OUTPUT IV: INSTITUTIONAL CAPACITY
4.1. CIAT Capacity Enhancement
4.1.A. Method - Natural Resource Management - D. Pachico
Natural resource management research emerged in the mid 1980's as a major concern for
the global agricultural research system, both national programs and international centers.
Assessing the impact of this research inevitably became a critical issue, both for
prioritizing research and also for monitoring progress.
Impact assessment has become an especially thorny issue because natural resource
management research has represented an expansion of the research agenda for many
institutions a time when resources have become increasingly scarce. A variety of different
methods and tools for assessing impact have been utilized, including indicators; spatial
models and geographical information systems (GIS); economic surplus models; farmer
participatory techniques; and empirical field surveys.
This paper explores some conceptual and methodological issues related to the assessment
of impact of natural resource management (NRM) research and presents a brief overview of
some research approaches. The experience of CIAT (Centro Internacional de Agricultura
Tropical), with emphasis on Latin America, is used as a vehicle of this paper because it
illustrates a number of the issues that are confronting a wide range of institutions and
scientists today.
While CIAT can not offer a fully developed model that provides a comprehensive solution
to all the issues faced in impact assessment of NRM research today, examination of its
experience may serve as a useful platform to stimulate an exchange of experiences and
viewpoints in this workshop.
The first section of the paper attempts to appraise the broader context of impact
assessment for NRM research. It briefly reviews some dimensions, approaches, and aspects
of the reach of impact assessment of NRM research.
The second section of the paper sketches out the broad outlines of some methodological
approaches to NRM impact assessment. Ex ante approaches are considered, and some aspects
of poverty measurement, extrapolation, and cross scale analysis are noted.
The third and final section of the paper will present brief summaries of the findings
of some studies focussing on the impact of NRM research.
Impact Assessment: Desperately Seeking Donors
The high level of current attention that impact assessment of NRM research is
attracting, arises to a very substantial degree out of a concern by stakeholders to be
able to demonstrate to their constituents that investment in NRM research is yields
results in terms of the socio-economic development goals that motivate much research
investment.
Generally donors investing in NRM research are not interested in means such as
increased knowledge, better research methods or improved models. Rather, their investment
is usually driven to attain ultimate development impacts.
These impacts are diverse. They can include increased productivity and efficiency;
increased income or welfare for the poor; improved nutritional or health status; the
welfare of women or particular social groups; empowerment of the less privileged;
environmental quality; welfare of present or future generations.
Different investors in NRM research give different weights to these various ultimate
impact outcomes, so almost inevitably impact assessment means different things to
different people. These differences in what impact to assess, or how to measure it, are
rooted ultimately in differences in values or differences in utility functions as
economists might prefer to have it.
"At IDRC we do not strive for objectivity in the usual sense of being remote from
the project being evaluated. Evaluation reflects the values and perspectives of those who
design and implement the evaluation" (Love 1996). Thus, impact assessment, as a form
of project evaluation, like evaluation ultimately rests on a base of subjective values.
Even where there is consensus about what type of impact is being sought of NRM
research, the measurement of impacts can typically be approximated through a range of
different variables. For example, if increased income is the desired impact, this could be
measured as income in a current period, or future income over a number of periods, and
could involve asset accumulation or the measurement of income variability.
Likewise, if improved environmental quality is the desired impact, this can be measured
through a variety of indicators, for example, biodiversity might be assessed at the level
of the plant genome, species, or plant community or ecosystem. Similarly, soil quality
might relate to organic matter, profile depth, or soil texture. Gender welfare might be
appraised, for example, through access to income, control of assets, or participation in
decisions.
While impact assessment has to be based on some initial definition of impact regarding
what and how impact is to be measured, equally important is the specification of how the
output of NRM research will attain the desired ultimate impacts.
Development impacts like poverty alleviation, environmental quality, nutrition and
health, gender welfare, and empowerment, depend on a wide range of other factors besides
the outputs of NRM research. The overall performance of the general economy, episodes of
violence or warfare, national policy, and social and cultural structures and changes can
all have a major effect on the development impacts that NRM research is intended to
influence.
Frequently the effects of these other factors will mask or even dwarf those of NRM
research. Thus, even when NRM research may be making a contribution to the development
outcomes of interest to investors in research, assessment of this impact can be made
exceedingly difficult in the presence of other powerful intervening factors.
An Approach to Impact Assessment
Impact assessment does not optimally initiate as a retrospective exercise after a
research program has produced an innovation that has entered into use. Impact assessment
needs to be a crucial part of research planning and prioritization. Instead, optimal
impact assessment should be an integral part of an ongoing monitoring of research progress
and product development during the life of a research project. After the conclusion of a
research project, impact assessment provides valuable feedback on the effectiveness of a
research effort ( Pachico 1994).
