‘Producing’ institutions of climate change adaptation and food security in north eastern Ethiopia

Highlights

•	Institutions play an important role in adaptation and food security action coordination.
•	State actors could use institutions to contain the resource management.
•	Institutions need to be seen as dynamic and socially produced.
•	Institutional diagnostics needs to capture social processes that produce institutions.

Abstract

The paper presents institutional diagnostics, which is sensitive to dynamic social and political processes ‘producing’ institutions underlying practices in resource management, climate change adaptation, and food security. The paper is based on a qualitative case study on watershed development interventions conducted in two villages in Amhara Region, Ethiopia. The research showed that resource management, adaptation, and food security institutions in Ethiopia are a result of struggles between containment strategies of the Ethiopian state and counter containment strategies of local communities. While the state’s containment institutions allowed it to mobilize a large number of rural residents for its resource management interventions, the counter containment strategies from local communities limited the potential contribution of the interventions for adaptation and food security endeavors of the state. From an institutional diagnostic perspective two conclusions are made, one empirical and another theoretical. The empirical part of the paper concludes that the Ethiopian state is using institutions to contain its population towards state-driven development pathways, which is essential to understand watershed development and state-led natural resource management interventions. The theoretical portion concludes that although institutions are often portrayed as static elements of social life, in fact they are also dynamic, socially produced, and could be coopted by powerful actors.

Keywords:

• Institutions
• Diagnostics
• Food security
• Adaptation
• Resource management
• Power relations

1 Introduction

Climate change has put pressure on society through its negative impact on food systems. The fifth assessment report of the Intergovernmental Panel on Climate Change (IPCC) indicates that changes in global temperature and precipitation regimes will affect global food security, impacting all aspects of food security, namely food access, utilization, and price stability (Porter et al., 2014). Hence, climate change adaptation interventions are crucial in areas where food security is threatened either through direct impacts of climate change, such as an increase in temperature and change in precipitation on crops, livestock and fishery systems, or indirect impacts such as flooding on food system infrastructures (Lobell et al., 2008; Porter et al., 2014). It is important to note that in sub-Saharan countries, where livelihoods are dependent on agriculture, climate change adaptation is intimately linked to management of the natural resource basis of rural livelihoods (Niang et al., 2014; Branca et al., 2013; Toulmin, 1991). It is within this intricate relationship of climate change, natural resource management, adaptation, and food security that this paper sets out to diagnose the role of institutions in food security in sub-Sharan Africa in general and in northeastern Ethiopia in particular. This paper attempts to understand how institutions influence food security by diagnosing the way natural resources, which have a bearing on food security are governed.

Table

Description		Woyniye Village	Laste Gerado Village
Population	Total	11183	4029
	Household	2286	982
Farming system		Mixed farming	Mixed farming
Village size		2866 ha	2835
Agro-ecology		Moderate warm	Warm
Slop		Mostly steep slop	Mostly flat
Livestock (in TLU)		4929	6453

The role of institutions in natural resource management, climate change adaptation, and food security is well recognized. They are considered important mechanisms of ensuring collective action in the face of uncertainty (Agrawal, 2010; Mehta et al., 1999; Toulmin, 1991). Taking the common notion of institutions as ‘enablers’ or ‘constraints’, some consider institutions in functional terms such as rules, regulations, and conventions (Ostrom and Basurto, 2011). This often leads to a prescription of institutional design principles, which are assumed to overcome institutional path dependency and inefficiency and allow actors to work together for a common good (Huntjens et al., 2012; Anderies et al., 2004). While such understanding reveals the importance of institutions for food security, it also conceals a lot about it. One important aspect that is concealed is the power dynamics, which influence and are influenced by institutions (Johnson, 2004). In real life situation, power asymmetries explain how actors make use of institutions differently, with powerful actors taking most of the advantage of the production and use of institutions (Jessop, 2005, 2001).

This paper argues for understanding institutions as a product of social and political practices (Mehta et al., 1999). It presents an institutional diagnostic with a special focus on how institutions are produced, selectively used, and resisted in a state-led natural resource management intervention. A case study of a watershed development intervention implemented by the Government of Ethiopia in two villages shows how existing institutions regulate state-society interactions in natural resource management. The paper’s major aim is to make a theoretical contribution to institutional diagnostics, which is committed to understanding institutions within the context of asymmetric power relations between state actors and local communities and the implication of this for climate change adaptation and food security.

