
S
S
S
S
S
S
S
S

0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0

99.0
99.0
99.0
99.0
99.0
99.0
99.0
99.0
99.0
99.0

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

X
X
X
X
X
X
X
X
X
X

0.0
0.0

240.00
245.00

0.00
0.00

S
S

0.00
0.00

0.0
0.0

0.0
0.0

99.0
99.0

0.0
0.0

L
L

Internal Meta demand
Meta Demand - Animal 0.00
Meta Demand - Human 0.00

0.0
0.0

External supply region providing net imports to Meta
Imports – External
0.00
0.0
0.0 240.70
0.00
S
0.00
0.0
0.0 99.0
0.0
L
--------------------------------------------------------------------------------------------------Kmax is maximum absolute unit cost reduction (in price units). Product of Cols 2-5
Kvar is the variable unit cost reduction (%/year), if specified.
Shift type: S - Supply, D - Demand
Adoption Form: L - Linear, X - Logistic

Notes.
1.
2.
3.
4.
5.

Since the technology has already been developed, the probability of research success is 100 percent
Adoption levels are proportional to the share of technical production in the municipio. The maximum adoption level was taken as 80%
Kmax is the maximum unit cost reduction in Col$1000 per tonne that will be achieved in each municipio.
No disadoption was assumed. Thus, benefits were received for every year of the simulation period following adoption.
No technical change is assumed to be taking place in the external supply region. This assumption can be easily changed by providing R&D and adoption
data for that region.

141

Impact Assessment - Annual Report 2001

Dynamics of Market
The economic impact of technical change can be visualised in each market, by the
changes brought about prices and quantities. Specifically, the time paths of prices and
production levels can be compared with and without the introduction of new
technologies. This step proceeds the analysis of the estimation of benefits. Thus changes
in; prices, quantities produced and consumed, trade, and the values of production and
consumption, are calculated in each year as a basis of calculating the benefits of the
R&D.
Initially, there is no difference between production and price "with" and "without"
technology. However, once the new technology is released and adopted, the “with” and
“without” values begin to vary, to an extent that is determined by the cost reducing
potential of the new technology and the rate and level of adoption in each region.

Figure 13: Production of maize in Meta
Quantity with R&D

Quantity without R&D

21000
Tonnes

20800
20600
20400
20200
20000
19800

The figures 13 and 14 show the tendency over time for the production of maize with and
without research in the Department of Meta. The area between the cur ves is a measure of
the potential benefits of R&D. The simulation predicts an expansion in the production of
technical maize for this Department (Figure 14).

142

Impact Assessment - Annual Report 2001

Metric tonnes

Figure 14: Increasing in maize production, Meta
Department
400
300
200
100

20
16

20
14

20
12

20
10

20
08

20
06

20
04

20
02

20
00

19
98

0

Simulation Period

It is important to notice that the prices and amounts in each region are related, since
adoption of new technology in one region impacts the situation of the market in other
regions through the effects of trade.
The simulation predicts a cumulative additional production in the quantity of maize
produced in Meta, attributable to R&D, to be 6,165 me tric tonnes during the simulation
period 1998-2017 providing benefits to the producers about Col$1.236.135,0 and benefits
to the consumer by Col$846.719,5 because of the decrease in prices (See Tables 4, 5 and
Figure 15).
Benefits by Region
As was described in the section on the conceptual economic frame, we can use the
concepts of the economic surplus and time preference (discounting values) to turn the
changes into prices and amounts over time to a single value of the economic benefit (and
costs) of R&D. Thus, for each region we can turn the annual flow of impacts into
measures such as the present value of gross benefits (B) and, if the data is available, of
costs (C), and other related criteria of investment.
The present value of gross benefits of the producer, consumer and totals considered for
each the commodity maize R&D are shown in Figure 15, including details of regions
defined. The inspection of these results suggests, that (according to the assumptions
described in the Section: Basic Research Factors, the market and the trade performance),
consumers always benefit from the R&D. In most of the cases, the producers in regions
with significant levels of adoption also benefit from the impact of new technologies.
For the commodity maize the consuming regions receive high net benefits during the
period of analysis Col$846719,5 equivalent to 64% of total benefits, due to the increase
of the productivity in the defined regions, and to the reduction of the prices. The benefits
for all producers are roughly 58% of the total benefits (Table 4 and 5).

143

Impact Assessment - Annual Report 2001

Figure 6.9. Meta Department Present Values of Gross Benefits by Producer and
Consumer Groups

2.100.000

1000 Colombian Pesos

Total Benefits

1.600.000
Technical

All Producers

1.100.000

Human
Consumption

600.000

100.000

All Consumption

Animal
Consumption
Traditional

-400.000

Beneficiary Group

The benefits to the consumers were constrained because the quantity of maize imported
represents a big proportion of the total consumption, and it was assumed that in the
external region was no technological change during the simulation. With the introduction
of any technical change different to zero in the external region, the benefits to producers
in the Meta would drop simultaneously, while the benefits to consumers would increase.
It is important to understand the meaning of the negative benefits to the producer (that is
losses for the producer). Such results mean that the gains from the production would have
been greater without the new technology (Table 5). This often means that since they
could no longer produce competitively, some producers have reduced or stopped
production of the affected commodity. We can easily see this effect in the production
regions where it is assumed that the new technology will not be available for the adoption
(or will not be adoptable, in this case by the traditional maize producers).
The calculated benefits to non-adopting producers are always negative. This is because
the price diminishes as a result of the adoption of new technologies in other regions, but
the costs of production has not changed for those who do not adopt. This general
reduction of the price reduces the net gains - until unacceptable levels. In such cases,
rational producers will change to their next best option of production (or, in extreme
cases, they stop the production if they are left with no other economic options).

