dealing with complexity, adaptability and continuity in agricultural research for development...

33
1 ILAC Working Paper 15 Dealing with complexity, adaptability and continuity in agricultural research for development organizations Javier Ekboir, Boru Douthwaite, Cristina Sette and Sophie Alvarez March 2013 Institutional Learning and Change (ILAC) Initiative - c/o Bioversity International Via dei Tre Denari 472°, 00057 Maccarese (Fiumicino), Rome, Italy Tel: (39) 0661181, Fax: (39) 0661979661, email: [email protected], URL: www.cgiar-ilac.org

Upload: independent

Post on 17-Nov-2023

0 views

Category:

Documents


0 download

TRANSCRIPT

1

ILAC Working Paper 15

Dealing with complexity, adaptability and continuity

in agricultural research for development

organizations

Javier Ekboir, Boru Douthwaite, Cristina Sette and Sophie Alvarez

March 2013

Institutional Learning and Change (ILAC) Initiative - c/o Bioversity International

Via dei Tre Denari 472°, 00057 Maccarese (Fiumicino), Rome, Italy

Tel: (39) 0661181, Fax: (39) 0661979661, email: [email protected], URL: www.cgiar-ilac.org

2

The ILAC initiative fosters learning from experience and use of the lessons learned to

improve the design and implementation of agricultural research and development programs.

The mission of the ILAC initiative is to develop, field test and introduce methods and tools

that promote organizational learning and institutional change in the CGIAR, and to expand

the contributions of agricultural research to the achievement of the Millennium Development

Goals.

Citation: Ekboir, J. et al 2013. Dealing with complexity, adaptability and continuity in

agricultural research for development organizations. ILAC Working Paper 15, Rome, Italy:

Institutional Learning and Change Initiative.

3

Table of Contents

Abstract ..................................................................................................................................... 4

Introduction ............................................................................................................................... 5

What are complex research organizations? ....................................................................... 7

Methodology ........................................................................................................................... 11

The evolution of international agricultural research and the origins of the CPWF ..... 12

The structure of the CPWF .................................................................................................. 15

What type of complex organization is the CPWF? .......................................................... 19

1.1 The unstable nature of the CPWF ............................................................................. 20

1.2 Dealing with organizational complexity ................................................................... 21

1.3 The strategy for organizational learning ................................................................... 22

Final remarks ......................................................................................................................... 27

References ............................................................................................................................. 30

4

Abstract

Complexity of research activities is increasing as multidisciplinary, multi-institutional time-

bounded research teams become more common. Some research organizations have responded

by introducing more flexible approaches to manage complex processes, creating new issues

related to team building, adaptive management and organizational learning. We explore a

complex research organization, the Challenge Program for Water and Food, a consortium that

researches multidimensional problems related to water management and poverty alleviation

in six large river basins. To deal with this complexity, the consortium created novel flexible

structures, but their implementation is not straightforward. Hastening adaptation of research

systems requires: balancing change and continuity, providing sufficient time for processes to

mature; building effective learning mechanisms that operate at different levels; recognizing

the constraints to adaptive management; aligning the incentives offered by the funding

agencies and the research organizations; building trust and commitment; and establishing

fluid communication among the partners.

Key words: fluid teams, adaptive management of research, complex organizations,

multidisciplinary research, CGIAR change process; organizational learning

5

Introduction

The complexity of research activities is increasing and multidisciplinary, multi-institutional,

time-bounded research organizations and teams are becoming more common. Some research

organizations have responded by introducing more flexible approaches to manage these

processes, which in turn has created new issues related to team building, adaptive

management and organizational learning. This paper analyzes the nature of fluid research

organizations and teams and their implications for research policies and management of

research. Additionally, it combines different disciplines to develop the basis of a framework

for the analysis of new types of research organizations.

Research systems have become more complex due to the emergence of research and

innovation networks that include private firms, public research institutes, universities and

civil society organizations (1), (2), (3); and a shift from disciplinary activities conducted in a

single laboratory to multidisciplinary, multi-institutional transient teams (4), (5).

Complexity theories (6) provide new perspectives on the implications of these changes. In

particular, they indicate that increasing the effectiveness of evolving research projects

requires coordination of several actors (stakeholders, administrators and researchers), team

building, appropriate incentives and procedures for resource allocation, committed

leadership, the creation of dedicated structures that foster collective action, organizational

learning structures that help to deal with unexpected events and time to allow the processes to

mature. These theories also indicate that even if all these factors are present, strengthening

the effectiveness of complex research organizations is not guaranteed due to path dependence

and the uncertainty members face about the lasting impact of the transient organization.

In complex systems coordination occurs through interactions between several components.

For example, human actions and interactions among variables that change at different speeds

(e.g., cultures, routines, and policies) provide the signals that induce individuals and

institutions to behave in particular ways (7). In research systems, in particular, the

interactions occur among stakeholders, research organizations and the researchers

themselves, and between rules and routines including policies, organizational dynamics and

changes in the way research is conducted. In turn, the interactions are influenced by a number

of factors including science and economic policies, markets, technical change and

globalization.

6

A fluid research organization is a particular type of complex organization meant to exist only

for a predetermined period to develop new knowledge about some particular issue, and its

members may change during that time (8). Fluid teams and organizations are becoming

increasingly common in research and international aid (9), (10), (11). Fluidity has advantages

and disadvantages. On the positive side, it can help to access new capabilities and more easily

tackle new tasks (12); on the negative side, it can lead to organizational forgetting (13), lack

of trust (9) and higher transaction costs. The actual effect of organizational complexity and

fluidity depends on an organization’s ability to facilitate collective action among partners

who have their own objectives, routines and incentives.

Complex research organizations have also been analyzed from different perspectives. (14)

analyze the effectiveness of a funding program that supports interdisciplinary research in

reorienting research themes and in strengthening accountability. The only instruments

mentioned are competitive grants mechanisms and a communication strategy to engage

stakeholders of research. (15) analyze how the selection mechanisms of a funding program

contributed to building trust among research partners. Finally, (16), (17) and (18) use

Principal Agent theory to analyze interactions between several actors in research system. All

three conclude that Principal Agent theory is not an appropriate theoretical framework to

analyze complex interactions in research systems.

All these papers focus mainly at the different impacts on funding. Our analysis, on the other

hand, has a broader focus because we look at multiple dimensions of organizational

complexity. Specifically, we analyze how a fluid research organization, the Challenge

Program for Water and Food (CPWF), embraced its complexity to a) coordinate different

partners, b) balance the benefits and costs of collective action, and c) explore new approaches

for research based on partnerships, adaptive management and participation.1 The CPWF is a

complex organization because a) it combines several stakeholders in formal and informal

interactions; therefore, different organizational cultures, incentives, routines and goals must

be coordinated; b) the participating researchers, organizations and stakeholders changed

during implementation, and are still changing; c) it has implemented procedures to learn (not

just monitor) how the new structure is working; d) it created a mechanism to adapt research

approaches on the basis of emerging information; and e) the CPWF has induced changes in

the research routines used by researchers from partnering organizations.