Institutional diagnosis for food security and climate change adaptation involves identification of various institutional varieties and their functioning in different settings, with the aim of probing opportunities for transformational change (Röling et al., 2012). It involves understanding the critical constraints of the system in a particular setting and experimenting with selected solutions to change the system to a desired state (Jiggins, 2012; Rodrik, 2010). This paper provides an insight into a state-centric institutional variety and explores what this variety is cable of doing and its predicaments in the social and political context of the study areas. Rather than looking at institutions as local rule making toolboxes of fixing free riding problems in collective actions or multi-stakeholder platforms (Faysse, 2006; Warner, 2005), this paper centers on how institutions are used by those who want to contain the action of others. In addition, it explains the different mechanisms that those under control display to resist containment strategies imposed on them (Jessop, 2005). The diagnostic aspects of this paper are limited to institutions that stem from the state. Thus, although this paper presents the role of power asymmetry in influencing the dynamics of institutions, it does not conclude that other forms of institutions that do not stem from the state are immune from power grabs.

This paper makes use of a variant of institutionalism called Strategic Relations Approach (SRA), which offers heuristic tools for state centric institutional diagnostic. SRA strives to understand the duality of agency and structures from a relational perspective (Jessop, 2005). It takes institutions as relational mechanisms that explain the relationship between structure and agency. Such a relational view of structure and agency unveils the role of power and power relations in the functioning of institutions. In SRA, structures are seen as strategically selective of action of agents. This means structures favor some actions while discriminating against others; while they favor the action of those with power, they discriminate the powerless. However, agents are not passive victims of structures. Rather, they make strategic calculations to make use of structures for their benefits. Hence, individual actors approach structural constraints in a way that maximizes their satisfaction (Jessop, 2005, 2001).

Section two of the paper presents the theoretical framework of the study which is grounded in Strategic Relation Approach (SRA), which is complemented by other theories to illuminate containment and counter containment strategies of the state and local communities respectively. Section three presents methodological approaches of the study including sampling procedures, data collection and analysis. Section four presents the findings of the study, including the vulnerability context of the study villages, diagnostics of institutions of containment of the state and counter containment responses of local communities. Section five distils the implications of the study for institutional diagnostics. The paper ends with a conclusion, which calls for a combination of ‘politicized’ approaches to institutional diagnostics with other forms institutional analysis for a compressive understanding of institutions of climate change adaptation and food security conditions in sub-Saharan Africa.

2 Theoretical framework

The theoretical point of departure for the study is that resource management institutions are dynamic and ‘produced’ through social and political practice of actors (Mehta et al., 1999). It develops an approach, which is sensitive to political, cultural and historical contexts that produce institutions (Clement, 2013; Johnson, 2004). The processes of ‘production of institutions’ involve creating and enforcing a particular institutional arrangement, which may range from a bricolage of formal as well as informal, to mutually overlapping and at times contradicting institutions (Cleaver, 2012). Actors involved in production of a new institutional arrangement are often not on equal footing. Rather, collaboration involves actors with asymmetric power relations, such as the state and local communities (Adger, 2003). In such situations, there is a tendency for the actors to compete for dominance over the collaborative process, resulting in an attempt by the dominant group to restrain the actions of the less powerful group in order to take advantage of the collaboration process (Walker and Hurley, 2004). In so doing, dominant groups in collaborative processes use their power and discursive instruments to make sure that their pre-planned interventions are not hampered by competing actors (Few, 2001). However, those with less power are not simply passive victims. Even in coercive state-society relationships, less powerful groups have some ability to counter the containment imposed on them (Mann, 2009). This interplay between actors of differential power relation influences institutional arrangements of resource management.

As a heuristic institutional diagnostic tool, SRA helps to uncover the following aspects of institutions of food security and natural resource management. First, it treats institutions as ‘emergent phenomena whose reproduction is incomplete, provisional, and unstable’ (Jessop, 2001:1230). Hence, it opens an opportunity to analyze the processes that produces a particular institutional arrangement in a particular time and place. Second, it looks beyond a particular institutional arrangement to understand the social relations and contexts that produce institutions (Jessop, 2001:1231). This insight is useful to understand the political context behind institutions of resource management and food security. Third, it highlights two inherent nature of institutions. A) They are strategically selective in their constraints as they privilege some at the expense of others. b) Actors make strategic calculations to take advantage of institutional arrangements (Jessop, 2001:1224). These insights are useful to uncover the role of power and power relations in resource management institutional settings. The paper incorporates insights from political ecology into the ways state-society relations are used to understand the political context in Ethiopia that produces institutions of resource management, which is in line with historical and sociological approaches to understanding of institutions (Clement, 2013; Johnson, 2004).

