An analysis of power dynamics within innovation platforms for natural resource management

Abstract

Innovation systems thinking is increasingly influencing approaches to sustainable agricultural development in developing world contexts. This represents a shift away from technology transfer towards recognition that agricultural change entails complex interactions among multiple actors and a range of technical, social and institutional factors. One option for practically applying innovation systems thinking involves the establishment of innovation platforms (IPs). Such platforms are designed to bring together a variety of different stakeholders to exchange knowledge and resources and take action to solve common problems. Yet relatively little is known about how IPs operate in practice, particularly how power dynamics influence platform processes. This paper focuses on a research-for-development project in the Ethiopian highlands which established three IPs for improved natural resource management. The ‘power cube’ is used to retrospectively analyse the spaces, forms and levels of power within these platforms and the impact on platform processes and resulting interventions. The overall aim is to highlight the importance of power issues in order to better assess the strengths and limitations of IPs as a model for inclusive innovation. Findings suggest that while IPs may achieve some short-term success in creating spaces for wider participation in decision-making processes, they may be significantly influenced by forms of power which may not always be visible or easily challenged.

Keywords:

• innovation platform
• power dynamics
• natural resource management
• power cube
• Ethiopia

1. Introduction

In the last 40 years, there has been a shift from linear approaches to technology and knowledge transfer exemplified by the World Bank Training and Visit Program (Anderson, Feder, and Ganguly 2006) to participatory approaches with the Farmer First initiative (Chambers 1994) and recently innovation system approaches. This change is largely due to a growing recognition that agricultural change is not just about adopting new technologies; it entails complex interactions among multiple actors and a range of technical, social and institutional factors. In addition, there is greater awareness that the diffusion of agricultural technologies requires localized adaptation and socio-technical innovation to ensure the appropriateness of technologies in specific contexts. This has led to the development of the agricultural innovation systems (AIS) concept, influenced by ideas from Lundvall (1992) on national systems of innovation and developed by Hall et al. (2001, 2006) within the agricultural research arena (Klerkx, van Mierlo, and Leeuwis 2012, 463).

An innovation system refers to a cluster of individuals and organizations involved in knowledge generation, diffusion and use, together with the processes required to turn knowledge into useful economic or social benefits. Innovation is often defined as ‘new information introduced into and utilised in an economic or social process’ (Spielman 2005, 12), which may include new combinations of existing knowledge. Innovations may be technological, organizational, institutional, managerial, related to service delivery or to policy. Most importantly, this definition implies that knowledge or technology does not become an innovation unless it is used. Advocates of AIS believe that innovation follows a non-linear process, and that the ‘system’ capacity depends on the ‘density and quality of relationships’ between the innovation agents and supporting institutions (Altenburg, Schmitz, and Stamm 2008, 327). Therefore, there is an explicit focus on linkages between actors within the system, and the influence of institutions and infrastructures on these actors’ ability to innovate. The focus on actor linkages implies that improved interaction will result in better information exchange, ideas and opportunities and ultimately lead to innovation. However, there have been questions about the best way to achieve these linkages.

AIS thinking prompted experimentation with different methods for establishing and supporting multi-stakeholder innovation. Innovation platforms (IPs) (also referred to as multi-stakeholder platforms, innovation networks or learning alliances) offer one potential approach. IPs are forums that are designed to bring together stakeholders from different interest groups, disciplines, sectors and organizations to exchange knowledge, ideas and resources and take action to solve common problems in order to bring about a desired change. The combination of these different actors is seen as a potential catalyst for addressing problems within a given system, and a way of ensuring that various groups – including those traditionally marginalized from innovation – can contribute to the change process. As such, IPs are often promoted as a means of addressing power imbalances between farming communities and decision-makers, and they can be regarded as a new model of inclusive innovation (see also Swaans et al. 2014).

IPs may also help address the lack of coordination and communication between stakeholders; seen, alongside power imbalances, as a key bottleneck to the process of fostering change in rural systems. Although there is an extensive literature which discusses innovation systems and IPs from a theoretical standpoint to date, there has been limited analysis of how IPs operate in practice (Nederlof, Wongtschowski, and van der Lee 2011, 11). The issue of power dynamics in particular tends to be overlooked. Considering the aims of IPs, and their current popularity, it is important to understand how power dynamics can potentially influence platform processes and in particular the participation of marginalized groups. This may also provide an insight into power relations within inclusive innovation more broadly.

This paper explores the use of IPs for natural resource management (NRM) in the Ethiopian highlands. Three platforms were established as part of a research-for-development project called the Nile Basin Development Challenge (NBDC), funded by the CGIAR¹ Challenge Program on Water and Food (CPWF). The NBDC project aims to improve the resilience of rural livelihoods in the Ethiopian highlands through a landscape approach to NRM. The project approach is based on the premise that development of integrated strategies, by a range of stakeholders, which consider technologies, policies and institutions, will lead to improved NRM. IPs were aimed to provide a forum for negotiation and dialogue to encourage: joint identification of issues; improved linkages between actors; increased community participation in planning processes; and co-design of interventions tailored to local livelihoods, environmental conditions and the needs of different stakeholders.