Logically impact assessment begins with a definition of the desired impact; where it is
expected; and for whom it is anticipated. "Outputs can not be assessed fairly unless
they are identified from the start," (Goldsmith 1993). Clarifying from the beginning
the nature of the expected impact not only provides a basis on which the impact of NRM
research will be assessed during the course of the research and at its conclusion, but
also it can assist in the specification of research directions most suitable for attaining
the desired impact.
Thus, whether ex ante impact assessment is done through structured formal methods or is
based simply on the judgements and preferences of research managers and scientists, it
largely conditions what impacts can and will ultimately be achieved. Consequently, a
precise specification of the outputs expected of a research project outputs and a clear
tracing of the logical link between these outputs and particular desired impacts, is the
optimal base both for future impact assessment as well as for research planning and
prioritization.
Once the expected impacts of NRM research have been determined, then the challenge for
applied impact assessment then becomes the definition of how this impact will be measured.
Some impacts are relatively straightforward to measure, for example, the direct
consequences of adoption of an innovation such as changes in the frequency of spraying or
types of agrochemicals used due to an IPM program or changes in yield due to a new
variety.
However, the impacts of innovations on ultimate development outcomes like nutrition,
poverty, or gender welfare, are far more difficult and often quite costly to measure with
any precision. Since these are inherently complex phenomena, resort to indicators has
become a key tool for policy and decision-makers interested in impact assessment (Gallopin
1995).
The advantage of indicators is that they simplify information about complex phenomena
and the are "easily detectable, relatively simple, and cost effective, and should
resort, if possible to existing information"(Harrington et al 1995). An indicator can
be usefully defined as "an easily observed variable that may be measured at low cost
and is highly correlated with the state of a complex system of interest for decision
making", in this case, for impact assessment (Pachico 1996).
To be convincing, indicators need to be theoretically and logically linked, preferably
in some causal relationship, with the behavior of the complex system of interest. There
are a number of methodological issues involved in utilizing indicators for impact
assessment, some of which are touched upon below. Nonetheless, early identification of the
variables or indicators that will be used to measure impact is a key step in the impact
assessment process.
After a research program has been designed, including specification of its expected
outputs and their associated impacts on development objectives, prototype or intermediate
outputs will emerge from the research in process. The assessment of the performance of
these intermediate outputs will yield further information about the likely or potential
impact of a line of research.
Ongoing assessment of research in progress can serve both as a crucial feedback loop to
adjust or refine a scientific research project, and it can also lead to a reappraisal of
the impact of a research project, for example that it performs different from expected in
a particular environment or among farmers of a particular resource endowment. A variety of
methods including regional trials, on-farm trials, farmer evaluations, participatory
research and surveys can be useful in these continuing assessment of research in process.
As finished products emerge from a research program and diffuse among users, ex post
studies of the impact that has resulted from this research is a culmination of the impact
assessment process. It is important to note that measurement of the adoption or use of an
innovation is not the same as measuring its impact on natural resources management or
development goals. Measurement of outputs does not constitute measurement of outcomes. For
example, farmers may be observed to use soil conservation measures, but this is not the
same as measuring the impact of the conservation measures on the quality of the resource
base, or the impact of the improved resource base on farmer welfare, be it farm
productivity, income, or nutrition.
In sum, impact assessment is best conceived as an integrated dimension of the entire
research process, ranging from planning and prioritizing research, to monitoring research
progress, to appraisal of the ultimate outcomes or impacts of research. These are related
in that information from intermediate evaluations and adoption or impact studies can feed
back to modify ex ante assessment of ongoing or future research.
Ex ante assessment helps to appraise the selected indicators that can be used to
measure the expected impacts of a research project. Continuing assessment monitors
progress towards the production of a research output and can involve a preliminary
assessment of whether it is likely to deliver the expected impacts. Ex post impact
assessment addresses whether planned research outputs have in fact been generated and
used, and the degree to which this use has led to the expected impacts among various
potential beneficiaries
Reach and Impact Assessment
The discussion so far has focussed on what impacts are to be measured,
how they can be assessed, and when impact assessment is relevant with respect to the stage
of a research program. Equally important is the issue of the "reach" of the
outputs of a research program. Reach has been defined by Smutylo and Carden of IDRC as
"the groups that are touched by the results of a program".
Thus, the benefits of a NRM research project may reach producers and consumers, males
and females, upstream and downstream resource users, current and future generations, land
owners and laborers, and farms of different sizes, social status, or agro-environments.
Despite the substantial weight that many investors in NRM research may place on knowing
the distribution of benefits, the relatively greater tractability of measuring the
aggregate total of benefits of research programs, even though this is often far from a
trivial matter, has tended to absorb more attention than fully tracing out the
distribution of benefits. Although this may be more difficult and costly of research
resources, stakeholder interest in these issues may demand that they receive more
attention in the future.