1 http://www.waterandfood.org/

7

Even though CPWF’s reforms have only been implemented for about two years (at the time

of writing), they are already yielding lessons of relevance. Due to the reforms’ short life,

many of the lessons reviewed in this paper refer to emerging problems. While most of the

benefits of the new approaches are expected to become clear later, some positive results have

also been identified. The CPWF leadership has made it clear that the program itself is an

experiment and, therefore, not everything that is tried is expected to work.

Section 0 reviews the literature on complex research organizations; section 0 describes the

methods used; section 0 describes the evolution of international agricultural research and the

creation of the CPWF; section 0 describes the current structure of the CPWF; section 0 uses

the information presented in the previous sections to analyze the issues emerging in the

implementation of the CPWF and section 0 discusses the lessons learned about the

management of complex research organizations.

What are complex research organizations?

We use a broad definition of an organization: a set of actors who collaborates over a

sustained period and develops shared cultures, communication codes and routines (3). The

collaborations may take different forms, including exchanges of information, joint priority

setting, and joint projects. It follows from the definition that organizations can be formal

(e.g., a National Research Council) or informal (for example, innovation networks). In fact,

organizations take many forms along a continuum that goes from a strict hierarchy to the

absolute horizontal. Informal organizations can emerge within formal structures; actually,

informal organizations complement and interlock formal organizations through a dense web

of personal relations (20). The definition implies that the organizational boundaries of

complex research organizations are defined not only by formal contracts but also by sustained

informal interactions (see below).

Complex research organizations take many forms, including innovation networks (2), fluid

teams (9), (8), inter-institutional research teams (10), (11), (5), and public - private

partnerships (21), (22). Complex organizations are multi-layered, i.e., several actors, teams

and individuals interact at different levels, such as partner organizations, senior management,

mid-managers and research teams. At the lowest level, complex organizations are formed by

teams integrated by individuals from one or several organizations.

8

Several factors have induced the emergence of fluid research teams and complex

organizations:

• To increase the relevance of research outputs, managers and funders have pressed for

stronger interactions among a wider range of stakeholders, including consultations, joint

research projects and participatory priority setting and evaluation.

• The reduction of public financing for agricultural research since the 1980s weakened

public research organizations (see section 0), forcing them to look for new partners.

Simultaneously, stakeholders demanded that public research organizations integrate into

innovation processes.

• Due to the increasing cost and complexity of innovation and research processes, no actor

commands all the resources it needs to innovate. Complex organizations enable partners to

share expensive equipment, specialists and the risks associated with research and

innovations.

• Interactions among professionals with diverse backgrounds enable access to a larger pool

of information and cross-fertilization among different disciplines, reducing the chances for

group-thinking and enabling a better exploration of alternative research paths. At the same

time, greater specialization of disciplinary research can hamper communication and

collaboration between researchers from different disciplines because they lack a common

language. Recent research has shown that the success of multidisciplinary teams depends

on a) the collaboration between a few generalists, who operate as “boundary spanners”,

and qualified open-minded specialists willing to explore the boundaries of their

disciplines, and b) the introduction of incentives that value participation in multi-actor

teams.

Despite their many benefits, complex research organizations are often unstable due to the

need to coordinate actors active at different levels (e.g., senior managers, middle managers

and researchers), changing priorities, short horizons, staff rotations and changing funding

opportunities. Staff rotations, in particular, are unavoidable because a) skill requirements

change as the research progresses, inducing some professionals to leave the team and others

to join it; b) efficiency considerations dictate that professionals participate simultaneously in

more than one project when they are required for less than full time in each of them; and c)

individuals leave in search of better professional opportunities. Organizational instability has

several problems (9), (12).

9

• When complex research organizations are structured around relatively short projects

(three to five years for agricultural research), members often do not have the time and/or

incentives to develop shared communication codes and trust in teammates. This, in turn,

increases the motivation for opportunistic behavior. Team building requires individuals to

invest time and effort, but the return to their investment becomes more uncertain when

the joint activities have a limited lifespan. This problem affects new and old team

members.

• When researchers have different expectations regarding the effort other team members

should exert, tensions and mistrust may emerge if the team does not have time to develop

shared expectations. The mistrust is greater when the researchers belong to different

organizations that offer conflicting incentives.

• Relatively short projects may not give new members time to integrate into existing teams,

even if they are willing to make the effort.

• Each team member has his/her own career expectations. These goals determine an

individual’s participation in complex research organizations, and their individual learning

needs and interests. The transitory nature of the partnership does not induce partners to

adjust their individual objectives.

• When researchers participate in complex research organizations, their primary allegiance

is to the organization that offers the best career opportunities, especially when career

advances are not linked to performance in multi-institutional teams.

• When a partnership involves several organizations, conflicts in objectives, cultures,

routines and incentives must be resolved, which, in turn, requires effective leadership,

negotiating skills, time and flexibility in the participating organizations.

• Coordination depends on developing shared objectives and routines. Instability increases

the transaction cost of coordination.

• Some research areas take long periods to produce results; for example, the study of forest

dynamics under different management strategies may take more than a decade. Such

research is very difficult to conduct in time-bound organizations or projects.

Recent publications have explored ways to strengthen complex research organizations (9),

(12), (10). In particular, it has been found that these organizations are more effective when

they have clearly defined objectives, such as developing a new medicine or a technology, but

10

they face great hurdles when they work on ill-defined issues such as management of natural

resources for poverty alleviation.

The effectiveness of complex organizations can be increased by (9):

1. Creating pools of people with similar capabilities that are supervised by a manager,

independently of the project they work on. The pool enables the members to share

experiences and develop trust, while the manager mentors the individuals and looks after

their career development. The pool allows coordination within the group and facilitates

coordination with other pools. Such pools are relatively common in fluid teams that

belong to permanent organizations that have a matrix structure (e.g., a national research

institute structured by regions and products), but are less common in transitory complex

organizations formed by several independent organizations.

2. Standardizing procedures, simplifying tasks and reducing interdependencies between

subprojects facilitates the integration of new people into existing teams and reduces the

impact of departing partners on the work of remaining members. While this approach can

be used in routine operations and in projects that are divisible (e.g., designing an

airplane), it is less effective when used in activities that are deeply integrated like

research to improve water use in agriculture.

3. Reducing interdependencies between project components and sub-teams also reduces the

need for coordination and for people working on different project components to develop

a shared vision of the project. For example, a telephone network technician does not need

to deeply understand the design of the whole network to repair individual pieces of

equipment (23). This approach is possible only if the processes are stable, modular and

well understood.

4. Creating the projects’ memories, for example, by developing knowledge management

platforms, and maintaining over time a pool of people with accumulated knowledge on

the processes.

5. Facilitating communication among team members, especially between incumbents and

newcomers. This can be done by explaining what a new member can contribute to the

team, establishing mentoring roles and joint planning and foresight activities.

The first three points are appropriate for repetitive, stable activities and cannot be easily

applied to agricultural research and innovation projects. Implementation of the last two points

requires a strong institutional commitment, effective leadership and sustained investments.