Power dynamics between platform members, competing interests, difficulties identifying and aligning incentives for collective action, as well as the wider political and institutional context, presented particular challenges for platform facilitation and had significant repercussions for the IP interventions. Below, these are described in order to highlight why a more in-depth understanding of power dynamics is important for future IP processes. This will provide valuable insights for NRM and for inclusive innovation, in both of which power remains an often neglected topic. Yet in both cases, power shapes institutional arrangements which create differentiated access to and control over resources. This, in turn, can create differentiated outcomes of innovation and development processes.

The paper is organized in the following way: Section 2 reviews work to date on power dynamics within multi-stakeholder platforms and briefly introduces the power cube conceptual framework. Sections 3 and 4 describe the research methods and the Ethiopian context. Section 5, containing the results and discussion, uses the power cube as an organizing framework and is divided into subsections on ‘IPs as spaces for participation’, ‘forms of power within the platform space’ and ‘levels of power’. The concluding Section 6 summarizes the findings and their implications.

2. Conceptual framework

The politics of participation and the role of power dynamics within participatory processes have been extensively written about. Indeed, the twenty-first century has witnessed a growing backlash against ‘participation’, on the basis that it has often failed to achieve meaningful social change, largely due to a failure to address power imbalances and political realities on the ground (Cooke and Kothari 2001). However, these debates, and the subsequent theoretical and methodological advances, do not seem to have been much applied to arena of inclusive innovation generally, nor to the establishment and facilitation of IPs specifically. This may be partially due to the fact that IPs have evolved from industrial and commercial settings where issues of participation and empowerment are not necessarily a concern.

A fundamental assumption behind the IP approach is that the ‘innovation system’ in question is not working effectively and that actors within the system need an initial push or opportunity to engage in joint discussion, action, sharing and learning. However, even if IPs succeed in creating a shared forum, they may fail to address the underlying issues which cause weak actor linkages in the first place. Power dynamics can be identified as a key factor governing the nature of interactions between stakeholders. Agricultural IPs bring less powerful actors, such as farmers, together with more influential actors, such as government officers or traders. In theory, platforms enable the members to articulate their needs and work together to achieve a common goal on equal terms. In reality, the goals, interests and perspectives of actors are likely to diverge and may be even in conflict with one another. Achieving consensus can be difficult and sometimes unlikely. Thus, platforms need to find ways of enabling less powerful actors to influence decision-making; a key tenet of inclusivity.

Various works have touched on the issue of power dynamics within platforms. Steins and Edwards (1999) focus on power dynamics and stakeholder representation within platforms on resource use in complex common-pool resource management scenarios. They highlight the importance of empowering platform participants in order to challenge inequalities and dominant power relations. Faysse (2006) describes a number of challenges involved in establishing multi-stakeholder platforms within unfavourable contexts, and asymmetries of power emerge as a key issue in the selected case studies. Faysse, therefore, argues that platforms provide an imperfect negotiation process. Cornwall (2002) also highlights that spaces for citizen participation are not neutral, but are shaped by power relations; and supporting inclusive participation of citizens in decision-making arenas calls for greater understanding of the micro-politics of participation as situated practice. Edmunds and Wollenberg (2002) examine the impact of multi-stakeholder platforms for disadvantaged groups and argue that they often mask abuses of power and more structural and enduring inequity. These works indicate that dealing with power dynamics is critical if IPs are to be successful mechanisms for inclusive innovation. However, power dynamics are complex and multi-faceted and vary depending on the context, presenting a challenge for facilitators. There is, therefore, a need for analytical frameworks and tools that can be used to identify these dynamics and investigate their influence on platform processes.

This paper uses the power cube, developed by Gaventa (2006), to critically analyse the IPs used in the Nile Basin Development Challenge. The framework – which has been used as a tool for practitioner reflection – was developed as a means for understanding the ways in which power operates, particularly within spaces which aim to increase citizen engagement in policy processes. The power cube enables important questions to be asked such as: Do new mechanisms for engagement, such as IPs, represent real shifts in power? Do they open up spaces where participation and citizen voice can have an influence? Will increased engagement in platforms risk simply re-legitimizing the status quo, or will it contribute to transforming patterns of exclusion and to challenging power relationships (Gaventa 2006, 23)? The framework (Figure 1) outlines three dimensions of power:

• spaces which refers to the potential arenas for participation and action including closed, invited and claimed spaces;

• forms which refers to the ways in which power manifests itself including its visible, hidden and invisible forms and

• levels which refers to the differing layers of decision-making and authority, including the local, national and global.

Figure 1. The ‘power cube’: the levels, spaces and forms of power (Gaventa 2006).

The framework highlights that these dimensions are not fixed or static and are constantly interrelating with each other. The three dimensions are used to organize the paper and are explained in greater detail in Section 5.

3. Research methods

The NBDC project formed district-level IPs in three study sites in the Blue Nile Basin in Ethiopia: Fogera district in Amhara Regional State, and Diga and Jeldu districts in Oromia Regional State. The platforms began in 2011 and are ongoing at the time of writing. We synthesize lessons from the initial phase of platform operationalization. Our analysis is based on the data collected using a mixed method approach.