Impact is also distributed spatially. Impacts occur at certain geographic locations,
each associated with particular environmental conditions and social groups. Utilizing
geographical information systems (GIS) can assist impact assessment, for example, by
helping to provide a framework for sample selection for impact study. Similarly, GIS can
assist in the extrapolation of impacts from specific locations to wider areas. Some
methodological aspects of using GIS in impact assessment are discussed below.
Impact occurs over many years, with significant time lags both between the initiation
of a research program and the production of the research output, and between the delivery
of a research output and its widespread diffusion among users. While NRM research is
likely to take significant time, it should be remembered that the genetic improvement
research of the CGIAR system that has had such an impact, took a minimum of a decade from
the outset of a program to the delivery of its first outputs, and in many cases it took
much longer. NRM research may not require a significantly longer horizon to produce
results than did the initiation of international crop improvement efforts.
Moreover, frequently there are significant time lags between the availability of a new
agricultural technology and its adoption. For a relatively simple technology in a
well-organized sector like irrigated rice in Latin America, peak adoption can be attained
within six years. For more complex technologies which lead to greater changes in the
production system or are more capital intensive, like forage technologies in Australia,
peak adoption may be reached only after 20 years. Because of their generally more complex
nature and the more gradual onset of observable impacts, the diffusion of NRM innovations
may more closely approximate the path of forages than rice.
Nevertheless, the impact of NRM innovations are often less immediate and therefore less
easily observed or measured. For example, decreased soil erosion or lessened loss of
biodiversity typically will not have an immediately observable impact on agricultural
productivity simply because the effects of resource degradation or its reversal occur over
extended periods of time.
Impact Assessment of NRM: CIAT’s Methodological Experience
In the context of the issues discussed above, this paper will now turn to a review of
CIAT’s experience with impact assessment of NRM research. This is done to share,
review and critique approaches and methods, without any intention to present a definitive
or normative model.
This section will be comprised of three parts. First, approaches to ex ante impact
assessment of NRM research will be briefly reviewed. Second, a method of measuring poverty
and extrapolating up from local perceptions of poverty to regional poverty profiles will
be briefly described. Third, methods for using the pressure-state-impact-response
framework across different scales of analysis will be considered.
Methods for Ex Ante Impact Assessment
The cornerstone of CIAT's Strategic Plan for the 1990's and Beyond was the integration
of a major research effort on resource management with its traditional efforts on
germplasm development. This was premised on the proposition that traditional research
paradigms based productivity considerations needed to give way to new technology design
conceptualizations that met ecological performance criteria in an agroecosystems
context(CIAT 1991).
To operationalize this vision, a major study was undertaken to identify research
problems and opportunities in natural resource management research (Jones et al 1991).
Broad environmental classes and within them, land use patterns (agroecosystems) were
defined using GIS. These agroecosystems were then prioritized for research based on a
number of criteria:
- Economic growth and resource potential
- Research ability of resource and agricultural problems
The first three criteria are essentially measures of the magnitude of the potential
impact of NRM research, while the last criteria is a measure of the probability of
achieving the potential impact. Taken together, the criteria would assess the expected
impact of natural resource management research. Scoring systems based on the above
criteria with different weighting schemes were used to assess the impact of conducting NRM
research for alternative agroecosystems (Jones et al 1991).
While there was broad acceptance within CIAT of this approach to set the initial
priorities for NRM research, it has been suggested that, "scoring should be used
sparingly. The results are unreliable and potentially very misleading," (Alston et al
1995 p. 487). In this view there is no adequate substitute for the use of economic surplus
models for impact assessment.
Such an approach was used to estimate the expected impact from alternative crop pasture
systems for the savannas of South America (Pachico et al 1993). Rates of return to
investment in crop systems were derived from shifts in the supply functions of the
commodities produced in the systems.
Nonetheless, it was judged that decision-makers were interested as much in poverty
alleviation and sustainability as in the crop productivity consequences of management
systems. Consequently, scoring was used to assess the effect of project outputs on
biodiversity, soil quality, water resources, pollution and pest ecology.
Extrapolating Local Perceptions of Poverty
Because poverty alleviation ranks high among the objectives of many investors in NRM
research, a better understanding of how improved NRM impacts on poverty is a crucial
issue. This in turn rests critically on the definition and measurement of poverty. Poverty
measurement confronts a number of difficulties in terms of the measurement of income, of
the importance of different measures of income relative to wealth or assets, and the
intra-household and especial inter-gender distribution of control over income and assets.
While external definitions of poverty such as income or expenditure are frequently used
as a basis for measuring poverty, and thus the impact of NRM research on poverty,
alternative approaches to poverty assessment are being explored (Ravnborg and Guerrero
1997).
Local perceptions are tapped through farmer interviews to develop indicators of well
being which can be used as indicators of poverty according to local perceptions rather
than externally derived understandings of poverty. In a case study in Honduras interviews
in a sample of 90 communities identified a number of indicators of well-being of
widespread relevance. These include indicators such as agricultural laboring, non-farm
income, land tenancy, livestock ownership, food availability, housing quality and crop
choice. These various indicators were combined into a well-being index (Ravnborg et al
1997).