Hence, SRA conceptualizes institutions of collaborative resource management as instruments of state containment strategies. Gramsci’s concept of hegemony and Foucault’s concept of governmentality add to this perspective an approach explaining how the state could achieve this. Literature on the Gramscian approach to state-nature relations explicate how the state uses its hegemony manifested principally through ‘ideology’ and ‘common sense’ and selective use of coercion and consent to a state-dominated resource management plans (Bridge, 2014:4; Mann, 2009). Ideologies of the powerful, which are a result of cemented experiences and experimentations of a narrow group of actors, are portrayed as applying to a wider range of actors. In this way, dominant groups obtain consent and legitimacy for their intervention even when they work against the interests of subordinate groups (Lears, 1985). Such hegemonic processes involve wide array of actors including: business people, teachers, religious leaders, and others who would work to their own subjugation, albeit unconsciously (Lears, 1985:572; Bates, 1975). The implication of this for collaborative resource management is that once the state establishes legitimacy of its political ideologies and gets the consent of its citizens, it gets the upper hand to decide on the goals, standards, and norms of the collaborative intervention.

Foucault’s concept of ‘governmentality’ is used to excavate how state’s ideologies penetrate collaborative resource management arrangements by creating governable spaces using power and knowledge (Bridge, 2014). Knowledge production, such as collection of regular statistical data and maps, rule setting and reorganization of subjects are used to govern individuals and collectives whereby power is used to create discursive rationalities (Rose-Redwood, 2006). Rule setting in resource management would involve mobilization of existing institutions such as constitutional provisions, resource use policies, regulations, and the creation of new institutions such as local by-laws. This way, state actors create mechanisms, which allow them to govern collaborative institutions from a distance (Agrawal, 2005).

The paper added a third dimension on SRA to better understand the response of local communities against state containment strategies. Those with less power are not simply passive victims of state containment strategies as they have some ability to counter the containment imposed on them (Mann, 2009). Counter-containment strategies of the powerless could involve subtle resistance, disruptive action, and/or open opposition to containment strategies (Few, 2001; Scott, 1985). This has important implications for collaborative resource management institutions as their effectiveness depends on the cumulative effect of containment strategies of the state and counter-containment strategies of local communities. Hence, what the collaborative resource management institutions are capable of doing is in part determined by the social and political practices that produce these institutions.

In summary the theoretical framework used in this paper for institutional diagnostics is firmly grounded in SRA, which looks at institution as emergent, embedded in social context and malleable to influence of power asymmetry. On top of insights of SRA, the paper used Gramscian notion of hegemony and Foucauldian notions of governmentality in order to detect actions/practices of the state that have the power to produce institutions of containment in local settings. The theoretical framework added one other dimension, counter-containment, to understand responses by local communities to the state containment strategies. The paper uses these theoretical insights to diagnose the institutional arrangements of watershed development in two villages of Northern Ethiopia.

3 Material and methods

3.1 Research areas

The state-led intervention selected for this paper is the Ethiopian Integrated Natural Resource Management Program. It is a nationwide program of rehabilitating degraded watersheds with soil and water conservation measures, enclosing degraded hillsides, afforestation, and water harvesting. As we will see in the discussions that follow, things were not as integrated as the name of the program imply. As a result, the focus of this study was on the main component of the program, the annual watershed development campaign that includes construction of soil and water conservation structures and hillside enclosures.

The study was conducted in Guba Lafto District, North Wollo Administrative Zone, Amhara Region, Ethiopia. The district was chosen because of its high exposure to climate related hazards as well as long years of experience in state-led natural resource management interventions. The district lies between 11⁰ 36′ and 11⁰ 58′ North latitude and 39⁰ 12′ to 39⁰ 50′ East longitude (Gesese and Mberengwa, 2012). According to data obtained from the district agriculture office, the total area of the district is about 130,000 ha, out of which 34.1% is cultivable land, 17.9% is grazing land, 27.1% is forest and bushland, 9.9% makes up human settlement, 6% is wetland and water bodies and 5% is degraded land. The 2007 population census shows that the district has a total population of 139,825, out of which 70750 are male and 69075 are female. More than 95% of the population live in rural areas (CSA, 2007). The district has three agro-ecological zones, 17% lowland (Kolla), 46% mid-altitude (Weynadega), and 37% highland (Dega). In terms of rainfall, the highland areas receive an average annual rainfall of 2500 mm, the midlands get 1500–2500 mm and the lowlands receive 500–1500 mm average annual rainfall (Gesese and Mberengwa, 2012).

From the district, two villages were purposely selected for a case study (see table below). The choice was made to capture diversity in performance of the case study villages in terms of area coverage and quality of soil and water conservation works. Data obtained from the district showed that the first study village, Woyniye, is one of the 34 villages with better performance of watershed development intervention. On the other hand, the second study village, Laste Gerado, is among the villages with the lowest performance. Despite experiencing similar political setting and institutional arrangement, the two villages performed differently.