Prior to IP establishment, 45 farmer focus group discussions were conducted to gather information about NRM planning and implementation processes: three focus groups – male, female and mixed sex – were organized in five research kebeles² per woreda.³ Fifteen to 20 farmers were involved per focus group and included individuals from wealthy, poor and landless households. Kebeles were selected to represent different elevations and cropping systems within the woredas. Researchers also undertook community mapping exercises, transect walks and historical timelines in each kebele and held in-depth interviews with extension agents, district and regional agricultural experts, government administrators and national researchers.

After platforms had been established, participatory community engagement exercises were carried out in three kebeles in each woreda to identify and prioritize community NRM issues. A total of 48 farmers were involved in this process per woreda, 16 farmers per kebele (eight males and eight females). Approximately 10 official IP meetings were held in each site between 2011 and 2013 involving 20–30 people per meeting (platform membership is discussed in more detail in Section 5.2.1). Minutes were recorded for each meeting and stakeholder feedback was documented by NBDC researchers using a combination of semi-structured questionnaires, comment cards and group discussions. In addition to official minutes, researchers recorded their observations of the platform process and pilot interventions.

From 2012 to 2013, each platform undertook pilot interventions in selected kebeles within the study woredas. The interventions were evaluated by platform members and researchers at the end of each season and reports produced. An independent mid-term review of the platforms was conducted between October and December 2012, which involved a review of secondary data, focus group discussions and key informant interviews with platform members, members of the IP technical groups, participant and non-participant farmers.

The authors of this paper were all involved as researchers in the NBDC project and several of the authors were involved in the establishment and facilitation of the IPs. Researcher reflections were critical for documenting platform activities, as well as capturing unspoken or tacit knowledge of the process and the often complex dynamics between stakeholders. It should be highlighted that due to the sensitive political context much of our evidence was generated through informal conversations as much as formal interviews.

4. Ethiopian context

It is important to outline the context within which the platforms are operating. Ethiopia is often cited as an example of severe natural resource degradation. Various land and water management programmes have been implemented on farms and community lands over the past four decades, undertaken by government agencies in collaboration with national and international organizations. However, these have mostly been top-down in nature and failed to take into account the needs, aspirations, constraints and livelihood realities of farming communities (Hoben 1996; Pankhurst 2001; Merrey & Gebreselassie 2011; Ludi et al. 2013). More bottom-up approaches are deemed essential if NRM interventions are to be owned by farmers, be sustainable, and make a meaningful contribution to improved environmental management and livelihoods.

Recently, attempts have been made to incorporate participatory approaches into NRM planning and implementation, with some successes. However, government in Ethiopia is extremely strong and maintains tight control of local decisions, despite formal decentralization; a situation which has persisted since the Derg regime of the 1970s and 1980s. Hagmann and Abbink observe that, ‘Although formally a federation, Ethiopia's central power holders keep a tight leash on sub-national entities’, and that ‘despite its participatory rhetoric … development is state-centred and state-driven’ (2011, 584–585). As a result, there is a contradiction between national plans and output targets which generally take a top-down approach, and attempts to devolve and decentralize planning processes using a participatory approach to the lowest administrative levels. Local government officials have very little experience of stakeholder consultation, locally tailored planning or other cornerstones of innovation. In addition to the political environment, planning and implementation processes are not sufficiently coordinated and institutional structures are often weak with little or no contact between stakeholders. Other challenges include a lack of bottom-up farmer organization, poorly developed markets, poor infrastructure, limited access to information and inadequate extension. As a result, the overall success of NRM interventions to date has been limited.

In recognition of these challenges, IPs were initiated to stimulate new approaches, with particular emphasis on changing the relationships and interactions between stakeholders. However, the Ethiopian institutional and political context proved to be a challenging environment for this model of inclusive innovation. There has been much debate about whether and to what extent external agents can bring about change in such environments through the use of multi-stakeholder processes. Faysse (2006, 220) notes that the nature of the state significantly affects the prospects of successful multi-stakeholder processes; a state which is ‘either too strong or too weak to support a multi-stakeholder platform process and decisions’ creates an unfavourable environment for such approaches. In order to monitor the platforms’ effectiveness, it was critical to understand the nature of stakeholder interactions; the impact of these dynamics on platform activities and to reflect on the extent to which platforms were able to shift these dynamics.

5. Results and discussion

The three dimensions of the power cube, as described in Section 2, are used as an organizing framework for our analysis. First, we define IPs as spaces for participation, and examine who created these spaces, who was invited to participate, who controlled the space and the quality of participation within the space. In order to then understand the forms of power operating within the platforms, we analyse who participates and who is excluded, who makes decisions, who implements decisions and who benefits and loses from these actions. Finally, we look at how the local-level platforms relate to the different levels of power in the Ethiopian context and how actors and policies at local, regional and national levels influence the platform process. We also explore whether or not platforms could influence higher political levels. Site-specific details of the ‘spaces’, ‘forms’ and ‘levels’ of power are not always discussed below, as our analysis aggregates what we identified as similar patterns across the three platforms. These similarities may partly be due to the project approach as well as the uniformity of the Ethiopian NRM planning and implementation process.