Pair wise corrections were found between the ranking of indicators in 87 out of the 90
communities, thereby confirming that local definitions of well being are highly consistent
across communities. Furthermore, GIS analysis was used to extrapolate the poverty measures
nationally.
The sample communities were characterized on the basis of altitude, accessibility,
public services, ethnicity, gender, and population density. The same combinations of
sampling factors are found in a large number of communities outside the sampling area.
Thus, the GIS demonstrates where the poverty measures developed at the local community
level are likely to be valid given similarities in community characteristics.
The methodology described above attempts to provide insight into local perceptions of
poverty and to ascertain the extent to which such local perceptions can be extrapolated
nationally. Such an approach offers the prospect of a better capacity to link natural
resource utilization issues to the distribution of poverty through geo-referenced
databases. This in turn provides a powerful tool for assessing the impact on welfare of
innovations in natural resource management.
Measurement of Impact Across Scales
The measurement of impact across scales is a central issue in the assessment of the
impact of NRM research. For example, the impact of erosion on the natural resource base
varies from the plot to the slope face to the watershed to the river basin. It is
essentially a question of measuring impact across different levels of a systems hierarchy.
While this is most easily envisioned as aggregation across scales in a spatial hierarchy,
it is also an important phenomenon in other hierarchical systems. For example,
biodiversity occurs at a genetic level within a species, among species in a plant
community, and across plant communities within an ecosystem.
Figure 1 displays some cross scale vertical and horizontal linkages in the case of land
use. Vertically the scales extend from the regional to the global, while horizontally
impacts are shown from the environmental to the economic to the social. In the case of
land use shown here, a land use change such as deforestation at a local level has a series
of consequences for the local level (erosion), to the national level (loss of forests), to
the regional level (increased frequency of droughts and floods, to the global level
(climate change). Thus, both the nature and the magnitude of the impact of resource
management decisions varies by the scale of the analysis.
To enable decision makers to grapple effectively with this complexity, not just for
technology design but also for policy formulation and policy making, there is a need to
integrate these various scales and dimensions in a way that links the issues operating at
different scales while at the same time managing the complexity of the system and
communicating it in a comphrensible form. The challenge then becomes to develop
information systems that allow users the freedom to show many indicators at the same time,
to choose the scale and to make their own assessments, conclusions and decisions regarding
impacts.
A particularly appropriate conceptual framework for undertaking this challenge is the
pressure-state-impact-response model. This use of this approach is illustrated with
examples at two scales, one regional (the Amazon basin), the other national (Peruvian
forest margins). As the user increases the scale (i.e. moves from region to nation) the
nature of the indicators displayed changes to reflect the new scale.
For example, at the regional level in this land use example the pressure is increased
accessibility driven by improvements of the road transport network (Figure 2). This
accelerates colonization in the forests which changes the state, that is the surface area
of the forests. The impact of this is fragmentation of forests, and a possible policy
response is the development of legally protected areas. In each case indicators can be
developed for the pressure-state-impact-response.
Within the same Amazon basin at the national level of the Peruvian forest margins, a
different set of indicators for the pressure-state-impact-response model is displayed
(Figure 3).
Using Participatory Research for Impact Assessment
The importance for impact assessment of defining the "reach" of an
intervention in order to design the impact assessment strategy involves understanding who
the stakeholders are. Participatory approaches to impact assessment take this stakeholder
analysis one step further, in providing methods for impact assessment which actively
involve the stakeholder.
Involving stakeholders in assessing the impact of NRM interventions can be done from
the perspective of two main approaches which have different objectives. In the first
approach which can de referred to as functional participation, stakeholder involvement in
impact assessment is a means to an end --and is usually initiated by agents in need of a
quick and relatively easy way to obtain an impact assessment, which can be qualitatively
appealing. The second approach which is referred to as empowering, decision-making or
capacity building participation, is an integral component of a process of building
participatory and innovative management of natural resources, which requires stakeholders
to have a capacity to monitor and assess changes in the status of those resources.
Involving local organizations in the design of technical innovations for soil water or
forest conservation and in planning where to locate them in a landscape, is now widely
recognized as a key element of successful innovation. When local organizations involve
farmers in experimenting with principles of conservation and adapting these to meet their
own needs and constraints, then innovative improvements in NRM are rapidly developed and
adopted. Examples are the EPAGRI microwatershed catchment committees in Santa Catarina,
Brazil, the Catchment Committees of the Kenyan Ministry of Agriculture, the Agha Khan Rual
support Program in Pakistan, the CIPASLA watershed consortium in Colombia, Landcare in
Australia (Ashby and Ravnborg, 1998 ).