144

Impact Assessment - Annual Report 2001

Table 4: Present Value of Gross Benefits for Each Region and Region Group
Technical Production

1
6
7
13
14
15
16
17
19
8

---------- Present Value of R&D Benefits ----------><---Costs---><------ Returns ------------>
Region
Producer
Consumer
Government
Total
(B-C)
B/C
IRR
-----------------------------------------------------------------------------------------------------------------Mesetas
105756.4
0.0
0.0
105756.4
0.0
105756.4
--.---.Villavicencio
137317.7
0.0
0.0
137317.7
0.0
137317.7
--.---.Vistahermosa
42170.6
0.0
0.0
42170.6
0.0
42170.6
--.---.Fuente de Oro
263320.4
0.0
0.0
263320.4
0.0
263320.4
--.---.Cumaral
36125.1
0.0
0.0
36125.1
0.0
36125.1
--.---.San Juan de Arama
35938.5
0.0
0.0
35938.5
0.0
35938.5
--.---.Granada
415033.5
0.0
0.0
415033.5
0.0
415033.5
--.---.Lejanias
77906.7
0.0
0.0
77906.7
0.0
77906.7
--.---.El Castillo
60320.9
0.0
0.0
60320.9
0.0
60320.9
--.---.Otros Tecnificado
62245.7
0.0
0.0
62245.7
0.0
62245.7
--.---.-

-----------------------------------------------------------------------------------------------------------------Total NPV Benefits
1236136.0
0.0
0.0
1236136.0
0.0
1236136.0
--.---.Traditional Production
---------- Present Value of R&D Benefits ----------><---Costs---><------ Returns ------------>
Region
Producer
Consumer
Government
Total
(B-C)
B/C
IRR
-----------------------------------------------------------------------------------------------------------------2 Puerto Concordia
-7711.6
0.0
0.0
-7711.6
0.0
-7711.6
--.---.3 Puerto Rico
-3845.6
0.0
0.0
-3845.6
0.0
-3845.6
--.---.4 La Macarena
-4485.3
0.0
0.0
-4485.3
0.0
-4485.3
--.---.5 Uribe
-29412.4
0.0
0.0
-29412.4
0.0
-29412.4
--.---.18 Mapiripan
-17018.6
0.0
0.0
-17018.6
0.0
-17018.6
--.---.9 Otros Tradicional
-16385.7
0.0
0.0
-16385.7
0.0
-16385.7
--.---.-----------------------------------------------------------------------------------------------------------------Total NPV Benefits
-78859.6
0.0
0.0
-78859.6
0.0
-78859.6
--.---.Demand
---------- Present Value of R&D Benefits ----------><---Costs---><------ Returns ------------>
Region
Producer
Consumer
Government
Total
(B-C)
B/C
IRR
----------------------------------------------------------------------------------------------------------------10 Meta Consumo Anima
0.0
123641.2
0.0
123641.2
0.0
123641.2
--.-- --.11 Meta Consumo Human
0.0
723078.2
0.0
723078.2
0.0
723078.2
--.-- --.-----------------------------------------------------------------------------------------------------------------Total NPV Benefits
0.0
846719.5
0.0
846719.5
0.0
846719.5
--.-- --.External Supply
---------- Present Value of R&D Benefits ----------><---Costs---><------ Returns ------------>
Region
Producer
Consumer
Government
Total
(B-C)
B/C
IRR
-----------------------------------------------------------------------------------------------------------------20 Meta Oferta Extern
-544752.5
0.0
0.0
-544752.5
0.0
-544752.5
--.---.-----------------------------------------------------------------------------------------------------------------Total NPV Benefits
-544752.5
0.0
0.0
-544752.5
0.0
-544752.5
--.---.-

Note: The interpretation of negative benefits. This indicates how much the sector has shrunk (reduced revenue, lost market share) relative to
the situation without research. Thus, because the cost of production of traditional producers has not changed, they will lose revenue and
profits as prices fall (because of the expansion of technical production). In reality, the traditional prodcucers will move into their next
best farming activity – more cattle, cotton etc. The loss is the loss to the maize production sector.