11

Complex processes, including complex organizations, have relatively long periods of stability

punctuated by major discontinuities and reorganizations; the reorganizations define possible

paths the process can follow when it stabilizes (7), (6). After a major reorganization,

managers are constrained by the new structure and can only make adjustments within those

boundaries, in other words; organizational processes are path dependent. This means that

organizations cannot change at will, and, adaptive management is constrained by past

changes and the organization’s culture, incentives and capacities.

Methodology

The question this paper seeks to answer is how a fluid research organization dealt with the

issues that emerged when it embraced its complexity to a) coordinate different partners, b)

balance the benefits of diversity with the costs of implementing collective action, and c)

explore new approaches for research.

Due to the complexity of the process, case study methodologies were followed (19). The

study comprised a single case, the CPWF. A conceptual framework was built with previous

knowledge of the CPWF and a thorough revision of the literature on the organization of

research, complex organizational change, and informal and fluid organizations. Empirical

information was used to revise the conceptual framework or when new issues were identified,

which in turn, led us to collect additional information and review of the existing evidence.

This cycle was repeated until a coherent case was built.

The empirical information was gathered from published documents, participation in

meetings, twenty-one semi-structured interviews with key stakeholders (including senior

managers, researchers and partner organizations), field visits and the authors’ own

experience. At the time of writing, one author was a member of the CPWF management team

and another participated in the knowledge management team. To ensure candid responses,

interviewees were promised anonymity. The interviews were conducted in 2011 and lasted

between one and two hours; some people were interviewed more than once.

The information was triangulated and double-checked with CPWF stakeholders to identify

inconsistencies and emerging questions. It must be stressed that we studied only a few

aspects of the CPWF. Important issues, such as its impact on poverty, food supply and water

management, and the generation of new research methods are left for further research.

12

The evolution of international agricultural research and the origins of the CPWF

To understand the dynamics of the CPWF it is necessary to review its parent organization, the

CGIAR, which itself has gone through several major reorganizations.2 The current change

process, started in 2008, is introducing new procedures for decision making and resource

allocation, is defining new principles to guide research activities, and is expanding the roles

stakeholders play in the system’s operations.

The origins of the CGIAR can be traced to the late 1950s when the Rockefeller and the Ford

Foundations teamed up with the governments of Mexico and the Philippines to create two

institutes to develop high yielding varieties of wheat, maize and rice. These institutes

developed the varieties that were a critical component of the Green Revolution (24). Other

centers were created in the 1960s to research issues related to tropical agriculture and new

donors, including the World Bank and governments from several developed countries, started

to contribute to the system. In 1971 the donors and centers created the CGIAR and expanded

the system’s goals and activities to include increasing the productivity of food legumes and

ruminant livestock; improving the management of natural resources; helping countries to

design and implement food, agricultural and research policies; building capacity and

strengthening National Agricultural Research Systems (NARS); and germplasm conservation

(25). Over the years, the number of centers under the CGIAR umbrella rose to 16.

The initial success of the CGIAR resulted from the collective effort of high-quality

researchers working on a narrowly defined problem (i.e., improved productivity of wheat,

maize and rice) and policy-makers providing the incentives (e.g., subsidies for fertilizer and

irrigation) and support services (e.g., public extension and seed delivery systems) to induce

adoption (26). In contrast with the initial focused mission, the objectives of the expanded

system were more diffused and spread the resources over more activities. Still, the vision

prevailing among stakeholders was that excellent, highly motivated and committed

researchers working with sufficient resources would develop superior technologies that

NARS could adapt to local conditions, and extension agents would then transfer to farmers

who would adopt them because they were more profitable (24).

2 The CGIAR is a global partnership that unites organizations engaged in research for sustainable development with the funders of this work; http://www.cgiar.org/index.html.

13

As the limitations of the model of the Green Revolution became apparent, new activities were

added to deal with particular problems; thus, agronomy programs were created to address

issues of local adaptation of high yielding varieties and social scientists were hired to study

adoption (or lack of) of modern agricultural technologies. GCIAR centers did not have to

justify their existence and donors consistently expanded their contributions to the

international agricultural research centers.

In the 1980s, research policies in many countries underwent major transformations, including

a shift from ‘blind’ funding of research institutions to project funding; competitive grants, in

particular, were seen as the golden standard for financing research (27). This change was

accompanied by a demand for research institutions to demonstrate the impact of their

activities, as evidenced by the discussions that followed the 1993 US Government

Performance and Results Act (28). CGIAR funders also adopted these principles, and funding

for agricultural national and international agricultural research contracted severely. In the

mid-1990s the centers began to transform themselves into output- and impact-driven

organizations (29); this move increased the system’s transaction costs, hampered long-term

research programs (26) and forced the centers to commit important resources to demonstrate

impact (30).

Several stakeholders (especially foreign donors) criticized the NARS for their lack of

participation in the emergence of high-value markets and the failure of modern technologies

to eradicate poverty. This led to a substantial downsizing of public research and extension

institutions (31). The CGIAR centers found that they could no longer rely exclusively on

weakened traditional partners, and started to work with private firms, NGOs, and farmer

organizations.

The CGIAR’s mandate expanded still further. The new activities included managing research

networks to facilitate research performed by others, some in conjunction with CGIAR centers

(32); rehabilitating seed stocks in natural disaster- or war-ravaged countries; managing

natural resources; and developing niche markets. The expansion in the scope of activities

carried out by the centers and in the number and types of partners the centers worked with

made most of the networks even more complex.

Following the weakening of the partners and a perception of limited impact, donors started to

demand that the CGIAR centers change their research approaches, especially developing a

more unified research portfolio, strengthening collaborations among them and demonstrating

14

their impact on poor farmers’ livelihoods. In the late 1990s the CGIAR launched the

“Ecoregional Programs”; due to changing priorities among the donors, this reform did not

have time to mature.

In 2001 the CGIAR launched the Challenge Programs (CPs), meant to provide additional and

stable funding to consortia of research organizations so that they would conduct

multidisciplinary, inter-institutional research on specific major issues affecting agriculture in

developing countries to generate more impact on development. The justification for setting

up the CPs included the expectation that they would improve the CGIAR’s relevance and

impact, help better target and integrate existing activities, achieve greater efficiency and

cohesion among CGIAR Centers, widen and improve partnerships with non-CGIAR partners

and mobilize more stable and long term financing. Some CPs were large-scale experiments in

using a multiple-source innovation model in a system more used to a centralized model (33).

Although the CGIAR centers were expected to play major roles in the CPs (each CP was

hosted by a CGIAR center), they had a management structures that were relatively

independent of the hosting centers, and they contracted established organizations to conduct

research that contributed to the CP’s goals.

In the early 2000s, several studies mentioned the progressive weakening of the CGIAR (34).