An important feature of local organizations in NRM is that they provide a conduit for
receiving and exchanging information about the impact of innovations which brings down the
cost for the individual stakeholder of experimenting and of enforcing collective
conservation. Participatory impact assessment, when it involves all the relevant
stakeholders making an impact on each other, provides information about transboundary
effects of NRM interventions enabling stakeholders to arrive at a joint plan of action.
Participatory impact assessment therefore, becomes a vital ingredient in a feedback or
learning process that increases the effectiveness of the overall participatory NRM
process.
Combining conventional and participatory monitoring and assessment of the transboundary
or off-site impact of NRM interventions into an ongoing NRM program can be especially
important when mobile as opposed to stationary resource flows are involved, or when
multiple competing uses exist for a given resource. For example in common property
resources with mobile flows that fluctuate unpredictably -- such as stream flow in a
watershed-- it is very difficult for users to assess the effects of use by one stakeholder
on the amount or quality of the resource (e.g. water) available to another stakeholder, or
the benefits to either user from a conservation intervention ( Schlager et al, 1992). When
cause and effect cannot be determined readily, conflict over usufruct rights is more
likely, and it is easier to free ride.
Thus the implementation of a locally managed participatory process of impact assessment
is an important element in sustained success of collective action to improve NRM. Landcare
in Australia is an example, in which conservation extension groups involving a broad cross
section of rural people with a stake in catchment planning are using techniques such as
GIS and aerial surveys in an extensive voluntary participatory environmental monitoring
and impact assessment.
CIAT's experience includes both functional and capacity building approaches to
participation in impact assessment. Participatory methods have been developed at CIAT to
provide ex ante assessment of the acceptability to farmers of conservation practices, and
these methods have been shown to be good at predicting future adoption behavior.
A capacity building participatory approach is central to CIAT's experience with the
development of community-based management of watershed resources for the Tropical American
hillsides. In this approach, stakeholder planning involves the definition of indicators
for monitoring the implementation and impact of NRM interventions. At the watershed level,
beneficiary assessment of improvement in their quality of life as a result of the changes
introduced in NRM is a basic feature of the approach. For specific projects which are
co-sponsored and cofinanced by the local watershed inhabitants and by external agencies, a
"commission" or task force of stakeholders is appointed by the watershed user
association to make regular site visits, using indicators for monitoring and assessment
which use local knowledge and also may draw on the GIS system developed by the
participating research agencies, such as the poverty (or wellbeing) mapping referred to
earlier in this paper. For example, CIAT is testing a "soil quality health kit"
designed for use by farmers (level 1) and by extensionists (level 2 of the kit) providing
simple diagnostic tools which enable users to build a rapid appraisal of the state of the
soils in a microcatchment when there are no accurate soils maps available.
Differentiating among groups of stakeholders in relation to poverty or gender and
understanding how impact is distributed among them is important for assessing the effect
of NRM interventions on equity, and it is integral to participatory impact assessment.
When stakeholder analysis is incomplete, and the relevant stakeholders are not represented
in a participatory impact assessment, then not only is the picture of impact obtained
likely to be incomplete, but the action based on the assessment is likely to be
ineffectual because some important actors have been left out.
CIAT's experience with stakeholder analysis illustrates this. Community-based
management of buffer zones around hillsides watercourses in Colombia failed to identify
semi-landless migrant farmers as Stakeholders. As a result, forest fires spread into the
buffer zones as a result of the traditional slash and burn land clearance by these
farmers. Subsequent stakeholder analysis enabled the participants to identify several
legitimate but conflicting interests. Burning on the neighboring agricultural land was
understood to be a rational practice for the farmers, but which had harmful externalities
for the community. A set of norms were developed specifying when and how burning could be
conducted and in some communities, groups formed to ensure compliance in making firebreaks
to protect the bufferzones (Ravnborg and Ashby, 1996).
CIAT's experience shows that stakeholder analysis needs to be conducted with two
objectives, each of which may be important and each of which needs to be evaluated
independently for its importance to impact assessment. One is whether the relevant
stakeholders in relation to a specified NRM problem are fully identified and represented.
The other is whether stakeholders bring relevant expertise to the problem. Representation
is important for the accuracy of any stakeholder-led participatory impact assessment, and
for future ownership by all stakeholders of any action based on that assessment. However,
the representative group may not be the most appropriate or most effective group for
tapping specialized knowledge, and it may be essential to select stakeholders with special
knowledge for certain assessments.
Some Case Studies in Ex Post Impact Assessment
While NRM research is a relatively new area of research, agricultural research has been
having an impact—both positive and negative—on the environment since its
beginning. The following examples of the impact of CIAT research on the natural resource
base demonstrate that the CG centers’ do in fact have a history of NRM impact from
which to draw upon. They also show that NRM research cannot easily be separated from
traditional commodity research, nor is such a distinction really necessary. Since many of
the most important NRM decisions are and will continue to be made in the context of
agricultural production, a more holistic approach to impact assessment may be most
appropriate.