145

Impact Assessment - Annual Report 2001

Table 5: Present Value of Gross Benefits for the Meta Department (excluding external supply region)
Total Scenario

1
2
3
4
5
6
7
13
14
15
16
17
18
19
8
9
10
11

---------- Present Value of R&D Benefits ----------><---Costs---><------ Returns ------------>
Region
Producer
Consumer
Government
Total
(B-C)
B/C
IRR
-----------------------------------------------------------------------------------------------------------------Mesetas
105756.4
0.0
0.0
105756.4
0.0
105756.4
--.---.Puerto Concordia
-7711.6
0.0
0.0
-7711.6
0.0
-7711.6
--.---.Puerto Rico
-3845.6
0.0
0.0
-3845.6
0.0
-3845.6
--.---.La Macarena
-4485.3
0.0
0.0
-4485.3
0.0
-4485.3
--.---.Uribe
-29412.4
0.0
0.0
-29412.4
0.0
-29412.4
--.---.Villavicencio
137317.7
0.0
0.0
137317.7
0.0
137317.7
--.---.Vistahermosa
42170.6
0.0
0.0
42170.6
0.0
42170.6
--.---.Fuente de Oro
263320.4
0.0
0.0
263320.4
0.0
263320.4
--.---.Cumaral
36125.1
0.0
0.0
36125.1
0.0
36125.1
--.---.San Juan de Arama
35938.5
0.0
0.0
35938.5
0.0
35938.5
--.---.Granada
415033.5
0.0
0.0
415033.5
0.0
415033.5
--.---.Lejanias
77906.7
0.0
0.0
77906.7
0.0
77906.7
--.---.Mapiripan
-17018.6
0.0
0.0
-17018.6
0.0
-17018.6
--.---.El Castillo
60320.9
0.0
0.0
60320.9
0.0
60320.9
--.---.Otros Tecnificado
62245.7
0.0
0.0
62245.7
0.0
62245.7
--.---.Otros Tradicional
-16385.7
0.0
0.0
-16385.7
0.0
-16385.7
--.---.Meta Consumo Anima
0.0
123641.2
0.0
123641.2
0.0
123641.2
--.---.Meta Consumo Human
0.0
723078.2
0.0
723078.2
0.0
723078.2
--.---.-

-----------------------------------------------------------------------------------------------------------------Total NPV Benefits
1157276.4
846719.5
0.0
2003995.9
2003995.9
--.---.-Annual Benefit Summary Across All Regions in the Scenario
<--------------- Benefits ----------------->
Year
Producer
Consumer
Government
Total
---------------------------------------------------1998
2933.6
3980.7
0.0
6914.4
1999
12524.5
17618.8
0.0
30143.4
2000
39651.6
55346.0
0.0
94997.7
2001
70854.2
98028.3
0.0
168882.6
2002
84365.0
116308.3
0.0
200673.3
2003
87781.6
120946.8
0.0
208728.5
2004
87784.3
121000.2
0.0
208784.6
2005
87774.3
121038.2
0.0
208812.5
2006
87746.6
121053.7
0.0
208800.3
2007
87705.4
121052.8
0.0
208758.3
2008
87647.3
121030.7
0.0
208678.0
2009
87576.0
120992.1
0.0
208568.2
2010
87487.2
120931.5
0.0
208418.7
2011
87386.0
120855.6
0.0
208241.7
2012
87267.5
120757.8
0.0
208025.3
2013
87136.8
120645.0
0.0
207781.9
2014
86989.1
120510.4
0.0
207499.6
2015
86828.8
120360.2
0.0
207189.0
2016
86652.3
120189.5
0.0
206841.8
2017
86463.5
120003.4
0.0
206466.9

146

Impact Assessment - Annual Report 2001

Table 6: Example simulation results for a single technical maize municipio – Mesetas
Region: 1