To address this issue, in 2008 the CGIAR launched the Change Management Initiative to

define a new business model. The core of the reform was the creation of a space where

donors could harmonize their policies (the Funders Forum); a Consortium that would unite

the centers under one legal entity and provide one entry point for funders to contact the

centers; and the Consortium Office that would coordinate and oversee the operations of the

individual centers. The final important element of the process was the creation of the CGIAR

Research Programs (CRPs). Similar to the CPs, the CRPs are partnerships created to address

specific issues related to the CGIAR mandate. Unlike the CPs, the CGIAR centers play a

major role in the management and implementation of the CRPs. The CPs are being

progressively absorbed by the CRPs.

The successive reforms of the CGIAR have increased the system’s complexity because the

mandate expanded, new partnerships were developed, new interactions among different

partners emerged, new institutional and governance arrangements were created and greater

coordination was imposed on the centers and researchers.

15

The structure of the CPWF

The CPWF is a CGIAR CPs. The CPWF is quite unusual in the sense that it is both a funder

of research and a research organization itself. The emphasis on its own organizational

learning to improve its performance makes the CPWF atypical.

The CPWF was launched in 2002 with the expectation that it would have three 5 year stages.

Phase 1 ran from 2004 to 2008 while phase 2 runs from 2009 to 2013 when the CPWF will

be absorbed by a CRP. Table 1 describes the CPWF’s evolution.

Table 1. Milestones in the evolution of the CPWF

2002 Formal creation of the CPWF 2003 Phase 1 of CPWF effectively starts 2004 First call for research proposals 2005 he management structure and operating rules are completed; second call for

research proposals 2006 First global meeting of CPWF researchers designed to facilitate exchanges of

experiences across basins 2007 and 2008

Participatory and demand-driven definition of Basin Development Challenges for phase 2

2008 Second global meeting of CPWF researchers designed to facilitate exchanges of experiences across basins

2008 End of phase 1 2008 Launching of the CGIAR change process 2009 Beginning of phase 2; new management structure established 2009 New partnerships and research approaches are established 2010 The decision to integrate the CPWF into a CRP is made 2011 Third global meeting of CPWF researchers designed to facilitate exchanges of

experiences across basins 2011 Integration of the CPWF board into a CGIAR center board 2012 Formal creation of the CGIAR CRPs 2012 The CRPs establish their own operational models

Currently, the CPWF works in six basins to address compelling basin development

challenges (BDCs) identified during phase 1. The CPWF set up work programs to address the

BDCs and all six are similarly organized: three to four research projects explore how the

improved management of water resources at different scales (farm, local, landscape and

institutional) can address the local BDC. An additional project coordinates the others,

facilitates learning and interacts with policy makers. Although all BDCs are similarly

structured, the actual research topics differ, depending on the challenges they tackle. For

example, two BDCs are “To reduce poverty and foster development by optimizing the use of

water in reservoirs” along the Mekong river and “Improving governance and management of

16

rainwater and small water infrastructure in the Limpopo basin to raise productivity, reduce

poverty, and improve livelihoods resilience”.

The CPWF’s formal organization is structured into three hierarchical levels and one cross-

cutting mechanism. The hierarchy is:

• A central team, made up of:

o The CPWF management team (CPMT) that sets and implements the overall

CPWF strategy

o A research management team that provides research leadership, mentoring and

oversight to the BDC programs

o A knowledge management (KM) team made up of specialists in communications,

information management and monitoring and evaluation that works to ensure an

evaluative, knowledge sharing and learning culture across the CPWF

o A secretariat that oversees activities and expenditures in all six basins

• The basin leaders (BLs) who interact upwards with the CPMT, sideways with other BLs

and downwards with the project leaders and project teams; they also promote learning

within the basin teams, and across the basins. Each BL is supported by a Coordination

and Change (C&C) project team.

• The project leaders (PLs) and their research teams

The cross-cutting mechanism is the Topic Working Groups (TWGs) which are communities

of practice where all members of the CPWF can discuss specific issues, regardless of their

geographic location or level of responsibility.

This structure was developed to support adaptive management and organizational learning,

shown in Figure 1. A key component of this structure is the C&C projects, led by the BL and

supported by the central team. Each C&C project identifies emergent problems and

opportunities in their geographic area of work, and makes sense of issues observed in the

research projects as these progress along their outcome pathways.

Figure 1: Networked structure of the CPWF in support of adaptive management and

organizational learning

17

Note: the text in red shows the learning actors and the text in black indicates the learning activities

Source: Adapted from personal communication with Rick Davies 2006.

Also based on phase 1 experience, the CPWF uses theory of change as a framework to

support organizational learning. A theory of change is a causal logic model that links research

activities to the desired changes in the actors’ behaviors that a project or program is targeting

to change. It describes the tactics and strategies thought necessary to achieve the desired

changes. A theory of change provides a road map of where the entity (e.g., research project,

BDC, global program) is trying to reach and how it expects to get there. The CPWF makes

theory of change explicit through outcome pathways (35). The CPWF posits that by being

explicit about the theory of change it is easier to see when progress towards development

outcomes starts to diverge from what was expected and early awareness of opportunity or

problems will support adaptive management.

Project

leader

learning

BDC

leader

and C&C

Cross-BDC

learning

through

CPWF central

team General

principles

derived at CPWF

Cross-

project

Project leaders and

researchers learning

at subproject level

(experiments,

18

Four parallel but interdependent types of research and learning activities are conducted within

the CPWF. Firstly, researchers research issues in their particular disciplines, often in the

context of multidisciplinary, multi-institutional teams (e.g., agronomy, ecology or

hydrology). Secondly, the research projects combine disciplinary research with social

sciences and action research to identify and test their theories of change. Thirdly, the C&C

projects, supported by CPWF Central Team, explore ways to build up organizational learning

and adaptive capabilities in the research projects and stakeholders. The C&C projects draw

on the projects’ experience to understand the contexts and mechanisms that enable the uptake

and use of research; they are also supported by the Learning-to-Innovate Topic Working

Group to carry out their research on adaptive management. Fourthly, the central team works

with the C&C teams to foster learning across basins and to learn itself how to adaptively

manage a large research for development project.

The creation of learning capabilities includes activities for reflection (e.g., bi-annual

reflection workshops in each basin; the International Water Forums, annual meetings of the

BLs and TWG leaders; semi-annual and annual reports, peer-assist and a flexible monitoring

system), mentoring, providing funds for learning activities, IT platforms to exchange

information within and between BDCs and visits to out-of-project sites. These activities also

have the objective of fostering integration of the research teams within and between basins,

but it is too soon to evaluate their effectiveness.

During inception of the BDCs emphasis was put on identifying inter-project dependencies on

data and method development. In practice, the CPWF found that projects needed to start

working before they could clarify what they could and could not do, and what collaboration

did and did not make sense. This experience was explicitly taken into account by some CRPs

which demanded an inception period before they could develop detailed multi-annual work

plans. While donors initially objected to the inception period, it was finally accepted.

Initial interactions among projects in the same basin were less than expected because:

a) Each project worked on a specific topic at a specific scale, and initially it was not clear to

the project leaders (PLs) how their projects fitted into the larger basin program.

b) Some projects could not wait for other projects to generate the information they needed;

therefore, they had to generate their own information.