1. Traditional commodity research with unanticipated environmental impacts
Traditional crop research has significant direct and indirect impacts on the natural
resource base. Some of these benefits are positive. For example, improvements in disease
resistance of irrigated rice has led to reduction in the use of pesticides and fungicides
in Colombia, Venezuela and Brazil (CIAT Rice Program Annual Reports). To the extent that
use of these chemicals had negative environmental impacts, the improved varieties had a
positive impact on management of natural resources.
In addition to the direct effects on chemical use, rice research has also had an
indirect effect on land use. Between 1966 and 1996 rice production in Latin America grew
by 2.5 percent a year, while price s fell by 30 percent. Eighthly percent of the
production increases came from yield increases rather than from expansion of area planted.
Irrigated rice yields almost doubled from 3 to 5.5 tons per hectare, while upland rice
yields remained unchanged. The result was that upland rice production became increasingly
unprofitable, leading to a reduction in rice production in the Brazilian Cerrados and
forest margins. Area planted to rice in these fragile areas declined from a peak of 6
million hectares in 1966 to 2.75 million hectares today (EMBRAPA data).
CIAT's research on improved forages for the South American savannas has included
substantial work on deep-rooted grasses of African origin. These grasses--Andropogon
gayanus and Brachiaria humidicola--were found to sequester significant amounts of organic
carbon deep in the soil. The researchers concluded that the process could account for the
sequestration of 100-507 Mt carbon per year in such improved pastures throughout South
America (Fisher et al). As such, it is quite likely that the research initiated largely
with the goal of increasing agricultural production is finding benefits in terms of
offsetting CO2 emissions from forest burning in the Amazon. Such an impact is of global
relevance, as well as ironic to the extent that these pastures may offset some of the
negative environmental results of conversion of tropical forests into pastures for cattle
production.
In other cases, the impact of commodity research may not be so unequivocally positive.
For example, irrigated rice and other improved crops are often heavy users of chemical
fertilizers, which can have direct, negative effects on environmental quality.
An impact assessment of an integrated production/processing project for cassava on the
north coast of Colombia found indirect, negative impacts on land use in the short run.
Adoption of the processing technology designed to help small farmers capture some of the
added value in processed cassava increased the demand for cassava, which led to an
expansion in cassava production at the expense of pasture and fallow. In the absence of
fertilizer or sound crop management practices, this intensification of production would
contribute to a decline in soil fertility.
The story doesn’t end here, however. Over time it was also found that farmers
invested their cassava earnings in livestock, resulting in the eventual reconversion of
cassava land back to pasture. It appears that farmers engaged in short term
"mining" of a resource in order to get money for a long term investment that
could lead to more sustainable production in the future as well as higher levels of
income. What this examples shows is that both short and long term impacts need to be
considered, along with the role played by factors such as income in decisions regarding
short run and long run natural resource use (Henry, Izquierdo and Gottret; Gottret and
Henry).
As these examples show, traditional crop research has important environmental
implications. While such impacts are often included in ex post project evaluations, what
remains to be addressed is the extent to which these types of concerns should be
considered ex ante and should form part of the traditional commodity research agenda.
Especially in the case of indirect effects, the causality that is relatively clear in
hindsight might be difficult to establish ex ante because the outcomes depend heavily on
the economic, political, and social parameters of the broader system in which agricultural
production takes place.
2. Joint NRM and commodity research
One way to incorporate the NRM concerns into traditional crop research is through joint
research projects. The important overlaps between the two research agendas offer many
opportunities for collaboration.
Cooperation among forage, rice and soils programs has yielded interesting results about
maintaining and enhancing soil quality in the tropical lowlands. Joint work by CIAT, CIRAD
and EMPRAPA’s Rice research program (CNPAF) showed that monocropping of upland rice
has been shown to have a devastating effect on the populations of earthworms, which play
an important role in maintaining soil structure. Rotations, on the other hand, can improve
macro and micro fauna in the soil. Improved pastures are also associated with increases in
the quantity and quality of worms (CIAT pastures and soils programs). What is clear is
that adoption of new practices and varieties has a positive impact on the natural resource
base. What remains is to measure the value of this improved soil quality through its
contribution to crop productivity.
Similar cooperative research is being conducted in Asia. Cassava farmers in Indonesia,
Vietnam, China, and Thailand are involved in different combinations of agronomic trials,
testing new cassava germplasm and ways to conserve soils through the use of vegetative
contour strips. Using simple plastic lined trenches instead of expensive and often complex
concrete trenches, collectors, and volume counters to measure soil losses, farmers can
see, measure, compare, and discuss results in the field. Participating farmers have
adopted different combinations of new cassava varieties, agronomic practices, and soil
conservation measures (PE-5 Annual Report, 1997.)