(Source: Village Farmer Training Centers in the two villages)

The field work for this study was done in two rounds, in June-July 2013 and January-March 2014. The researcher went to the field with the initial theoretical perspective, trying to understand the platforms created for watershed development and management and their challenges. From this initial abstraction the analysis went on identifying the structures and generating mechanisms of the watershed management interventions through iterative movement between field observation of concrete experience and theoretical reflections. The field work involved a study of watershed intervention in two villages in Guba Lafto District, North Wollo Adminstrative Zone, and Amhara Region, Ethiopia.

3.2 Sampling procedure, data collection and analysis

The study made use of a purposive sampling procedure. The selection process involved multiple stages. For the first village, the villagers were first divided into two groups: those who live at the upper catchment of the village and those who live at the lower catchment of the village. These two groups of villagers often face different sets of climate and other livelihood risks because of the striking elevation differences between them. Then, the village development teams were used to narrow down the sampling population. Development teams are government organized teams that comprise 20–30 neighboring farmers. Accordingly, the 65 development teams in the village were divided between upper catchment and lower catchment. Three development teams from each catchment were selected randomly. From each development team two respondents were selected. Accordingly, six of the respondents were from the upper catchment and six were from the lower catchment. Out of these respondents, seven were political party members and five were non-party members. Political party membership was taken as a sampling criterion because of the political nature of the case study intervention and the explicit use of party membership by the government for the purposes of mobilizing local communities.

For the second study village, sampling began by stratifying the village into the three sub-villages. Development teams were chosen from each sub-village. Party membership and gender were also used as selection criteria. Accordingly, there were 12 respondents selected for individual interview, four from each neighborhood. Out of this 12, five were party members and seven were non-party members. Two of the respondents were women.

The research used both intensive and extensive data collection methods (Yeung, 1997; Pratt, 1995). The intensive method involved five key informant interviews with experts from regional, zonal and district level, one focus group discussion with experts of Guba Lafto District, 24 individual interviews with farmers (12 from each study village). The extensive method involved review of over 16 documents from regional, zonal and district agricultural offices.

The data analysis involved on-the-field pre-coding and identification of structures and causal mechanisms and theoretical reflection as well as back-to-office analyses. After finishing the field work, the researcher transcribed the data using f4 transcription software and uploaded it on MAXQDA version 11 qualitative data analysis software. The coding to identify structures and generative mechanisms was done in light of boarder natural resource management and state-society-nature relationship. The initial codes were repeatedly refined to develop a meta-analysis of field observations into a broader explanatory argument involving society, nature and the state in natural resource management interventions in the current Ethiopian context. As is the case in most qualitative studies, the analysis developed into more concrete arguments as the writing process continues (Miles and Huberman, 1994). The study involves iteration between abstraction and concrete experiences. Institutions at work were discerned only after initial conceptualization of state-society-nature relations, which had to be tested with field observation and reconceptualization.

4 Results

This section presents the findings of the study vis-à-vis the theoretical framework set in section two. The section starts with laying the basis for climate change adaptation and food security situations in the study villages by providing a brief account of multiple sources of vulnerability. This is followed by a sub-section which illustrate the watershed development intervention in Ethiopia under a state led ‘Integrated Natural Resource Intervention.’ The sub-section ends by showing disparities in the performance of the two case study villages both in area coverage and quality of soil and water conservation interventions and asking what explains such disparity. Sub- section three and four presents the actions/practices of the state that ‘produces’ institutions of containment. The final sub-section presents the counter containment response of local communities.

4.1 Vulnerability as a starting point for institutional diagnostics

The core livelihood risks facing local communities in the study areas are food insecurity and poverty, called here livelihood risks. These are referred as “core risks” because they are a result of other intermediary risks such as moisture stress, drought, degradation, and shrinking income. Identification of these core risks is important as the ultimate outcome of any successful risk management intervention should be improving food security and livelihood conditions of local communities. The sources of livelihood risks are multiple. Since experts and local communities attribute overlapping and conflicting risk framing to an area, one notes that the material as well as discursive dimensions of risk are important in adaptation decision (O'Brien and Wolf, 2010). The multiplicity of livelihood risks also implies that livelihood hazards are produced through the interaction of multiple sources of risks (Marino and Ribot, 2012). This paper considers the above points while assessing the vulnerability of people in the study areas to climate change impacts and food insecurity and understand the role of institutions set by the government to address climate change impacts and food security.

Accordingly, in the first study village both experts and local communities identified that the short as well as long rainy seasons have been problematic for crop production for a while. The short rainy season has been disappearing all together, forcing farmers to abandon production during this season. The long rainy season, on the other hand, continued to withdraw earlier than it should affecting crops during flowering and seed setting stages. On top of these climate related risks, the village had a high population density with the average farmland per household at only 0.25 ha. As a result, many of the rural youth in the village were landless. Another challenge was although around a quarter of the village households had access to irrigation, they faced fluctuating prices for their produce. On top of these risk settings, one could add crop failures due to inappropriate agricultural technologies introduced by the government. Additionally, the political leadership and the village experts failed to mobilize farmers to take up improved technologies that prevent soil degradation, improve soil moisture, and increase crop productivity and production.