19

c) The subprojects that composed the research projects were commissioned to teams who

had not worked together previously; therefore, communication channels among them had

to be developed.

d) The research projects in the first two BDCs that became operative were selected through

competitive processes; this made it difficult to maintain the logic on which the BDCs

were designed. In subsequent BDCs, the projects were commissioned, helping to maintain

the design logic; still, coordination problems persisted.

On reflection after two years of operations, the CPMT realized that it had probably been too

simplistic in its expectations for collaborations among projects in the same basin. The first

annual BDC reflection events showed that inter-project links and programmatic added value

was emerging naturally through: (i) having some people working in multiple projects; and,

(ii) developing common frameworks within the basin teams, e.g., for understanding

livelihoods (Mekong basin) and for integrated rainwater management (Nile basin).

No BDC expects long-term funding from the CPWF (i.e., beyond the end of phase 2).

Although the activities of the CPWF are being absorbed by a CRP, there is uncertainty about

how this will affect research activities, which in turn is affecting the commitment of

researchers and managers to the adaptive model of research.

In addition to the internal learning activities, the BLs and the communications team foster

learning by stakeholders in their basins, especially by policymakers. The learning instruments

include frequent personal interactions, participation in multi-stakeholder forums, document

repositories, web-based information exchanges and providing information.

What type of complex organization is the CPWF?

Five features make the CPWF a particularly complex organization: a) the research teams do

not belong to a single institution; i.e., the CPWF is a collection of teams from different

institutions linked by a common theme, formal contracts and the learning structure; b) the

research teams did not start with common objectives or a shared history; c) the CPWF is both

a research organization and an administrator of funds; d) the CPWF designed a novel learning

strategy to be implemented by a highly fluid (transient) structure; and e) the different teams

are coordinated by setting operating rules and routines, by actively exchanging information,

and by the support they get from the CPWF Central Team for organizational learning,

including reflection on how the research is contributing to the BDC’s goals.

20

While most of the issues raised in section 0 can be found in the CPWF, the solutions

proposed in the literature cannot be easily applied due to the program’s unique features. The

CPWF has addressed some of the issues in its own way, but others remain significant

challenges. This section reviews how three issues that emerged from the CPWF’s complexity

are influencing the research activities.

1.1 The unstable nature of the CPWF

The history of the CPWF is characterized by major discontinuities followed by periods of

relative stability (see table 1). In phase 1 the CPWF came under pressure to focus its

activities (26) and, as a result, at the start of phase 2 it dropped three of its original basins on

the basis of incidence of poverty and importance of transboundary issues. Then, the CPMT

used a number of sources of information (i.e., phase 1 results, a Comprehensive Assessment,

basin expert consultations, and an international forum) to identify the BDCs. Selection criteria

included broad stakeholder agreement on importance of the challenge, their motivation to

work on it, opportunity for the CPWF to contribute, and high impact potential. The transition

between phases was characterized by profound changes in management, research teams,

geographical focus and the way research was organized. While phase 2 was still in the

launching stage, the program started to be integrated into a CRP, which itself was in the

process of being launched (see section 0). The early transition has made the future of the

CPWF activities beyond 2013 highly dependent on priorities established by the new CRPs

and on funds made available by donors for specific basins’ activities.

Two other major changes in the CPWF’s operations were introduced in the transition from

phase 1 to phase 2. First, to bring more focus and relevance to research, this was reorganized

around the BDCs. Second, the research teams were also reorganized. Most members of the

management team in phase 2 participated in phase 1 either as researchers or as managers and

four out of the six BLs participated in phase 1. However, only about 18 percent of the

researchers participated in both phases, with just a few taking an active role in the

organization of phase 2.

In phase 2, the CPWF invested important resources in fostering interactions within and

between the projects (see section 1.3), but during the inception period, some PLs and

researchers had great difficulty in understanding how the different projects in each basin were

related. Two opposing attitudes have been observed; some researchers have fully embraced

the adaptive management concept and the need to interact with other projects and

21

stakeholders, while other researchers prefer to work in a more traditional organizational

setting; we could not assess what proportion of researchers belong to each group. Two major

questions will have to be answered in the near future. First, whether the efforts to foster

interactions and learning will induce more researchers to adopt the new research approaches,

given that the contracts are time-bound, the incentives offered to researchers by different

institutions often conflict among themselves, and the administrative requirements of some

contracting organizations are quite inflexible. Second, whether the new organizational

arrangements will actually result in different development outcomes and new research

methods and routines.

The CPWF plans to answer both questions in 2014. At the time this paper was being written

(second half of 2012) the CPMT and the BLs started a major ‘institutional history’ initiative

to learn from the CPWF experience, treating each BDC program as a case study.

1.2 Dealing with organizational complexity

The literature on organizational change and the CPWF experience indicates that a) even if

diversity among members of an organization is useful in theory (36), it may be less so in

practice due to coordination challenges and the looming end of CPWF’s phase 2; b) the fact

that a team has relevant knowledge does not guarantee its success in using it (12); and c) to

facilitate learning and the exchange of information, interactions within and between

components of the research program (basins in the case of the CPWF) need to be developed

at all levels of implementation (i.e., by a majority of researchers and stakeholders), but the

incentives a fluid research organization can offer may not be sufficient to induce changes in

ingrained behaviors.. The CPWF had to deal with these issues. It is common for

organizations to have more than one management structure (i.e., matrix structures); for

example, many firms have functional departments, such as sales or research, and thematic

groups, such as product lines. In the CPWF, however, these management structures do not

belong to the same organization, creating the potential for conflicting objectives and

incentives, on the one hand, and differing expectations about the members’ commitments on

the other. Similar conflicts have been reported in national research organizations where

researchers manage resources provided by external funders. This conflict is already present in

the CPWF where some managers expect a greater commitment from researchers who only

invest the minimal effort required to fulfill the contracts. The conflict is compounded because

the transitory nature of the CPWF and the CRPs conspires against building researchers’

22

allegiance to them (or at least, some of them), and because the BLs and PLs have few

instruments to induce greater participation, except the threats of cancelling the contracts or

not awarding future contracts.

The CPMT and the BLs have played the role of “boundary spanners” facilitating the

exchange of information within and between projects. The effectiveness of these activities

required substantial commitment and depended on the priorities of the individual BLs. In

some basins, all research teams have effectively shared information while in others some

projects were relatively neglected. Because the CPWF only finances a share of the basin

teams operations and it is a federated organization, the CPMT has a limited ability to

influence these behaviors.

During the design of phase 2, the CPWF negotiated with partnering organizations their

planned contributions. After the inception phase, however, the original design had to be

reviewed, including the participation of partners. The adjustment was not easy because,

among other factors, it involved budgetary reallocations. This, in turn, required the partners

to review their planned budgets and the number of researchers they could employ. As many

research positions in international research organizations are funded by projects, the

adjustment had immediate implications on the centers research capabilities. The partners’

dependence on short term funding was an important hurdle for adaptive management and

required strong negotiating skills.