Mesetas

Price wedge information: Relative to Mesetas: Price difference = 0 $1000

Transmission = 1.00

---------------- Producers -------------- ------------- Consumers ------------------ ----- Government ---- ---------<---no R&D---> <------- with R&D ------> <-- no R&D ---> <------- with R&D ------>
< Tax Rev.>
Research
Year Price Quantity Price Quantity Benefits Price Quantity Price
Quantity Benefits "Prod" "Cons" Benefits
Costs
----------------------------------------------------------------------------------------------------------------------------1998
240.7
2241.6 240.6
2242.6
483.1 240.7
0.0 240.6
0.0
0.0
0.0
0.0
0.0
0.0
1999
241.2
2244.0 240.8
2248.3
2104.5 241.2
0.0 240.8
0.0
0.0
0.0
0.0
0.0
0.0
2000
241.6
2246.0 240.5
2259.7
6659.4 241.6
0.0 240.5
0.0
0.0
0.0
0.0
0.0
0.0
2001
242.0
2247.7 240.1
2272.1
11883.3 242.0
0.0 240.1
0.0
0.0
0.0
0.0
0.0
0.0
2002
242.2
2248.9 240.1
2278.0
14190.1 242.2
0.0 240.1
0.0
0.0
0.0
0.0
0.0
0.0
2003
242.4
2249.9 240.2
2280.3
14842.0 242.4
0.0 240.2
0.0
0.0
0.0
0.0
0.0
0.0
2004
242.5
2250.4 240.4
2280.9
14931.2 242.5
0.0 240.4
0.0
0.0
0.0
0.0
0.0
0.0
2005
242.6
2250.6 240.4
2281.3
15017.9 242.6
0.0 240.4
0.0
0.0
0.0
0.0
0.0
0.0
2006
242.5
2250.4 240.4
2281.3
15101.1 242.5
0.0 240.4
0.0
0.0
0.0
0.0
0.0
0.0
2007
242.4
2249.8 240.4
2280.9
15181.5 242.4
0.0 240.4
0.0
0.0
0.0
0.0
0.0
0.0
2008
242.2
2248.9 240.2
2280.2
15258.5 242.2
0.0 240.2
0.0
0.0
0.0
0.0
0.0
0.0
2009
242.0
2247.6 240.0
2279.1
15332.6 242.0
0.0 240.0
0.0
0.0
0.0
0.0
0.0
0.0
2010
241.6
2246.0 239.6
2277.6
15403.1 241.6
0.0 239.6
0.0
0.0
0.0
0.0
0.0
0.0
2011
241.2
2244.0 239.2
2275.8
15470.7 241.2
0.0 239.2
0.0
0.0
0.0
0.0
0.0
0.0
2012
240.7
2241.6 238.8
2273.6
15534.5 240.7
0.0 238.8
0.0
0.0
0.0
0.0
0.0
0.0
2013
240.1
2238.9 238.2
2271.1
15595.5 240.1
0.0 238.2
0.0
0.0
0.0
0.0
0.0
0.0
2014
239.4
2235.8 237.6
2268.2
15652.6 239.4
0.0 237.6
0.0
0.0
0.0
0.0
0.0
0.0
2015
238.7
2232.4 236.9
2265.0
15706.5 238.7
0.0 236.9
0.0
0.0
0.0
0.0
0.0
0.0
2016
237.9
2228.6 236.1
2261.4
15756.7 237.9
0.0 236.1
0.0
0.0
0.0
0.0
0.0
0.0
2017
237.0
2224.5 235.3
2257.6
15803.8 237.0
0.0 235.3
0.0
0.0
0.0
0.0
0.0
0.0
-----------------------------------------------------------------------------------------------------------------------------Present value of benefits
105756.4
0.0
0.0
0.0
-- Price --- ------------- Production ----------- ----------- Consumption -----------Year
K
K
K
K
--R&D change in- Value of Benefits -R&D change inValue of Benefits
Total Base Spill Var NoR&D
R&D
Price Quantity Productn /VoP (%) Price Quantity Consumptn /VoC (%)
---------------------------------------------------------------------------------------------------------------------1998
0.29 0.29
0.0 0.00 240.7 240.6 -0.07
1.0
539622
0.0
-0.07
0.0
0
0.0
1999
1.28 1.28
0.0 0.00 241.2 240.8 -0.34
4.3
541568
0.3
-0.34
0.0
0
0.0
2000
4.02 4.02
0.0 0.00 241.6 240.5 -1.07
13.7
543643
1.2
-1.07
0.0
0
0.0
2001
7.12 7.12
0.0 0.00 242.0 240.1 -1.86
24.4
545635
2.1
-1.86
0.0
0
0.0
2002
8.44 8.44
0.0 0.00 242.2 240.1 -2.17
29.1
546967
2.5
-2.17
0.0
0
0.0
2003
8.77 8.77
0.0 0.00 242.4 240.2 -2.22
30.4
547863
2.7
-2.22
0.0
0
0.0
2004
8.77 8.77
0.0 0.00 242.5 240.4 -2.18
30.5
548366
2.7
-2.18
0.0
0
0.0
2005
8.77 8.77
0.0 0.00 242.6 240.4 -2.14
30.7
548649
2.7
-2.14
0.0
0
0.0
2006
8.77 8.77
0.0 0.00 242.5 240.4 -2.10
30.9
548619
2.7
-2.10
0.0
0
0.0
2007
8.77 8.77
0.0 0.00 242.4 240.4 -2.07
31.0
548355
2.7
-2.07
0.0
0
0.0
2008
8.77 8.77
0.0 0.00 242.2 240.2 -2.03
31.2
547794
2.7
-2.03
0.0
0
0.0
2009
8.77 8.77
0.0 0.00 242.0 240.0 -1.99
31.4
547001
2.8
-1.99
0.0
0
0.0
2010
8.77 8.77
0.0 0.00 241.6 239.6 -1.96
31.6
545901
2.8
-1.96
0.0
0
0.0
2011
8.77 8.77
0.0 0.00 241.2 239.2 -1.92
31.8
544588
2.8
-1.92
0.0
0
0.0
2012
8.77 8.77
0.0 0.00 240.7 238.8 -1.89
32.0
542972
2.8
-1.89
0.0
0
0.0
2013
8.77 8.77
0.0 0.00 240.1 238.2 -1.85
32.2
541149
2.8
-1.85
0.0
0
0.0
2014
8.77 8.77
0.0 0.00 239.4 237.6 -1.82
32.4
539030
2.9
-1.82
0.0
0
0.0
2015
8.77 8.77
0.0 0.00 238.7 236.9 -1.78
32.6
536699
2.9
-1.78
0.0
0
0.0
2016
8.77 8.77
0.0 0.00 237.9 236.1 -1.75
32.8
534094
2.9
-1.75
0.0
0
0.0
2017
8.77 8.77
0.0 0.00 237.0 235.3 -1.72
33.0
531284
2.9
-1.72
0.0
0
0.0