5.1. IPs as spaces for participation

‘Space’ refers to decision-making arenas and forums for action; in this case, the IPs themselves. Power shapes the boundaries of spaces, who may enter them and what is possible within them. The power cube refers to three kinds of spaces: ‘closed’ spaces which are controlled by a powerful group and difficult or impossible for outsiders to influence; ‘invited’ spaces where policy-makers invite outsiders to contribute their views to decision-making, allowing influence but within boundaries determined by the powerful and ‘claimed’ spaces where the less powerful can ‘develop their agendas and create solidarity without control from power holders’ (Luttrell and Quiroz 2009, 11). Previous studies as well as extensive situational analysis (Merrey and Gebreselassie 2011; Ludi et al. 2013; Snyder et al., 2014) indicate that that local NRM planning and implementation in Ethiopia is a ‘closed’ or at most nominally ‘invited’ space. One goal of the NBDC IPs was to create a genuinely invited space where a range of participants could influence how higher level NRM strategies were implemented in their locality, for example, the location, design and timing of interventions.

Who creates the space is critical to who has power within the space and which agendas are pursued. Therefore, an analysis of power and its expression in IPs cannot ignore the role of NBDC researchers. The IPs were instigated by the NBDC project to work on improved NRM. Thus, the agenda of NRM came from researchers, not from district staff or communities. NBDC staff also held visible power as resource providers, without which the platforms would not function. Researchers undertook a situational analysis in all three study sites before the platforms were established. They identified the predominance of top-down approaches to the implementation and the lack of sectoral integration and community participation as major challenges to successful NRM. NBDC researchers proposed IPs as a potential mechanism for dealing with these issues. During initial platform meetings, researchers explained the IP concept and the NBDC research agenda, and shared the results of the situational analysis and perceived challenges. Despite researcher efforts to ensure that IP members understood the aims and objectives of the platforms, often platform participants had a limited understanding of the concepts driving the platform work. This shortcoming may have been due to the fact that the challenges and proposed solutions were largely identified by researchers who analysed the system from an external perspective.

Variations in understanding meant that different actors within the platforms interpreted the platform aims in different ways and used the space to push their own agendas. For example, due to the synergies between the government soil water conservation agenda and the NRM focus of the NBDC research, government workers saw the platforms as an opportunity for achieving government targets. The need for external support in the area of NRM was frequently mentioned by government actors in early platform meetings and it was clear that local government agents were concerned about meeting targets set by regional- and national-level decision-makers. Their interest in the platforms could, therefore, be seen as an attempt to appropriate the platform space for their own purposes in order to capitalize on external expertise. In addition, researchers from the wider NBDC team also sought to utilize the IPs as vehicles for disseminating their research findings. Thus, while platforms focused on issues that were relevant to certain stakeholders, the platforms also presented an opportunity for dominant actors to achieve their goals.

NBDC researchers undertook the initial platform facilitation, including establishing the time and place of platform meetings, giving them a certain degree of control over the process. In addition to researcher facilitation, IP members were encouraged to establish technical committees for the implementation of pilot interventions. These included a range of IP members but were primarily led by government actors. Concerns about ownership and sustainability led NBDC researchers to hand over facilitation to local actors after a year. The process of devolving facilitation was not straightforward. NBDC researchers decided not to transfer facilitation to dominant local government actors – another example of the influence that researchers had over the platform process – and instead local NGOs were assigned due to their more ‘neutral’ role. However, NGO representatives still had to seek approval from and work alongside local government and in some cases struggled with facilitation capacity. This illustrates the difficulties that ‘insiders’ may face in navigating power dynamics, particularly if they are part of existing power structures. So although the NBDC platforms succeeded in establishing ‘invited’ spaces for decision-making, these spaces still operated within boundaries set by more powerful actors, namely project partners and local government. On reflection, creation of truly inclusive ‘claimed’ spaces for innovation may be unrealistic in a short time period in such a challenging context.

5.2. Forms of power within the platform space

It is important to look not only at the spaces for engagement and how they are constructed and managed, but also at what goes on inside them. The power cube outlines three forms of power: visible, hidden and invisible forms. Visible power refers to the observable aspects of decision-making which usually take place in public. It is possible to identify visible signs of power by looking at who participates and who gains and loses within decision-making arenas. However, often little attention is paid to hidden forms of power which refers to the ways in which powerful actors and institutions maintain their power and privilege (cf. Scott 1985). This can include deciding who participates in decision-making and what gets onto the agenda by excluding certain issues from discussion or controlling politics behind the scenes. Hidden forms of power also make certain voices or issues more important than others; this can be based on factors such as gender, age, ethnicity or expertise. Invisible power goes a step further and refers to the social and political culture which shapes the psychological and ideological boundaries of participation. This form of power serves to keep significant issues and problems away from the decision-making arena by influencing how individuals think about their place in the world; for example, their beliefs, sense of self and acceptance of the status quo (Gaventa 2006, 29). These forms of power can, therefore, be much harder to identify and describe.