1.3 The strategy for organizational learning

The organizational learning structure

A key element in the design of phase 2 was the creation of strong organizational learning

capabilities for adaptive management of research activities. The CPWF’s learning activities

are coordinated by the CPMT with the active involvement of the KM team and the basin

leaders (BLs) in the design, implementation and assessment of the activities. Although the

project leaders (PLs) and researchers participate in many learning activities, they have a

lesser role in their design and assessment. The reason for this pattern of participation is that

the CPWF was designed as a centralized network (Figure 1). The structure lends itself to a

learning strategy of decentralized experimentation with centralized learning (DECL). In this

structure, the basin projects implement their own decentralized research activities; the BLs

coordinate the work of the projects in each basin, and organize activities to foster the

emergence of a team spirit among the researchers; finally, the Coordination and Change

23

(C&C) projects support each BL in facilitating the analysis of these activities (i.e., making

sense of the process) and sharing of the lessons among basins. The CPWF central team

guides and supports the overall process.

Effectively, each research project is seen as an experiment, jointly assessed by the CPMT and

BLs. The learning structure is complemented by the topic working groups (TWGs).

Researchers’ time to engage in TWGs is not budgeted and depends only on their willingness

to participate. The first reaction of researchers to this novel structure has been mixed; some

researchers participate actively, others are fearful that their own organizations will not allow

them flexibility of operation and a third group sees it as too complicated and time-consuming.

In order to minimize the fear that the external reviews of activities for learning may cause,

the CPWF stresses that assessment for learning is different from evaluation for

accountability.

Action-research as an organizational learning strategy

Action- research on organizational learning in the CPWF is quite unique because it involves

self-experimentation, demanding from each stakeholder a serious effort to understand what is

expected from him and from his partners. As organizational changes require time, the process

moves slowly and is hampered by the diverse expectations of the different stakeholders. Also,

while this decentralized design allows for a broad exploration of approaches to do research

for development, it hinders the emergence of a sense of belonging to a joint innovation effort.

More than one year after the inception of phase 2, conversations with researchers made it

clear that while most basin teams have not yet developed a strong sense of unity or common

purpose, several teams are emerging.

Phase 2 was designed to facilitate organizational learning through experimentation, frequent

reflection and adaptive management. The innovation research carried out by both the Central

Team and the C&C projects addresses the same research question: how does research trigger

innovation? The Central Team seeks to answer the question by analyzing the six basin

development challenges as a set while the BLs work at the scale of individual projects.

Assessing the effectiveness of these organizational learning activities is particularly difficult

because their effects can only be gauged after the impacts of the funded research have

manifested themselves, usually well after the projects have ended. Hence the CPMT planned

to carry out case study analysis in 2014. Uncertainty of future funding has brought this

forward by one year.

24

The activities of the Central Team and the BLs learning are an example of double loop

learning, while some of the funded research is single loop learning. In double learning loops,

individuals and groups question assumptions, hypotheses, values, and policies that led to

actions taken in the first place, modifying what is necessary to increase effectiveness (37).

In essence, the Central Team and the BLs research by funding and overseeing learning

activities in the basins, inducing changes in the way the funded researchers operate, and

collectively assessing how those changes impact the ability of the research and management

teams to adapt their activities to emerging events. The five main research instruments used

for organizational learning are a) contracts that specify how the funded research should be

conducted; b) a set of activities to foster organizational learning (e.g., workshops, on-line

interaction tools and mentoring); c) close interactions between the basin teams and the

Central Team, d) coordination of research at different levels of the CPWF and e) stating from

the start of projects of the underlying theory of change in the form of outcome pathways that

is tested and updated during implementation as part of monitoring and evaluation

procedures.3

In addition to frequent direct communications among members of the basin teams, the BLs

interact assiduously with the Central Team. Each BL is directly supervised by a member of

the management team; as well as fostering organizational learning (essentially, by unifying

criteria and concepts), the interactions help to develop a sense of belonging to a larger

organization.

Hurdles for the implementation of Decentralized Experimentation and Centralized Learning

(DECL)

The Central Team and the BLs have invested substantial efforts in implementing the learning

structure and in fostering organizational learning. However, conversations with researchers

and PLs have shown two different responses to this organizational innovation. Some

researchers and PLs have embraced the idea of a learning organization, while others are

reluctant to participate in these activities and to manage their research adaptively.

Researchers in the second group mentioned that their main motivations for participating in

the CPWF were to obtain funds to cover part of their salaries, fine-tune known methods or

obtain data for scientific publications (required to advance their careers) rather than learning

3 These activities will be analyzed in a separate paper.

25

new ways to conduct research for development. They also manifested that the learning

activities required substantial amounts of time that were “diverted from research”. Future

research should monitor which of the two responses becomes prevalent among researchers.

The Central Team and BLs have been sensitive to the ‘transaction cost’ of putting in place a

learning structure. Both groups have reflected together on getting the balance right between

the demands of the research activities in the projects and the knowledge management

processes. The conclusion has been that both must be integrated and planned together from

program inception while the amount of effort put in towards ‘research’ versus ‘knowledge

management’ depends on where you are in the project cycle. The KM team has concluded

from feedback from project staff that while planning for future impact (through developing a

theory of change) is very useful during proposal and project inception, projects are more

receptive to research support during start-up. Once the research becomes clearer then projects

can be more specific about their theory of change, the outcome monitoring that is sensible to

carry out and the specific communication and outreach support they require. As time passes,

researchers become less amenable to change their research topics and methods.

The organizational learning mechanisms include the possibility of renegotiating the contracts

between the CPWF and the research organizations when the researchers can make a

compelling argument that a change has the potential to increase the CPWF’s impact. The

change in the contracts requires that the researchers negotiate both with the CPWF and their

permanent organizations. The flexibility in managing contracts can increase the CPWF’s

impacts but the double negotiation can raise transaction costs.

Finally, the incentives offered by the researchers’ centers can discourage adaptation because

administrators demand that researchers deliver the products promised in the original contract;

otherwise, the researchers can expect to receive a poor performance review. In these cases,

adaptation requires the formal modification of the original contract between the center and

the CPWF, which entails going through the centers’ contract approval procedures. The

researchers who are not inclined to change their research methods see many of the learning

activities underpinning adaptive management as bureaucratic requirements.

The loci of organizational learning

The structure for DECL, shown in Figure 1, is complex in the sense that complexity refers to

the nature of the process and the properties it exhibits and not to the existence of redundant

structures and bureaucratic requirements. Implementing it was a challenge even for the BLs

26

who were closely supported by the Central Team; in other words, even in teams supported by

strong mentoring, developing new routines requires time and effort. This phenomenon has

been explained in several ways. For example, organizational cultures emerge slowly and are

hard to change, but are not immutable (38). Alternatively, once an organization has found an

effective routine to achieve its objectives, it has few incentives to change what is working

until the environment has evolved enough to force the organization to change (39). Finally,

organizational change requires the abandonment of old capabilities and the development of

new ones, which demands time and resources (40).