Study village two shared the climate related risks with study village one. The area is closer to the lowlands; hence, it suffers from moisture stress for most part of the year. Apart from climate related risks, the village had high number of livestock with free grazing production system, which made it practically impossible to plant short-rain season crops even when it rains. The village used to benefit from remittances sent by youth who travelled to Arab countries. However, a crackdown on illegal migrants in Saudi Arabia made many youth return home, causing families to lose critical livelihood sources. Accordingly, the livelihood of the people in this village was “at risk” because of multiple risk settings from different sources within and outside the village.

4.2 Watershed development interventions in Ethiopia

The Ethiopian Integrated Natural Resource Management Program is a multi-year government-led program, which is currently under implementation in all regions. The study focuses on this national program, also called watershed development program, undertaken during the Growth and Transformation Plan (GTP) period (2010/11-2014/15). The core interventions of the program include utilization of various soil and water conservation methods such as physical soil conservation structures, biological conservation structures and area enclosure of mountainous and degraded areas. These ‘hard’ interventions are supplemented with ‘soft’ interventions such as organization of local communities into user associations and development and control of resource use by-laws.

Accordingly, looking at the aggregate coverage of the work that has been done nationally between 2010 and 2014, around 16 million ha of land was treated with different soil and water conservation structures in 19,807 micro watersheds during the four years of the GTP (MoA, 2014). At regional level, data obtained from the Amhara Regional Agriculture Office indicated that the integrated natural resource program constructed physical soil and water conservation structures on terraces on 3.83 million ha on farm plots and communal lands; 1500, 252 ha of enclosure on degraded hillsides; and 50468 ha of gully treatment between the years 2010 and 2013. In the study district, Guba Lafto, 154 micro watersheds were treated with 30,840 ha of farm terraces, close to 5 million moisture conservation structures as well as over 1.6 million m³ of trenches were constructed, and over 13 million seedlings were provided to farmers. A feedback report from the district showed that in 2013/14, a total of over 1.9 million residents were organized for the work and 91% (over 1.6 million) turned out for the work.

It was challenging to estimate the work done during the GTP period at the study villages since comprehensive data was not available. The performance estimates for the five-year GTP period were calculated by extrapolating the available data; hence, it is only indicative and cannot give a full picture. According to data obtained from village Farmers’ Training Centre of the first study village, 2076 male and 1414 female members of the village were organized for watershed development work. An eight-day report by the district during the 2013/14 watershed campaign showed that 96.28% of the labor force was organized for the campaign, giving it a rank of 8th among the 34 villages in the district. In terms of work performance, 350 km of hillside terrace, 1095 km of farm terraces, 6347m³ of trenches and 41,110 moisture harvesting and water conservation structures were constructed in this study village over the five-year GTP period.

In study village two, according to data obtained from the village Farmers Training Center, 811 men and 600 women members of the village were organized for the watershed development work. An eight-day report by the district showed that only 48% of the labor force was organized for the campaign, giving it the last rank out of the 34 villages in the district. Most of the work done was on farm terraces. Accordingly, over the five years period an estimated 350 km of farm terraces, 28 ha of gully treatment, 3,500m³ of trenches were constructed in two micro-watersheds in the village.

The above figures are not meant to accurately measure the achievement of the national intervention, rather they are meant to indicate the huge level of coverage of the interventions both at the study areas and nationally. The achievements gave the Ethiopian Government positive credit from the international community, as some claim that due to these interventions Ethiopia is ‘food secure and greener’ than 140 years ago (see for example Reij, 2015). It is also important to note that, although the aggregate figures were indicative, there was a significant difference in performance between the first and second study villages, as the first study village performed better both in development and management of its micro watersheds compared to the second study village. The sections that follow explain the reason behind the large coverage of the watershed intervention and the difference in performance of the two villages.

4.3 Institutions as a state hegemonic instrument

This paper uses the Gramscian notion of ‘hegemony’ to illustrate the Ethiopian state’s use of the ‘developmental state ideology.’ The state uses develomentalism¹ to install its ideology as common sense among its citizens and direct their everyday life towards the realization of state-led initiatives. Hegemony here refers to the way powerful forces use their ideology to invade the thought processes and everyday practices of their subjects by elevating the status of their ideology as a commonsense that everyone should buy in (Bates, 1975). Accordingly, the ruling party, the Ethiopian People’s Revolutionary Democratic Front (EPRDF), has been using its developmental state ideology to create a social consensus among citizens to establish its political ideology as a hegemonic and ‘natural’ pathway to escape poverty and walk on the path of prosperity (de Waal, 2013).