147

Impact Assessment - Annual Report 2001

Table 7: Example simulation results for a single traditional maize municipio – Puerto Concordia
Region: 2

Puerto Concordia

Price wedge information: Relative to Mesetas: Price difference -0.7 $1000

Transmission = 1.00

---------------- Producers -------------- ------------- Consumers ------------------ ----- Government ---- ---------<---no R&D---> <------- with R&D ------> <-- no R&D ---> <------- with R&D ------>
< Tax Rev.>
Research
Year Price Quantity Price Quantity Benefits Price Quantity Price
Quantity Benefits "Prod" "Cons" Benefits
Costs
----------------------------------------------------------------------------------------------------------------------------1998
240.0
524.3 239.9
524.2
-41.9 240.0
0.0 239.9
0.0
0.0
0.0
0.0
0.0
0.0
1999
240.5
524.9 240.1
524.5
-182.3 240.5
0.0 240.1
0.0
0.0
0.0
0.0
0.0
0.0
2000
240.9
525.4 239.8
524.2
-562.7 240.9
0.0 239.8
0.0
0.0
0.0
0.0
0.0
0.0
2001
241.3
525.8 239.4
523.7
-978.7 241.3
0.0 239.4
0.0
0.0
0.0
0.0
0.0
0.0
2002
241.5
526.0 239.4
523.7
-1141.0 241.5
0.0 239.4
0.0
0.0
0.0
0.0
0.0
0.0
2003
241.7
526.3 239.5
523.9
-1166.2 241.7
0.0 239.5
0.0
0.0
0.0
0.0
0.0
0.0
2004
241.8
526.4 239.7
524.0
-1146.8 241.8
0.0 239.7
0.0
0.0
0.0
0.0
0.0
0.0
2005
241.9
526.4 239.7
524.1
-1127.3 241.9
0.0 239.7
0.0
0.0
0.0
0.0
0.0
0.0
2006
241.8
526.4 239.7
524.1
-1107.8 241.8
0.0 239.7
0.0
0.0
0.0
0.0
0.0
0.0
2007
241.7
526.3 239.7
524.0
-1088.2 241.7
0.0 239.7
0.0
0.0
0.0
0.0
0.0
0.0
2008
241.5
526.0 239.5
523.8
-1068.6 241.5
0.0 239.5
0.0
0.0
0.0
0.0
0.0
0.0
2009
241.3
525.8 239.3
523.6
-1049.1 241.3
0.0 239.3
0.0
0.0
0.0
0.0
0.0
0.0
2010
240.9
525.4 238.9
523.2
-1029.5 240.9
0.0 238.9
0.0
0.0
0.0
0.0
0.0
0.0
2011
240.5
524.9 238.5
522.8
-1010.0 240.5
0.0 238.5
0.0
0.0
0.0
0.0
0.0
0.0
2012
240.0
524.3 238.1
522.3
-990.5 240.0
0.0 238.1
0.0
0.0
0.0
0.0
0.0
0.0
2013
239.4
523.7 237.5
521.7
-971.1 239.4
0.0 237.5
0.0
0.0
0.0
0.0
0.0
0.0
2014
238.7
523.0 236.9
521.0
-951.7 238.7
0.0 236.9
0.0
0.0
0.0
0.0
0.0
0.0
2015
238.0
522.2 236.2
520.2
-932.5 238.0
0.0 236.2
0.0
0.0
0.0
0.0
0.0
0.0
2016
237.2
521.3 235.4
519.4
-913.3 237.2
0.0 235.4
0.0
0.0
0.0
0.0
0.0
0.0
2017
236.3
520.3 234.6
518.4
-894.3 236.3
0.0 234.6
0.0
0.0
0.0
0.0
0.0
0.0
-----------------------------------------------------------------------------------------------------------------------------Present value of benefits
-7711.6
0.0
0.0
0.0
-- Price --- ------------- Production ----------- ----------- Consumption -----------K
K
K
K
--R&D change in - Value of Benefits --R&D change in - Value of Benefits
Total Base Spill Var NoR&D
R&D
Price Quantity Productn /VoP (%) Price Quantity Consumptn /VoC (%)
---------------------------------------------------------------------------------------------------------------------1998
0.00 0.00
0.0 0.00 240.0 239.9 -0.07
-0.0
125788
-0.0
-0.07
0.0
0
0.0
1999
0.00 0.00
0.0 0.00 240.5 240.1 -0.34
-0.3
125983
-0.1
-0.34
0.0
0
0.0
2000
0.00 0.00
0.0 0.00 240.9 239.8 -1.07
-1.1
125751
-0.4
-1.07
0.0
0
0.0
2001
0.00 0.00
0.0 0.00 241.3 239.4 -1.86
-2.0
125412
-0.7
-1.86
0.0
0
0.0
2002
0.00 0.00
0.0 0.00 241.5 239.4 -2.17
-2.3
125381
-0.9
-2.17
0.0
0
0.0
2003
0.00 0.00
0.0 0.00 241.7 239.5 -2.22
-2.4
125504
-0.9
-2.22
0.0
0
0.0
2004
0.00 0.00
0.0 0.00 241.8 239.7 -2.18
-2.3
125622
-0.9
-2.18
0.0
0
0.0
2005
0.00 0.00
0.0 0.00 241.9 239.7 -2.14
-2.3
125688
-0.8
-2.14
0.0
0
0.0
2006
0.00 0.00
0.0 0.00 241.8 239.7 -2.10
-2.2
125681
-0.8
-2.10
0.0
0
0.0
2007
0.00 0.00
0.0 0.00 241.7 239.7 -2.07
-2.2
125620
-0.8
-2.07
0.0
0
0.0
2008
0.00 0.00
0.0 0.00 241.5 239.5 -2.03
-2.2
125489
-0.8
-2.03
0.0
0
0.0
2009
0.00 0.00
0.0 0.00 241.3 239.3 -1.99
-2.1
125305
-0.8
-1.99
0.0
0
0.0
2010
0.00 0.00
0.0 0.00 240.9 238.9 -1.96
-2.1
125050
-0.8
-1.96
0.0
0
0.0
2011
0.00 0.00
0.0 0.00 240.5 238.5 -1.92
-2.1
124745
-0.8
-1.92
0.0
0
0.0
2012
0.00 0.00
0.0 0.00 240.0 238.1 -1.89
-2.0
124369
-0.7
-1.89
0.0
0
0.0
2013
0.00 0.00
0.0 0.00 239.4 237.5 -1.85
-2.0
123945
-0.7
-1.85
0.0
0
0.0
2014
0.00 0.00
0.0 0.00 238.7 236.9 -1.82
-1.9
123453
-0.7
-1.82
0.0
0
0.0
2015
0.00 0.00
0.0 0.00 238.0 236.2 -1.78
-1.9
122911
-0.7
-1.78
0.0
0
0.0
2016
0.00 0.00
0.0 0.00 237.2 235.4 -1.75
-1.9
122306
-0.7
-1.75
0.0
0
0.0
2017
0.00 0.00
0.0 0.00 236.3 234.6 -1.72
-1.8
121653
-0.7
-1.72
0.0
0
0.0
Year