5.2.1. Platform membership and representation

Researchers played a significant role in determining initial platform membership. Due to the project aims, they focused on identifying stakeholders whose mandate was to address NRM issues and who had decision-making power over resources. These included: community representatives, district administrators, government experts from the Bureau of Agriculture, extension agents, national agricultural research centres (NARs), local universities and NGOs. Representatives from these stakeholder groups were invited to an inception meeting which consisted of roughly 12 government staff including extension agents, 4 NBDC researchers, 3 community representatives, 2 NARs researchers, 1 university researcher and 1 NGO representative. The numbers and composition fluctuated in subsequent meetings, but the dominance of government representatives persisted throughout, both in terms of numbers and the degree of power that these participants were able to exert on the process. The dominance of local government was also established very visibly in all three sites by the fact that platform meetings were held in woreda offices.

Stakeholder analysis before platform inception indicated public sector dominance of NRM, a lack of civil society actors and the weak role of the private sector. To some extent, this lack of diversity among actors may reflect the NRM focus of the research. However, an investigation of Ethiopian rural innovation systems and networks conducted by Spielman et al. (2010) also found that government extension and administration exert a strong influence over smallholder networks, crowding out market-based and civil society actors thereby limiting innovation processes. The dominance of certain stakeholders is an important consideration for IPs. This has been alluded to by Wennink and Ochola (2011), who state that government actors must be treated like any other platform members. However, this is not always achievable. Although NBDC researchers were aware of the potential power that government representatives could exert on the process, they were unable to take steps to ensure more balanced representation of stakeholders. This was partly due to the lack of diversity in the Ethiopian institutional context, but also due to the fact that government endorsement was a prerequisite for the project to become operational, and this included stakeholder selection.

In the NBDC platforms, the power of local government representatives was particularly visible in the selection of ‘community representatives’. In all three sites, it was decided that a kebele chairman or manager should participate in platform meetings, ideally alongside one or more community representatives. Kebele leaders work for the government at local level and act as intermediaries between the government and the wider community. Farmer representatives were selected according to certain criteria: they should be active, able to represent other farmers and report outcomes of meetings to the wider community. As a result those selected were often ‘model farmers’ who were handpicked by the district administration. Model farmers play a prominent role in the local administrative structure and are expected to persuade their neighbours to support government initiatives and to participate in developmental activities (Williamson 2011). The selection process for model farmers is also subject to local power dynamics. Research has found that such farmers are often selected by local leaders based on relationships and political patronage rather than on their knowledge, farming abilities and willingness to assist others (Lemma and Hoffmann 2005). It is, therefore, likely that both the kebele leaders and the selected ‘farmer representatives’ who were invited to platform meetings were supportive of government agendas and not necessarily representative of the wider community.

The ‘farmer representative’ selection process within the platforms can be seen as an example of false homogenization as outlined by Luttrell and Quiroz (2009), when stakeholders are grouped, and hence represented in ways which do not reflect their true diversity. NBDC researchers found it difficult to challenge the selection of farmer representatives as too much interference risked derailing the process. This meant that in reality, researchers had limited power over platform composition. The issue of who has control over stakeholder representation has important implications for IPs, particularly if they are established to encourage the participation and innovative capacity of the poor. The consequences of who is represented and who is excluded can be demonstrated when looking at the decisions and subsequent interventions undertaken by the NBDC platforms.

5.2.2. Decision-making

The starting point for each platform was the identification of a commonly agreed upon NRM issue. In Diga, the issue of land degradation was selected; in Fogera, unrestricted grazing was chosen and in Jeldu, soil erosion. Although these issues were evidently important for platform members, NBDC researchers suspected that the choice of these entry points was significantly influenced by a large-scale government Sustainable Land Management (SLM) campaign that was taking place throughout Ethiopia at the time. Due to concerns about dominant government interests and a lack of community representation in the prioritization of key issues, a series of ‘community engagement’ exercises were conducted in selected kebeles in each district. These exercises revealed significant differences in the prioritization of NRM issues between farmers and decision-makers. Farmer-selected issues, although linked to land degradation, were more focused on immediate livelihood concerns. They included termite infestation, crop disease, water shortage and lack of livestock feed. In contrast, issues selected by decision-makers reflected longer term concerns with landscape-wide soil and water management. While such discrepancies are to be expected, this serves to highlight the ways in which hidden forms of power can be exerted over multi-stakeholder negotiations in order to achieve the objectives of dominant actors. In this case, power was expressed through the influence of government sensitization programmes on the prioritization of key issues.

Conflicting perspectives came to a head when the results of community engagement exercises were presented to the platforms. In Diga, platform members prioritized land degradation rather than the issue of termite infestation identified by farmers from the wider community. In Fogera, although unrestricted grazing was identified as a common issue by members of the platform and community members, there were very different views about how this issue should be addressed. Government representatives wanted to take immediate action by confining livestock to homesteads, whereas farmers thought it should be tackled over a longer time frame through a range of interventions, partly due to concerns about producing sufficient fodder. In Jeldu, there was apparent consensus between members of the community and the platform members who both prioritized soil erosion, but this may have been because NBDC community engagement exercises coincided with the government SLM campaign, which entailed ‘awareness raising’ activities in every kebele of participating districts.