Farmers in Laos, Indonesia, the Philippines, Vietnam, Malaysia, Thailand, and China are
also participating in a "Forages for Smallholders Project." Farmers are able to
test forage germplasm for different uses ranging from improved dual-purpose livestock
feeding systems to improved fallows or legume cover crops. Researchers and farmers,
working as partners, are beginning to develop and adapt adaptable, problem-solving
forage-based technologies (PE-5 Annual Report, 1997).
CIAT is also a participant in the global project "Alternatives to
Slash-and-Burn". Much of the initial research has been to document the effects of
current land use patterns in terms of deforestation, carbon emissions, and biodiversity
losses. These measures will serve to help measure impacts as more sustainable land use
systems are developed and adopted. Farmer participatory research is being employed in the
Peruvian Amazon to develop such systems. Initial farmer testing of new rice and disease
resistant banana and plantain varieties is intended to address problems most mentioned by
farmers. As farmer and local national agricultural research system (NARS) confidence
increases in the use of participatory methods, the research will shift more to improved
fallows, secondary forest management, and the development of alternative crops and
products (Fujisaka et al, 1998).
CIAT and national scientists have identified and counted frequencies of plants in
different land uses--forest, cropping after forest, fallows of different ages, and
cropping after fallows--in Acre, Rondonia, Pucallpa, and Yurimaguas. These studies look at
plant species numbers, biodiversity losses in the affected areas, plant community
succession as land uses in slash-and-burn systems are forced to change over time, the
nature of weed invasions, and the effects of human exploitation of various forest products
on plant composition. Such research seeks to understand the impacts of current land use in
the forest margins in order to assist in the development of policies and technologies
meant to preserve biodiversity (Fujisaka et al, forthcoming).
The cooperative commodity/NRM research has lead to a better understanding of the causal
relationships between agricultural production practices and the environment. This
information is a critical input into the process of setting research and technology
development agendas. It also helps identify appropriate indicators of environmental impact
to use in the evaluation of research projects.
As mentioned earlier, a remaining challenge in this area of research is to identify
which of these biophysical relationships are also economically important. This is critical
for priority setting. CIAT and IFPRI are currently working on a project in which soil
erosion data will be incorporated into a traditional consumer-surplus economic impact
assessment model to try to capture environmental costs.
The emphasis on participatory methods in these cooperative projects also helps keep the
research focused on solving problems that farmers consider to be important. By targeting
research to the specific needs of farmers and by incorporating farmer’s knowledge and
experience into the technology development process, participatory research can lead to
shorter adoption lags and to higher levels of local adoption. However participatory
research methods have their own problems with regard to impact assessment, namely
difficulties associated with scaling-up the results beyond the initial study area. CIAT is
working on using GIS to identify regions which are similar in terms of key economic and
ecological characteristics to the study site. The results of such research would help to
target more precisely the release of new technologies and to increase the impact,
especially in the short run.
3. NRM research
Distinguishing between NRM and commodity research is difficult, however there are some
situations in which the short-term goals of traditional agricultural or environmental
research appear to diverge from those of NRM. Such situations demonstrate some interesting
characteristics of NRM research and highlight challenges to impact measurement.
In 1992, participatory evaluation of contour barriers to control erosion was carried
out in a pilot area in the Cabuyal watershed in southern Colombia. Many technological
options for controlling soil erosion are available to farmers, however adoption levels are
low. One reason is that such techniques are designed and recommended on the basis of
agronomic data, not necessarily on the basis of how well they meet farmers priorities. The
results of the farmers’ evaluations—which differed from those of the
scientists--were incorporated into recommendations given out by extension programs. Over
the next two years, adoption of contour barriers in this zone increased from 2 farmers to
261. More importantly, the farmers preferences as solicited in interviews were closely
correlated with what farmers actually adopted, suggesting that participatory evaluation
would be a valuable tool in estimating future adoption of technology and designing
technologies to maximize impact (Ashby et al, 1996).
The impact of local involvement at the policy level can also be seen in the Colombia
research site. In this watershed, an association was created which involves residents and
outside organizations working in the watershed. It provides a forum in which stakeholders
with diverse interests in the resources of the watershed can identify problems, set
priorities, and negotiate solutions. Through concerted action of this group, residents
achieved a change in the national policy regarding the maintenance of forested buffer
zones around watercourses. Colombian national environmental policy stipulated the creation
of buffer zones measuring 50 meters for springs and 30 for watercourses. Local residents
considered this to be too big, and since official enforcement capacity was weak,
compliance rare.