State-led developmentalism as a hegemonic ideology surfaced after EPRDF’s internal party split in 2001 and grew stronger during the threat of defeat in the 2005 elections. The party adopted liberal ideologies, emphasizing the central importance of the economy for the existence of the front and the nation at large (Bach, 2011). The front openly started showing its alliance with the developmental state ideology, using ‘modernization’ through state intervention as the main political ideology to attack its oppositions and legitimize subsequent policy directions (Gebresenbet, 2015; Bach, 2011; Vaughan, 2011; Lefort, 2012). The front uses poverty as the ultimate enemy of Ethiopia calling for an aggressive state economic intervention. It depicts apocalyptic consequences such as famine and national annihilation for failure to embrace developmentalism. It uses militaristic terminology such as “war against poverty”, “developmental army”, “development patriotism”, and “developmental hero/heroine” to describe the sort of hegemonic consensus that the party wants to see develop among party members and the public at large (Gebresenbet, 2015:70).

In the context of smallholder farmers, state developmentalism meant enlisting rural communities for selected priority development initiatives set by the government. The two notable recent five-year plans, the Plan for Accelerated and Sustainable Development to End Poverty (PASDEP) (MoARD, 2006) and the Growth and Transformation Plan (GTP) (MoFED, 2010), envisaged a large scale mobilization for national development. GTP was the most acclaimed and ambitious policy framework, which materialized in political endeavors by the state pressing citizens to fully embrace the state’s policy and work towards achieving the nationally set targets (de Waal, 2013).

One of the main strategies that the party uses to promote state developmentalism is to use its party members to promote its ideology. Party members are the vanguards of the developmentalist ideology of their party (Lefort, 2012). Party loyalty, especially in rural areas, is demonstrated by taking up agricultural practices promoted by the state to set examples for others. The party also expects its members not to question development recommendations from higher level political organization of the party and convince their neighbors to do the same (Lefort, 2012). Hence, the party expects all party members and by extension all citizens to take up development interventions by the state without question and work on how to achieve it. What this means is that the capacity and legitimacy of party members determines the level of penetration of the state ideology into the everyday practice of its citizens.

When looking at the study villages, village one had 550 party members and village two had 300 party members. These party members were also leaders of other social organizations in their village. In study village one, the party members that the researcher interacted with and observed in different party related meetings showed their strong allegiance to the party’s ideology and worked towards promoting it. The party members attended meetings and showed up to the watershed development work on a regular basis. Members were also rated as working well in adopting improved technology promoted by the village extension agents. For example, in one party members’ meeting in the village, it was reported that of all the farmers who used government-promoted technologies for that production season, 29% of fertilizer users, 34% of those who planted in row in general, 71% of those who planted wheat in row and 89% of those who planted teff (an indigenous staple food) in row were party members.

During the watershed development campaign, the party-members leading development teams at village level were responsible for mobilizing other farmers for the work. Those at the leadership level oversaw and actively participated in the watershed work. In all the days that the researcher was in the field during the watershed campaign in study village one, either the chairman of the village or his deputies were always present for the campaign work. It was also a regular practice that higher-level leaders within the party formed sub-committees, monitored quality of work and ranked the performance of their subordinates on a daily basis. Respondents who are party members said that especially after the 2005 election, party members were strictly advised not to use coercion as a means of mobilizing their community for development. In one observation of a village party meeting, the chair of the meeting who was from the district administration office pleaded with the attendants as follows:

“You are leaders and you should take responsibility not only for yourself but for others as well. You are called lead farmers and you should be concerned about others as well, not just yourself” (Field observation at Woyniye village).²

There were many factors, which constrained the effectiveness of party members in promoting the developmental state ideology of the state in the first study village. First, there was a general resistance by non-party members of the community to recognize party members as “vanguards” or “models”, arguing that they are not any different from the rest of the community apart from their party membership. One of the respondents stated:

“The party members are just like us; of course they bring lessons when they go to meetings. Otherwise, they are not any different from the rest of us” (Individual interview at Woyniye village).

Second, a common complaint against the party members by the government representatives was that the party members agree to do their best and lead others, but fail to perform their duties when they go back to their community. Party members also had doubts on their legitimacy to lead their community. The following quote from one of the focus group discussions illustrates this:

“We can’t force them [the community] because the people say that we are violating their constitutional rights. We are also told not to force people but to convince. If we had the power to coerce, they would have come out. Now it is up to their discretion. Because of this, we are not really putting the theory into practice” (Focus group discussion at Woyniye village).