The selection of NRM issues was a critical point in the IP process as these issues were used to identify entry points for interventions. Although there was a lack of consensus regarding the issues, platform members in all three sites decided to focus on fodder interventions as a way of meeting both the short-term needs of farmers and the longer term goals of decision-makers. This was largely due to researchers from the International Livestock Research Institute (ILRI) who played a key mediating role. In Diga, researchers identified a termite-resistant fodder grass that could provide livestock feed and rehabilitate degraded land. In Fogera, improved forage production was seen as a way of enabling farmers to meet their livestock's fodder needs and control grazing. In Jeldu, fodder interventions were seen as a way to increase livestock feed and improve soil and water conservation. While this highlights the contribution that external facilitators can make if they have an overview of the context and are able to highlight synergies between apparently conflicting perspectives, the fact that fodder interventions were pursued in all three sites cannot be ignored and illustrates another sphere of influence over the process.

Much has been written about the influence that donor organizations unwittingly have over ‘participatory’ selection processes. Described as the ‘development effect’, participants frame their needs in terms of what they know or assume that the implementing agency will be able to deliver, thereby securing ‘known benefits’ (Mosse 2001). Critiques of participation also highlight that researchers and development practitioners often use participatory approaches to achieve their own agendas or those of their organizations. In this case, it seems that the promotion of fodder interventions was not a conscious ploy by ILRI researchers but rather researchers drawing on their own knowledge and experience, which was limited to a certain area of expertise. Fodder interventions were in many ways a compromise between farmer and decision-maker interests, so to an extent the interventions met the demands of both groups. However, it is possible that it was easier for platform participants to mirror the discourse and preferred solutions of the ‘experts’, in this case, the researchers facilitating the platforms than making more complex and differentiated statements of preferences (Cornwall 2004, 84).

5.2.3. Implementation

The next step was to design innovative interventions. Due to the need to meet certain project objectives, NBDC researchers again influenced this process by stressing that piloting should permit exploration of the factors influencing adoption and effectiveness of the innovations. Three different approaches were, therefore, applied in each site: backyard fodder development by individuals at household level; planting of fodder on soil and water conservation structures and area enclosure of communal grazing areas. Improved forages were then chosen by experts to suit local agro-ecologies. After the innovations had been designed, farmers were selected by the woreda (district) Livestock Development Agency in collaboration with extension agents in target kebeles and invited to participate. In the first year, 40 farmers participated in Diga, 13 in Fogera and 96 in Jeldu. One of the selection criteria for participating farmers was that they should express an interest in fodder development. However, during the farmer selection process there was limited consideration for the needs of different types of farmers. Variations in livestock holdings, land size and wealth were not taken into account, and farmers without livestock were not included.

One of the most commonly raised issues in institutionalizing participatory innovation processes is how to incorporate the diverse and often competing needs of participants. Deeper engagement and greater farmer participation in the planning and design of the innovations could have helped rectify this deficiency. However, during the first year of the pilot, farmers were largely seen by IP members as ‘implementers’ rather than co-designers. As a result in most sites, participating farmers were not aware that the selection and design of the pilot interventions were part of an IP process. This led to different understandings of the innovation process between decision-makers and farmers. Platforms are designed for deliberation, which means that participants should have sufficient knowledge of the aims and objectives of the platform and the theories at work to engage in a meaningful way (Faysse 2006, 225). In the first year of the pilot, community members were not able to discuss or influence the design of innovations due to their exclusion from the process. It is likely that the platforms and the resulting innovations were seen as yet another top-down government programme by participating farmers. As a result, the IP activities reflected and reinforced existing gaps between community members and decision-makers rather than overcoming them.

In Fogera, this gap became apparent when initial interventions were destroyed shortly after activities began; farmers in the first pilot village uprooted the improved fodder plants they themselves had planted. During subsequent investigation by researchers, a number of farmers expressed suspicions that platform interventions were part of an agenda to take communal land for a government afforestation programme. Such perceptions are widespread in Ethiopia due to a history of insecure tenure and land redistribution initiatives under previous regimes. As a result, farmer resistance to NRM innovations is common and presents a significant barrier to the sustainability of NRM initiatives. Community members had not raised these issues with platform members during implementation, and it is debatable whether they would have revealed such concerns to government administrators. Eventually, the platform was forced to abandon activities and establish a new site elsewhere.

This illustrates that although it can be difficult for less powerful actors to influence platform dynamics, they still have significant power to resist innovations through non-engagement. Scott (1985) describes a number of tactics including foot dragging, dissimulation, false compliance, feigned ignorance and sabotage as ‘weapons of the weak’. Gaventa (2006) refers to such strategies as hidden forms of power, as they often take place outside formal and public decision-making arenas. Such forms of hidden power illustrate that power dynamics are often more complex than they initially appear and seemingly powerless actors can have more agency than is assumed. These forms of resistance also suggest that although the IPs brought farmers into the decision-making space and nominally involved them in planning and implementation processes, the platforms were not necessarily effective at influencing deep-structure power dynamics influencing stakeholder interactions.