The diversity of explanations demonstrates the complexity of implementing organizational

change. Despite these challenges, the Central Team was able to introduce important changes

in the ways the CPWF operated. For example, in the Andes, the second basin to become

operational, the projects were initially selected by a competitive mechanism. After it was

found that this mechanism hampered addressing the BDC with an integrated approach,

projects in other basins were commissioned.

The learning activities were more intense within the Central Team and among the BLs. The

activities that involved the basin teams have been more moderate and started later because

they required the BLs themselves to learn the new organizational model. It is too early yet to

assess organizational learning at the level of the basin teams but there are indications that the

model may need adjustments to take into account the issues mentioned in sections 1.1 and

1.2.

Knowledge management systems can partially compensate for the lack of a shared mental

model when the information is easily codified (9). The CPWF has invested substantial

resources in the creation of a communication structure, including electronic communications,

codification of tacit organizational routines and building document repositories. These

investments alone, however, have not been enough to substantially change many researchers’

approaches to research. The main reason for this result is that knowledge and research

routines cannot be easily codified nor are they repetitive; therefore, the effectiveness of

communication structures depends on a) the whole set of incentives for change offered to

researchers; and b) the implementation of other measures to foster organizational learning,

such as mentoring and peer review. It must be stressed that although KM systems by

themselves are not sufficient to induce organizational change, they should be an essential

component of any knowledge management strategy.

27

Final remarks

• The increasing complexity of research organizations is posing new challenges to policy

makers and research administrators. Adaptive management can coordinate the different

components of complex research organizations, but implementation of flexible strategies is

not straightforward. General policies, such as funding schemes, provide signals for

managers and researchers but do not induce effective coordination or rapid changes of

research cultures. To speed up adaptation of research systems it is necessary to implement

several simultaneous strategies including Balance change and continuity (6). Too much

change hampers learning as new experiences cannot mature; too much continuity makes the

organization unresponsive to emerging challenges and opportunities. The balance depends,

among other factors, on the maturity of the organization. In early stages, more exploration is

required to avoid being locked into inferior organizational arrangements; as information on

alternative arrangements is collected, the better ones can be identified and diffused across

the organization. In our case, the CPWF is still at the stage where it needs to change rapidly

to adapt the new learning strategy; the challenges will be to continue the strategy even after

the CP is absorbed by the CRP, and eventually develop more stable routines that still allow

for organizational learning.

• Provide sufficient time to allow processes to mature, which, in turn requires careful

consideration of whether goals and methods should be changed. In other words, how and

when should an organization decide that a particular goal or activity needs more time or that

it is the wrong approach? This decision involves careful assessment of the transaction costs

of any change that requires renegotiating contracts and incentives.

• Build effective learning mechanisms that combine interventions at different levels,

including support from the senior management, commitment of sufficient resources (not

only at the beginning of projects but also during implementation) and hiring professionals

with the appropriate capabilities. For example, the CPWF exploits the properties of complex

process with a DECL strategy, flexible contracting and active exploration of organizational

alternatives.

• Recognize the constraints that the dynamics of complex processes impose on adaptive

management. Big changes can be made only at long intervals, while relatively minor

adjustments can be introduced more often.

28

• Align the incentives offered by the funding agencies and the research organizations. If the

incentives offered by the organizations in the partnership are misaligned, the researchers

will chose the organization that offers the best career opportunities, including more job

security.

• Invest not only in strengthening human resources over the whole life of the partnership, but

also in effective programs to build trust and commitment, which in turn require activities

meant to create bonds that link researchers and managers. Particularly important is to a)

spend time and resources to jointly develop proposals with shared theories of change and

research approaches, b) jointly review the theories of change during implementation; and c)

use participatory and flexible management.

• Provide continuity despite the time limitations imposed by the funding cycles and external

shocks. While the CPWF was expected to have three phases, the CGIAR change process

introduced great uncertainty regarding the teams’ future, reducing the incentives for

researchers and partners to adopt new approaches for research,. Continuity of research

partnerships beyond the short projects also enables the creation of reputations, which, in

turn, has two important consequences. First, it allows researchers to build more innovative

and resourceful research programs to address complex issues. Second, performance in

previous projects can be an important consideration for funding projects that address

complex issues, replacing the requirement of specifying expected outputs and outcomes.

The impossibility of predicting future outcomes is an essential feature of complex processes

and the main rationale for adaptive management (6). The theory of contracts (41) and

principal-agent models of research (42) have shown that when the outcomes of a process

cannot be clearly specified in advance, incentives are an inefficient instrument to induce

effort and compliance; rewards based on past performance induce researchers to exert more

effort and to explore more innovative research paths.

• Evaluate the effectiveness of financing mechanisms in achieving the program’s goals; for

example, in the CPWF competitive funding did not effectively coordinate and aligned

research processes. Soon after starting phase 2, the CPWF switched from a competitive

mechanism to commissioned projects.

• Establish fluid communication among the partners, because operating on the enabling

environment, such as new funding mechanisms and incentives, is not enough to ensure

coordination. The reason is that actors participating in complex processes have their

29

individual goals, which they pursue conditioned by the enabling environment. Alignment of

diverse goals requires leadership, active management and negotiation.

• In complex research processes, advanced planning cannot identify all the issues that will

emerge after inception. The adjustment costs can be reduced by having an inception phase

to sort out emergent problems, especially reallocation of roles and resources among partner

organizations. Effective leadership is a key organizational resource to navigate the process.

It should be noticed that the interventions mentioned are not a checklist or a set that has to be

implemented in its entirety; rather, it intends to show the multiple dimensions of managing

complex research organizations.

30

References

(1) Mowery, D.C. and B.N. Sampat (2005) Universities in National Innovations Systems. In

Fagerberg, J., D.C. Mowery and R.R. Nelson (eds.), The Oxford Handbook of Innovation.

Oxford; New York: Oxford University Press.

(2) Powell, W.W. and S. Grodal (2005) Networks of Innovators. In Fagerberg, J., D.C.

Mowery and R.R. Nelson (eds.), The Oxford Handbook of Innovation. Oxford; New York:

Oxford University Press.

(3) Ekboir, J.M. and R. Rajalahti (2012) Coordination and Collective Action for Agricultural

Innovation. In Rajalahti, R. (ed.) Agricultural Innovation Sourcebook. Washington, D.C.:

The World Bank.

(4) Conraths, B. and H. Smidt (2005) The Funding of University-Based Research and

Innovation in Europe. An Exploratory Study. Brussels: European University Association. (5) Gibbons, M., C. Limoges, H. Nowotny, S. Schwartzman, P. Scott and M. Trow (1994)

The New Production of Knowledge. The Dynamics of Science and Research in

Contemporary Societies. London: Sage.

(6) Axelrod, R. and M.D. Cohen (1999) Harnessing Complexity. Organizational Implications

of a Scientific Frontier. NY: The Free Press.