In study village two, the party members, especially those at the leadership position seemed to be at odds with the people. Unlike the first study village, neither the village chairman nor the sub-village leaders appeared for the watershed development campaign work during the research field work. In village council meetings, at least in one occasion, the village chairman was absent and in another one he arrived two hours late. Out of the three sub-village leaders, only one was present at the three village meetings that the researcher attended. In one of the village meetings that the researcher attended, one farmer said:

“For me, I see that we have no leaders, especially at the cabinet and development team levels. It is just a waste. People are not coming out for the work. We agreed that we should meet and discuss with the people on why they are not turning out for the campaign, but we never did so” (Field Observation at Laste Gerado village).

The village leadership was also implicated in serious allegations of corruption. During a village council meeting, the members of the council questioned cabinet members about unaccounted community financial contributions to which the leaders failed to provide proper answers. Members of the council warned the cabinet that such embezzlement of community fund would erode the trust of the people in the leadership. The leadership also found itself cornered in a bigger political problem beyond its control. In a recent conflict with a neighboring village over grazing area, the leaders in the study village were instructed by the district government to accommodate the demand of their neighbors and convince their constituency accordingly. The villagers resented that the moment the leadership decided to side with the decision of the district, people in the village also decided not to cooperate with the leadership on anything, including the watershed development campaign. There was also tension between the leadership and members of the community over the sensitive issues as the village leaders were involved in cracking down what the district government called Islamic extremists operating in the village. Many of the villagers resented their leaders for these measures as the villagers felt the leaders were siding with the government at the expense of their own people.

Because of these complicated issues, the second study village can be characterized by a weak ability of party members and those in leadership positions to use their party’s development ideology to mobilize people for development work in general and the watershed development campaign in particular. Out of desperation, the village leadership resorted to coercion to get things done. The political vacuum among the party members as well as between the party members and the villagers was observable in village development activities. Unlike the first study village, participation in the watershed campaign was too low. Although the village development agents reported that there was an average turnout of close to 50% during the campaign period of 2013/14, the chairman of the village declared that the maximum turnout was never more than 400 people, out of the expected 1400. During the entire research fieldwork in the village, the turnout at the campaign site was always less than 100 people. There was even an incident when the work was halted all together because of low turnout. The amount of work done during the 2013/14 campaign was also too small and was done poorly.

4.4 Institutions as a form of governmentality

The hegemonic developmental state ideology requires complex governmentality mechanisms of social organization and control to translate state ideologies into concrete actions on the ground. In its long history, with some crucial reforms along the way, EPRDF secured popular support through mass mobilization and organization. The governing party used both state and party lines to promote and implement its ideology. To this effect, delivery and control of socio-economic advantages such as education, health, agricultural extension and micro-credits have been helping the party to keep its grip on popular support both in rural and urban areas (Vaughan, 2011). Next, the papers examines some of the governmentality strategies employed by the party that are relevant for the watershed development intervention: top-down target setting, social organization at different levels, public conferences, regular monitoring and evaluation, reporting and feedback mechanisms, collection of statistical facts and figures, and the creation of a spirit of competition at different levels. This section explains these mechanisms in detail.

One of the main instruments of the EPRDF to promote its developmental ideology is top-down target setting. Apart from the centralized party decision-making culture discussed above (Vaughan, 2011; Bach, 2011), the budget dependency of lower level administrative units to higher level units such as the regional and national governments made top-down planning a norm rather than an exception (Ayele, 2011). Moreover, members of the ruling party occupied almost all the leadership and council member positions in the local governments (Yilmaz and Venugopal, 2008). This control of public organizations nurtured top-down development planning. On the watershed development intervention, EPRDF developed the Growth and Transformation Plan at the federal level, which was expected to develop its own plan in order to meet the nationally set targets. On their part, the regions gave targets to each district in order to meet regional targets. This top down approach made meeting those targets an arbitrary and difficult task for lower level officials. Targets set at a higher level are usually difficult to achieve at the lower level, yet it is the lower level officials that were held responsible if they fail to achieve the targets set for them.

The top-down planning process gets to ordinary citizens through public conferences held at village level. Public consultations using conferences are part of the long tradition of EPRDF public engagement forums, mainly adopted from its armed struggle culture (Vaughan, 2011). Reaching ‘consensuses’ with the people on issues that the party deems important have always been a defining characteristic of the decision-making process of the party, with farmers conferences used as the main forum for engagement (Vaughan, 2011).

At the beginning of each year, the state organized two types of public conferences for villagers, one for the party member farmers and another one for the general public, which lasted five to seven days. The village cabinet organizes the party members’ conferences, with the facilitation of the conferences led by a political representative from the district. The conferences serve as a forum of evaluation of past years’ development performance and introduction to the plans of the upcoming production year. Key decisions on the watershed development work such as, the sites for watershed development work, the number of days and hours of work, work norms and control mechanisms were introduced during these conferences. The party member conferences were followed by conferences for the general public. These conferences were organized at the sub-village level, with facilitation from the village level experts. The party members were also expected to attend these conferences in their respective sub-villages to make sure decisions made were in line with the decisions made during the party members’ conference.