5.2.4. Interactions between stakeholders

Even if members of the wider community had been included in the IPs, they would not necessarily have been free to express their points of view. As Hildyard et al. (2001, 69) pointed out, identifying stakeholders and getting them around the table will only create effective dialogue if ‘all the actors involved are deemed to have equal bargaining power’. There was a sense that participating farmers were reluctant to express alternative views in platform meetings because these spaces were ‘infused with existing relations of power’ (Cornwall 2004, 80). Various studies in Ethiopia have observed that opposition or minority views are often not tolerated and can even result in people being denied access to resources (Pausewang, Tronvoll, and Aalen 2002; Aalen and Tronvoll 2009).

In addition to the restrictive political context, widely held negative attitudes towards farmers did not create a conducive environment for farmer participation. In all three platform meetings, decision-makers frequently complained about farmer ignorance of key issues. In early meetings, decision-makers highlighted ‘lack of knowledge’, ‘limited awareness of NRM’, ‘backward or inappropriate farming practices’ and ‘farmers’ short-term vision’ as some of the main challenges to improved NRM. These perceptions have undermined self-confidence among farmers, many of whom seem to have internalized the narratives presented by decision-makers. Farmers participating in platform meetings frequently stated that they ‘lacked awareness’, thereby reinforcing decision-maker perceptions.

This internalization of government narratives can be seen as an example of invisible power whereby people come to understand their situation as unchangeable or inevitable. Such attitudes and patterns of interaction between community members and decision-makers are intimately related to struggles over resources, governance and power and are often firmly ingrained. Vaughan and Tronvoll (2003a, vi) observe that ‘the process of socialization from birth often teaches Ethiopians that people are not equal’ which leads to deeply entrenched power relations that cannot easily be changed. The fact that similar patterns emerged in all three platforms, located in different parts of the country, suggests that these dynamics are widespread throughout Ethiopia. These forms of invisible power are important to identify because they have a great deal of influence over platform processes. One of the main aims of IPs is to facilitate knowledge sharing between actors; however, such attitudes inhibit farmers’ ability to share their perspectives, knowledge and experiences. This situation represents a major institutional barrier to inclusive innovation and is extremely difficult to address with platform members. Platform facilitators came to realize that platforms may have limited impact in terms of challenging entrenched mind-sets.

5.3. Levels of power

Much of this analysis so far has focused on power dynamics at the local level, namely the roles played by NBDC researchers, farmers and local-level decision makers. This focus reflects the research agenda and the fact that interactions between actors at the local level were identified as a critical factor for the success of NRM interventions. However, there is significant debate regarding which level of power is the most important to address. There are some who argue that changing power dynamics must begin locally, while others argue that the nation state is still the main source of power and public authority. The power cube recognizes that what is going on at all levels – local, national and global – is potentially significant, as well as the inter-relationships between people working at these levels. During the NBDC IP process, researchers became increasingly aware of the impact of interactions between different levels of power, namely between local-, regional- and national-level actors.

The hierarchical social and political environment in Ethiopia has been cited by a number of studies as an impediment to participatory approaches. As Vaughan and Tronvoll (2003b, 33) write, ‘the pattern of social interaction in Ethiopia sustains a strictly hierarchical stratification of society, where one is constrained by a large, invisible, but rigid system of common sanctions, to obey “orders from above”’. Despite rhetoric about decentralization, local government officials are subject to top-down power from higher levels of government which can impede local attempts at participatory, inclusive approaches to planning, innovation and implementation. Ayele (2011, 143) writes that ‘local government is not adequately institutionalized to exist as an autonomous level of government … explicit and implicit provisions in the regional constitutions and statutes render local government a subsidiary structure whose function is limited to implementing centrally adopted policies’. This is expressed in the quota-based NRM implementation framework, and the pressure placed on government staff to meet targets. Hoben (1996, iv) has also written that the ‘cumulative effect of top-down attitudes does not support the adoption of a participatory, error-embracing approach to agricultural development’. In this respect, there is often a lack of consideration for the challenges faced by local-level decision-makers in the analysis of power dynamics. Local-level government officials operating within this environment are constrained by higher level decision-making processes at regional level and national level which they have little power to change. They are acutely aware of these dynamics and as such there is a lack of incentive to do things differently; in other words, a lack of incentive for local innovation.