When this problem was taken up within the association, an agreement was reached with
the regional watershed authorities and the local water management board that sizes of
buffer zones would be determined on an individual basis, taking into consideration the
specific characteristics and features of the area. Technical advice would be provided to
help make these determinations. As a result, over the next 18 months, over 135 hectares
were enclosed and 150,000 trees planted. The community supplied 3,714 person days of labor
to the effort. The local-level consensus and support for the policy also allowed the
community to implement a highly effective enforcement mechanism. Anyone who did not
participate would have their water turned off by the local water authority. (Ashby et al,
1997 and Ravnborg and Ashby 1996)
Both these examples represent cases where the recommendations based on narrow
production or environmental protection considerations were rejected by farmers. On the
surface, this does not look like positive impact. However the in terms of the amount of
soil actually conserved or the number of hectares actually reforested, impact is higher
with the "second-best" practices because they were adopted. Natural resource
management problems are often characterized by the fact that the private and social costs
and benefits of economic decisions diverge. Since decisions are made by individuals,
special attention must be paid to the particular incentives facing each individual. An
environmental technology or policy that is incentive compatible—and therefore
feasible—may look quite different from one that derives from some concept of the
"social good" without regard for whether it can be feasibly implemented. In
reality, tradeoffs may have to be made between priorities of the different stakeholders in
order to assure implementation.
In terms of quantifying impact, NRM research faces many of the same valuation problems
mentioned in earlier sections. Since many resources do not have markets, it is difficult
to put a monetary value on them. Research carried out within the context of a particular
production system can be valued in terms of the contribution of the resource to
productivity. However much NRM research is not that site or system specific. Work is
underway by CIAT economists in Pullcalpa to use techniques such as contingent valuation to
try put a value on resources, and thereby provide a basis for assessing the impact of
projects that lead to changes in the resource base.
Finally, resource management associations such as the watershed user’s association
can have value to a community far beyond the achievement of their short-term NRM goals.
Improving the capacity of a community to organize and work cooperatively to achieve its
goals will help it to address other issues that collective action. Incentive problems
described above are not unique to NRM, and collective action is one way to help mitigate
their effects on economic efficiency. There is increasing evidence that high levels of
cooperation and organization can contribute to development (Johnson). Given the dynamic
nature of NRM and of the broader economic system, the real impact of strengthening the
ability of communities to better manage their own resources may be felt not through short
term conservation projects but through its contribution to the better long run use of all
community’s resources—natural, mechanical, and human. CIAT and the Rockefeller
Foundation are currently involved in research aimed at understanding and quantifying the
short and long run impact of this "social capital" on NRM and community economic
development.
Conclusions
Impact assessment of NRM research serves important functions of insuring the
effectiveness of NRM research while also demonstrating its returns to
stakeholder-investors. As such, impact assessment is best implemented as an integral part
of the entire research process rather than as a retrospective exercise that comes at the
end of a research project.
Investors in NRM research are motivated by a complex set of development goals, and
their interests are less on the outputs of the research process (scientific knowledge,
methods, even adoption of technology), than on the outcomes which occur as a result of
research. These impacts may be economic, social, environmental or institutional.
Careful linkage of the expected outputs of research with the intended development
outcomes is crucial to insure that research is planned so that it is directed at having an
impact on the outcomes of interest to investors, and also to insure that investors have
realistic expectations about how research can contribute to desired outcomes.
Particularly in cases of multi-stakeholder management of natural
resources, impact assessment can be of direct relevance to communities, not just to
scientists and research managers. Thus, participatory methods of impact assessment can
strengthen local capacity for improved resource management as well as function as feedback
to scientists and research managers.
Different indicators are appropriate for measuring various types of
impact. Ideally these indicators are specified at the outset of the research process; are
relatively inexpensive to measure; and correlate closely with the state of the underlying
complex development processes. Frequently impact assessment will need to use an array of
indicators to assess a variety of outcomes. These will differ in scale both geographically
and over time. Moreover, the distribution of impacts among social groups, including by
gender, is often of key interest.
A number of empirical studies can be cited where the impact of NRM
research has been assessed. These assessments have utilized a variety of indicators,
including biophysical (e.g. soil carbon), agronomic (e.g. yields, pesticide applications),
economic (e.g. profitability), and social (e.g. well being). Clearly impact assessment is
not the exclusive province of a particular set of specialized practitioners. It requires a
multi-disciplinary approach, and can well make use of a variety of methods including GIS,
participatory research, and bio-economic modeling.
Figure 1. Vertical and horizontal linkages: the case of land use
(Source: Winograd 1997)
Figure 2: Amazonian Forest - Pressure, State, Impact and Response
Indicators
2a. Pressure: Accessibility State:
Surface of Forests
2b.Impact: Fragmentation of Forests 2c. Response:
Protected Areas
(Source: Winograd 1998)
Figure 3: Peruvian Forest - Pressure, State, Impact and Response
Indicators
3a. Pressure: Frontier Forest under threat Accessibility.
Distribution of Fires (January 1993)
State: Surface of Agricultural or 'Altered' land
3b.Impact : Soil degradation - loss of topsoil/nutrients/organic matter,
Fragmentation of Forests
3c. Response: Protected Areas, Potential Yield
(Source: Winograd 1997)
|