148

Impact Assessment - Annual Report 2001

Conclusions and Recomendation
The primary determinant of the quality of the economic evaluation is the consistency of
the technical R&D data and of the adoption data. It is important to elicit appropriate
information from scientists, extension workers, and other informed stakeholders. The data
preparation phase includes, as a minimum, an analysis of the trends in production, area
and yield, and the structure of prices, adoption and cost of the production where it is
possible. This informs the scientific debate, that is an essential part of the evaluation
process on the probable impact of new technologies - particularly in order to establish the
appropriate with and without R&D cases.
For the calculation of the benefits for geographic and socio-economic targets, nowadays
it is important to also consider the socio-economic distribution of the benefits as well as
its geographic and agro-ecological distribution. By default the benefits are calculated for
producers, consumers, and government sectors, but these groups can broken down further
to help to a better understanding. The examples used in other studies include the split of
consumers into urban and rural groups, and of producers into different scale of operation
and production systems - such as commercial irrigation or small producers in dryland (i.e.
Sanint and Wood 1998).
The method employed in the research study appropriate for this task for several reasons.
As was demonstrated in this study, each municipality can be represented separately in
terms of its own consumption, production systems, trade, generation and adoption of
technology. The results of the analysis include the changes in prices, production,
consumption and commerce for each region, as well as the benefits (B) at level of the
region in this case. This is extremely valuable to establish incentives to participate in
initiatives of investigation at a regional and sub-regional level
In order to obtain results that reflect reality in the regions, it is important to have
appropria te base information for the model. For maize it was possible to get information
detailed at the municipal level, on which the pertinent analyses for the definition of the
regions were made.
This information was useful to make necessary analyses of the trends in production,
consumption and prices, in the scenarios proposed without and with adoption of
technology.
From the point of view of the region, for those municipalities in whic h the technology is
adopted, the benefits are derived broadly in proportion to technology adoption. In the
case of maize in the Meta Department, benefits are also obtained because the increase in
production allows a reduction in the volume of imports (at the present time imports are
roughly 40 kg/per-capita.)
The benefits gained by the improved maize (technical) producers reflect advantages of
adopting new technology. On the other hand, traditional producers lose out because of

149

Impact Assessment - Annual Report 2001

new technology. Traditional producers lose benefits, relative to the situation without
research, because the research does not increase their yields or decrease their costs, but
the use of new technologies elsewhere is expanding production and reducing prices
(relative to what they would be without research). Thus traditional producers lose profits
and share of production.
One of the limitations of the methodology used is that the effects of R&D on other
commodities cannot be valued because the model does not allow the simultaneous
analysis of more than one product.
There are some methodological areas that could deserve attention:
•

The capacity to make evaluations that consider interaction between impacts of R&D
in multiple commodities (general equilibrium effects, for instance, would integrate
the market simulation, important secondary mechanisms and reaction, which would
serve to dissipate or to intensify the realistic impacts of R&D).