Fiszbein (1997, 3 cited in Ribot 2003, 61) highlights that in some cases, the inability to conduct effective participatory approaches may not be due to a lack of capacity but, rather, poorly designed incentives. Intensive engagement with community members is often difficult to undertake and adds to already strenuous workloads. Local government administrators, therefore, need to see clear advantages of a participatory rather than top-down approach, yet in Ethiopia various structural factors militate against this. Discussions with Jeldu IP members during the first platform meeting revealed that NRM planning is largely top-down because the main priority of local administrators is to fulfil quotas set by national and regional decision-makers. They generally agreed that more bottom-up approaches to NRM planning and implementation could help to better tailor plans to local contexts but they queried whether they would still be able to achieve the ambitious targets outlined by the Government of Ethiopia's Growth and Transformation Plan. Although participatory approaches are outlined in NRM guidelines as part of recent decentralization efforts, in reality they are often not carried out. Authorization and incentives from central government may therefore be required in order to encourage participatory initiatives. Platform members in all three sites questioned why the NBDC project was not focusing attention on communicating these issues to regional- and national-level actors. Certainly, given the importance placed on meeting targets set out in national policy, it is critical for IP facilitators to frame the importance of participatory approaches and of local innovation in reference to these policy goals.

Although local-level IPs can help make these power dynamics visible, they are unlikely to transform them. In this case, it seems that engagement with regional and national actors is required to influence the way things are done at a higher level and on a longer term basis. The NBDC project aims to present evidence generated from the three local case studies to higher level actors through a national-level IP to advocate for change. Other research platforms in Ethiopia have successfully used evidence of local problems and the failure of existing ways of working to strengthen the hand of district officials in negotiating with regional-level authorities (Tucker, Le Borgne, and Lotti 2013). This approach highlights the potential role that interconnected IPs at multiple levels could play in influencing policy. However, political will is required for changes to take place in addition to evidence that policy changes can support the achievement of national objectives.

6. Conclusion

As the NBDC IPs are ongoing, this paper can only document lessons learned from the initial stages of the platform process. Many of the challenges faced by the NBDC platforms are context specific. Nonetheless, lessons can be drawn and perhaps used to inform other IP processes and other forms of inclusive innovation. Experience from the NBDC platforms demonstrates that power dynamics need to be acknowledged and dealt with explicitly. One of the first steps in dealing with power dynamics is to recognize that IPs are not neutral mechanisms; they aim to promote change which can bring benefits for some but may have negative effects on others. Often these changes fundamentally challenge the existing distribution of power and can thus be threatening to vested interests. The benefits that result from such innovation processes are likely to be determined by who initiates them, who participates and who has influence over the process. If such power dynamics are well understood, then perhaps ways can be found to mediate them, manage potential negative consequences and work for positive change.

In order to identify, analyse and challenge power dynamics, certain tools and frameworks are required. The power cube is one possible approach and highlights the multiple dimensions of power that should be considered. This framework has been useful in conducting a retrospective analysis of the power dynamics in the NBDC IPs, but it may also be applied to other models of inclusive innovation. Our findings illustrate that visible forms of power in particular may be easily identified, whereas hidden and invisible forms are more difficult to uncover and by their nature will be difficult to address. In the case of the NBDC project, it seems that IPs are unlikely to have any significant impact if they merely address visible forms of power at the local level without addressing hidden and invisible power dynamics as well as structural blockages at higher levels. So, although IPs may achieve some short-term successes at local level, significant changes to institutional arrangements and incentive mechanisms may be required at higher levels to make longer lasting change tenable at larger scales. As Gaventa (2006) argues, successful change requires thinking not only about individual dimensions of power but also about how each dimension relates to the other. This multidimensionality is a huge challenge and suggests that in order for meaningful change to take place in inclusive innovation systems then multiple, linked strategies at different levels must be designed that incorporate an appropriate time frame and build the capacity of concerned stakeholders.

Even if IPs cannot solve problems of power and representation, they may serve to make these issues more visible, and can potentially play a role in building people's capacities to change these dynamics (Hiemstra, Brouwer, and van Vugt 2012). However, the degree to which platforms are able to address power dynamics depends largely on how they are facilitated. As Swaans et al. (2013) argue (see also Swaans et al. 2014), the success of such platforms is heavily linked to the attitude, skills and capacities of the facilitator or ‘innovation broker’. Identifying and analysing power dynamics and their effects on inclusive innovation demand high levels of reflexivity and awareness. While reflexivity is perhaps possible within a research context, when IPs are mainstreamed as part of development interventions, facilitators may find this challenging.

In summary, those who instigate innovation platform processes need to recognize that altering power dynamics is a complex, difficult and lengthy process. However, if power dynamics are not taken into account in the formulation and facilitation of IPs, there is a danger that platforms give the illusion of increased participation whilst simultaneously replicating and masking existing conditions. It may therefore be necessary to question whether IPs are appropriate mechanisms for achieving inclusive innovation in certain conditions, and whether other models may also need to be considered. This is important to acknowledge because if power imbalances are not adequately addressed such processes may in fact aggravate poverty and environmental decline rather than providing innovative solutions.

Acknowledgements

The authors wish to thank Aberra Adie, Gerba Leta and Elias Damtew for their valuable inputs to this project. We also acknowledge the contribution of Aklilu Amsalu.

Funding

This work was supported by the Challenge Program on Water and Food.

Notes

1. Consultative Group on International Agricultural Research.

2. Administrative unit, equivalent to a neighbourhood.

3. Administrative unit, equivalent to a district.