(7) Holling, C.S., L.H. Gunderson and L.H. Ludwig (2002) In Quest of a Theory of Adaptive

Change. In Gunderson, L.H. and C.S. Holling (eds.). Panarchy. Understanding

Transformations in Human and Natural Systems. Washington, D.C.: Island Press.

(8) Huckman, R.S., B.R. Staats and D.M. Upton (2008) Team Familiarity, Role Experience,

and Performance: Evidence from Indian Software Services. Working paper 08-019.

Boston, MA.: Harvard Business School.

(9) Bushe, G.R. and A. Chu (2011) Fluid Teams: Solutions to the Problems of Unstable Team

Membership. Organizational Dynamics 40: 181-188.

(10) Bennett, L.M., H. Gadlin and S. Levine-Finley (2010) Collaboration and Team Science:

A Field Guide. Draft, Washington, D.C.: National Institutes of Health.

(11) Beardon, H. and K. Newman (2009) How wide are the ripples? The management and

use of information generated from participatory processes in international non-

governmental development organizations. IKM Working Paper No. 6, IKM Emergent

31

Programme; Bonn Germany: European Association of Development Research and

Training Institutes (EADI).

(12) Huckman, R.S. and B.R. Staats (2010) Fluid Tasks and Fluid Teams: The Impact of

Diversity in Experience and Team Familiarity on Team Performance. Working paper 09-

145. Boston, MA: Harvard Business School.

(13) David, G. and T. Brachet (2011) On the Determinants of Organizational Forgetting.

American Economic Journal: Microeconomics 3:100–123.

(14) Lowe P, and J. Phillipson (2006) Reflexive Interdisciplinary Research: The Making of a

Research Programme on the Rural Economy and Land Use. Journal of Agricultural

Economics 57: 165-184.

(15) Davenport S., J. Davies and C. Grimes (1998) Collaborative research programmes:

building trust from difference. Technovation 19: 31-40.

(16) Klerkx L. and C. Leeuwis (2008) Delegation of authority in research funding to

networks: experiences with a multiple goal boundary organization. Science and Public

Policy 35: 183-196.

(17) Braun D. (2003) Lasting tensions in research policy-making - a delegation problem.

Science and Public Policy 30: 309-321.

(18) Shove E. (2003) Principals, agents and research programmes. Science and Public Policy

30: 371-381

(19) Yin, R.K. (2003) Case Study Research. Design and Methods, 3rd. ed., Applied Social

Research Methods Series, California: Sage Publications.

(20) Hakansson, H. and D. Ford (2002) How should companies interact in business

networks? Journal of Business Research 55: 133– 139.

(21) Perkmann, M. and K. Walsh (2009) The two faces of collaboration: impacts of

university-industry relations on public research. Industrial and Corporate Change,

Industrial and Corporate Change, 18:1033-1065.

(22) Bozeman, B. and E. Corley (2004) Scientists’ collaboration strategies: implications for

scientific and technical human capital. Research Policy, 33:599–616.

32

(23) Narduzzo, A., E. Rocco and M. Warglien (2000) Talking About Routines in the Field. In

Dosi, G., R.R. Nelson. and S.G. Winter (eds.) The Nature and Dynamics of

Organizational Capabilities, New York: Oxford University Press, N.Y.

(24) Ekboir, J.M (2009) The CGIAR at a Crossroads: Assessing the role of international

agricultural research in poverty alleviation from an innovation systems perspective. ILAC

working paper 9. Rome: ILAC.

(25) Anderson, J. (1998) Selected Policy Issues in International Agricultural Research: On

Striving for International Public Goods in an Era of Donor Fatigue, World Development

26(6):1149-1162.

(26) CGIAR Science Council (2008) Report of the First External Review of the Challenge

Program on Water and Food (CPWF). Rome, Italy: Science Council Secretariat.

(27) Vera-Cruz, A., G. Dutrénit, J.M. Ekboir, G. Martinez V. and A. Torres Vargas (2008)

Virtues and limits of competitive funds to finance research and innovation in the

agricultural sector: the case of the Mexican Produce Foundations. Science and Public

Policy 35(7):501-513.

(28) Kraemer, S (2006) Science and Technology Policy in the United States. New Brunswick,

N.J.: Rutgers University Press.

(29) Kassam, A.H. (2006) Agricultural Institutions and Receptivity to Social Research: The

case of the CGIAR. In Cernea, M.M. and A.H. Kassam (eds.), Researching the Culture in

Agri-Culture. Social Research for International Development. Oxfordshire, U.K.: CABI

Publishing.

(30) Bellon, M.R., M. Morris, J.M. Ekboir, E. Meng, H. De Groote, G. and Saín (2006)

“Humanizing” Technology Development: Social Science Research in a Home of the

Green Revolution. In Cernea M. and A. Kassam (eds.), Researching the Culture in

Agriculture, U.K.:Oxon. CABI.

(31) Byerlee, D., G. Alex, and R.G. Echeverría (2002) The Evolution of Public Research

Systems in Developing Countries: Facing New Challenges. In Byerlee, D. and R.G.

Echeverría (eds.) Agricultural Research Policy in an Era of Privatization, Wallinford,

UK: CABI Publishing.

(32) Plucknett, D.L., N.J.H. Smith and S.Ozgediz (1990) International Agricultural

Research. A Database of Networks. Study Paper 26, Rome, Italy: CGIAR.

33

(33) Woolley, J. and B. Douthwaite (2011) Improving the resilience of agricultural systems

through research partnership: A review of evidence from CPWF projects. CPWF Impact

Assessment Series 10. Colombo, Sri Lanka: CGIAR Challenge Program for Water and

Food.

(34) Renkow, M. and D. Byerlee (2010) The impacts of CGIAR research: A review of recent

evidence. Food Policy 35(5): 391-402.

(35) CPWF (2011) M&E Guide. http://monitoring.cpwf.info/. Last visited 24/2/2012.

(36) Georgsdottir, A.S., T.I. Lubart and I. Getz (2003) The Role of Flexibility in Innovation.

In Shavinina, L.V. (ed.), The International Handbook on Innovation, UK: Pergamon, pp.

79-96.

(37) Argyris, C. and D. Schön (1996) Organizational learning II: Theory, method and

practice, Reading, Mass: Addison Wesley.

(38) Schein, E.H. (1991) Organizational Culture and Leadership, 2nd ed., San Francisco and

Oxford: Jossey-Bass Publishers.

(39) Christensen, C.M., S.D. Anthony and E.A. Roth (2004) Seeing What’s Next. Using the

Theories of Innovation to Predict Industry Change. Boston, Mass.: Harvard Business

School Press.

(40) Skarzynsky, P. and R. Gibson (2008) Innovation to the Core. A Blueprint for

Transforming the Way Your Company Innovates. Boston, Mass.: Harvard Business Press.

(41) MacLeod, W.B. 2007. Reputations, Relationships and Contract Enforcement, Journal of

Economic Literature XLV(3):595-628.

(42) Huffman, W. and R.E. Just (2000) Setting efficient incentives for agricultural research:

lessons from principal-agent theory. American Journal of Agricultural Economics

82(4):828-841.