•

The spatial presentation of prices, for example, to take into account the costs of
transport and other transactions. Given the confidence already put in GIS technology,
its use can be extended (together with other information) to estimate the effects in
price formation related to distance, this would bring more realism to the market
responses. The present supposition of transference without cost between markets can
exaggerate, some times significantly, the potential for trade.

From the ecological point of view, there are high benefits from the adoption of new
technology, since this allows greater productivity, reflected in the crop yields. The new
production system will allow soil conservation through the fixing of nitrogen (in the new
maize/soybean rotation) and the improvement of organic material in the soil, increasing
the life of the soil and promoting sustainable agriculture.
By the end of the simulation period the Meta department was producing 20.541 tonnes
per year (15.238 tonnes of which came from technical maize producers) which is
equivalent to the 31% of the department’s total consumption.
The adoption of new technology lead to an increase in the annual production of maize in
the Meta department of 6165 tonnes over the simulation period (that is, the difference
between the total production for the department with and without the new technology).
By 2017 an extra 369 tonnes per year would be produced, made up of an increase of 388
tonnes per year from technical producers, and a decrease of 19 tonnes per year from
traditional producers.
By the end of the simulation period the total annual benefits to technical producers were
Col$M 180.563.9 per year, to consumers were Col$M 120.003.4 per year, and the losses
suffered by traditional producers were Col$M 9.146.6 per year.

150

Impact Assessment - Annual Report 2001

If we calculate the present value of the stream of benefits (and losses) over the whole
simulation period, using a discount rate of 10 percent, we find that present value of
benefits to producers are Col$M 1.236,1 (US$ 630.167), to consumers are Col$M 846,7
(US$ 431.669), and the present value of losses to traditional producers is Col$M 78.9
(US$ 40.203). Thus, the total present value across all groups is some US$ 1.021.633. 7 We
acknowledge, however, that the benefits of this research may well be larger because of
adoption beyond the Meta department, but we have no means of estimating what that
additional adoption might be.
A benefit-cost analysis of this research investment would suggest that the investment
would be worthwhile if the total cost of development and diffusion of the new technology
was less than US$M 1.02. Since we were unable to get detailed cost and personnel
figures attributable only to carrying out the research that underlies this new technology,
we can only estimate a general level of effort below which the R&D investment would
still be worthwhile.
Based on information from CIMMYT and CIAT, the following assumptions can be made
about direct research personnel costs per year: Ph.D. US$50.000, Research Assistant
US$18.000, Field Worker: US$10.000 (* 2 workers = US$20.000). This gives a basic
research team cost of $88.000 per year. If we add an arbitrary 50 percent for other
indirect costs, the total marginal cost of a full-time field based research team would be
some US$132.000 per year. On this rough basis, we can say that if the research required
less than (1.021.633 / 132.000) about 7.7 full time equivalent research team years, it was
profitable. If the research took more than the equivalent of 7.7 full time years of research
team time, it was likely not profitable (although, as we suggest we have ignored potential
benefits of the new technology beyond the Meta department).
The analysis described here is an ex-ante evaluation, trying to estimate the future effects
of a new technology. To make this analysis we had to make many assumptions about, for
example the likely rates and levels of adoption, and the growth rate of demand.
Obviously our results are sensitive to those assumptions, but other assumptions can be
easily tested in the modelling approach used. We can see in Tables 2 and 3 some of the
factors we can take into account; such as level of adoption of the new technology,
adoption speed, the market scenario and the macroeconomic and sectoral policies applied
in the country or region.
The utility of the analysis is based on its capacity to evaluate the economic potential that
a new technology offers, and this allows research managers to balance this potential
benefit against the likely research costs. As the actual diffusion and adoption process, it is
necessary to make ex-post studies to quantify the real levels of adoption and impact, to
evaluate that process and to extract lessons. On this basis, R&D managers and policy
makers can aim to improve the payoff to farmers and food consumers, as well as improve
the efficiency of technology generation and diffusion processes.
Donor: Interamerican Development Bank
7

All conversions based on an exchange rate of US$1 =$Col 1.962.

151

Impact Assessment - Annual Report 2001

Institutions Involved:
• IFPRI (Environment and Production Technology Division) and CIAT (Impact
Assessment Unit)
Donor: Interamerican Development Bank
People and Institutions Involved:
• Stanley Wood, IFPRI (Environment and Production Technology Division)
• Rob Schipper, Wageningen Agricultural University (Department of Development
Economics)
• CIAT (Impact Assessment Unit)
References Cited in Report
Alston, J.M., G.W. Norton, and P. G. Pardey. Science Under Scarcity: Principles and
Practices for Agricultural Research Evaluation and Priority Setting. Cornell
University. Ithaca. 1995.
Ardila J., Trigo E., and Pineiro M. Los Recursos Humanos en la Investigación
Agropecuaria: Tres Casos en América Latina. Desarrollo Rural en las Américas12:
233:58. 1980.
Cremers, M. and J. Roseboom. Investment Trends in Public Agricultural Research in
Latin America. Unpublished. ISNAR. The Hague. 1997.
Crabb R. The Hybrid Corn Makers: Prophets of Plenty. Rutgers University Press. 1947
CIAT. Trends in CIAT commodities. 1992.
FAO. Digital Soil Map of the World and Derived Soil Properties (Version 3.5). Base de
Datos en Formato del SIG grabado en CD. Rome. 1996.
FAO. Hojas de Balance de Alimentos – Promedio 1992-94. Rome. 1996.
FAO. Report on the Agroecological Zones Projects: Volume 3. Methodology and Results
for South and Central America. World Soil Resources Report 48/3. Rome. 1981.
Griliches Z. Research Costs and Social Returns: Hybrid Corn and Related Innovations.
Journal of Political Economics 66: 419-31. 1958.
Griliches Z. Hybrid Corn: An Exploration and Hybrid Corn and the Economics of
Innovations. Rural Sociology 24: 381-83. 1959.
Holdrige, L. Ecología Basada en Zonas de Vida. IICA. San José. Costa Rica. 1978.

152

Impact Assessment - Annual Report 2001

Kendall Hayes H. A Professor’s Study of Hybrid Corn. Minneapolis: Burgess Publishing
Co. 1963.
Lara L. Social and Economic Factors Related to the Adoption of Technology in Maize
in the District of Ariari in Meta. 1990.
Lima, M. and G.W. Norton. Determinación de Prioridades de Investigación
Agropecuaria en Venezuela. Fondo Nacional de Investigación Agropecuarias and
Internacional Service for National Agricultural Research, Maracay, Venezuela and
The Hague. 1993.
Mangeldsdorf Paul C. Hybrid Corn, Scientific American 1985: 39-47. 1951.
Palomino, J. and G.W. Norton. Determinación de Prioridades de Investigación en
Ecuador. INIAP-FUNDAGRO. Unpublished report. ISNAR. The Hague. 1992.
Pardey, P.G., and S.R. Wood. “Targeting Research by Agricultural Environments”.
Chapter 31. J.R. Anderson ed. Agricultural Technology Policy: Issues for the
International Community. Wallingford: CAB International, 1994.
Psacharopoulos, G., S. Morley, A. Fiszbein, L.Haeduck, B., S.R. Wood. Poverty and
Income Distribution in Latin America. The Story of the 1980s. Technical Paper No.
351. The World Bank. Washington. 1997.
Rovira, J.A. and H.M. Eder. Sistemas y Regiones Agrícolas de Venezuela. Fundación
Polar y Ministerio de Agricultura y Cria. Caracas. 1986.
Rogers E. M. and Ha Vernon V. Ruttan. Reforming the Global Agricultural Research
Support System. Issues in Third World Development, ed. Kennth C. Nobe and
Rajan K. Sampath (Boulder, Colo: Westview Press). 1983.
Sanchez, A.J. L. Arias and J. Comerma. Delimitacion y Definition de Unidades Agro
ecológicas: Metodología. Serie C No. 1-02 FONAIAP-CENIAP. Maracay. 1982.
Schipper R. De Evaluatie van Landbouwkundig Onderzoek in Ontwikkelingslanden.
Wageningen 1976.
Sanint, L. and S.R. Wood “LAC benefits from rice research for the last three decades”.
Chapter in “The Socio-Economic Impacts of Rice Research”. Conference
Proceedings. IRRI. Manilla. 1997.
Scobie G.M. and Posada R. The Impact of Technological Change on Income
Distribution: The case of rice in Colombia. American Journal of Agricultural
Economics 60: 85-92. 1978.
Spijkers P. Rice Peasants and Rice Research in Colombia. Wageningen 1983.

153

Impact Assessment - Annual Report 2001

Wohlgenant, M.K., “The Nature of the Research – Induced Supply Shift” . Invited paper,
proceedings of the Global Agricultural Science Policy for the 21st Century,
conference, Deparment of Natural Resources and Environment. Melbourn. 1996.
Wood, S.R. y W. Baitx. DREAM: Manual Y Guía Operativa. Informe BID-IICA-IFPRI.
Proyecto de Fortalecimiento de Capacidades y Aplicaciones para Priorizar
Investigación Agropecuaria en América Latina y el Caribe. IICA. Costa Rica. 1998.
Wood, S.R. and P.G. Pardey. Agroecological Dimensions of Evaluating and
Prioritizing Research from a Regional Perspective: Latin America and the
Caribbean. Discussion Paper 93-15. ISNAR. The Hague. 1993.
Wood, S.R. and P.G. Pardey. Agricultural Aspects of Evaluating Agricultural R&D.
EPTD. Discussion paper No. 23. International Food Policy Research Institute.
Washington. 1997. (Agricultural Systems forthcoming).
Wood S.R. 1998. Evaluación económica de nuevas tecnología agropecuaria:
multimercados, zonificación agro ecológica y transferencia de tecnología. Fondo
de Tecnología Agropecuaria (FONTAGRO), Working Document No. 3.
Washington 1998.

154