In terms of participation, for study village one, the attendance was 70% for party members and 90% for the general public conferences. The debates for the party members were also more open compared to the general public conference. The public were cautious of openly resisting the conference agenda already decided by the party and the vanguard farmers. Although the researcher did not attend any conferences in the second study village, an interview with the local development agent revealed that the conference had to be adjourned repeatedly due to a lack of participants. Finally, the conferences were held under serious political pressure from the district with a threat of heavy fine for non-attendance. Informal discussions with some of the villagers also revealed that the discussions were more of informing than consulting. Many of the decisions taken were those made by the district.

Once decisions on the coming watershed development were introduced during village conferences, the state used a mix of public and political party forms of social organizations to implement the decisions. The formal social organization of the village includes the village cabinet, the village council, development teams and one-to-five teams. The study villages also had three sub-villages with one leader coordinating day-to-day activities in the sub-village. The cabinet represented the core political wing of the village with the highest decision making power, although in principle they were supposed to be answerable to the village council. The village council consisted of a group of farmers who were elected by their neighbors to represent them in the council. Since it is a political representative body, membership in the council was based on political affiliation. The village residents were further grouped into ‘development teams’. In general, the heads of each family in the village made up the core of the development teams with one team comprising of 20–30 family heads. Each ‘development team’ also contained four to six sub-teams that were called ‘one-to-five’ teams under it. One- to-five teams were the lowest level of political organization. They were called one-to-five teams because they comprised of one leader and five team members. In the study villages, study village one had 330 one- to-five teams and study village two had 210. A majority of the one-to-five team leaders were also party members and take orders from the development team leaders. The role and accountability of one-to-five teams and development teams as well as the village leadership was summarized by one of the respondents as follows:

“One-to-five teams are supposed to bring their team members to work. The development team leaders are supposed to link the government with the people. They follow up with developmental activities; they are the government of the 30 people under them. They are both responsible and accountable. The sub-village leaders are supposed to lead the development teams, give work for the teams, and evaluate the performance of the teams. We implement the village plan which we receive from the village experts” (Individual interview from Woyniye village)

Comparing the two study villages on the functioning of social organizations, study village one had better social organization compared to study village two. On all levels of leadership, those in study village one held regular meetings with much better attendance compared to their counter-parts in study village two. In study village one, each time the council meetings that the researcher attended was convened; it had a minimum of two third attendances of its members. In study village two, on the other hand, there was an instance where a council meeting had to be rescheduled three times because of low attendance. The third time it was held, it was held with only 50 of the 300 members in attendance. The same goes for the development teams and one-to-five teams.

While in study village one, villagers knew their developmental and one-to-five team leaders and members very well, in study village two they only knew their development team members and one-to-five teams are still a new institution. During the watershed development campaign work, the leadership in study village one was strict on distributing the work to the one-to-five teams. The sub-village leaders would move around the development teams and check if the work was distributed to the one-to-five teams. In study village two, on the other hand, the work was organized only in development teams. In study village one, the village cabinet held regular meetings, sometimes every day, during the watershed development campaign work. In study village two, there were hardly any command post meetings.

The day-to-day watershed development campaign period had been monitored and evaluated regularly at different levels. During the watershed campaign, the development teams held brief meetings among themselves after every campaign day. Later they joined bigger groups in their sub-village for an overall daily evaluation and discussion on work quality, attendance and control. Either sub-village leaders or village level experts facilitated these meetings. The village cabinet then met every day after the campaign work to evaluate the daily performance and prepare a report for the district. The evaluative nature of these meetings was also a result of the political culture of the ruling party, whereby it uses self-criticism as a way of promoting learning, mutual monitoring, and evaluation (Vaughan, 2011; Chinigò, 2014;Bach, 2011). These monitoring and evaluation meetings generated an exchange of regular statistical data on plans and performances organized daily, weekly and monthly throughout the watershed development campaign period. The monitoring and evaluation results were also used to award best performing villages, districts and zones within the region. The awards were meant to create a sense of competition among actors at different levels. With this regard, the first study village had a successive rank of 1–3 over the four year watershed development period out of the 34 villages in the district, while the second village was ranked last on successive years.

4.5 Countering containment institutions of the state

The two previous sections made it clear that the state used existing institutions to contain the collaborative watershed management process to its advantage. As a result, the state managed to mobilize huge number of its rural citizens nationwide for resource management.

The existing containment strategies were not in tune to address freeriding tendencies among community members, disappearing customary arrangements for collaborative work, land fragmentation, and the failure of some villagers to see the immediate benefits of the watershed development. For the government officials, a ‘village’ was the intervention unit. They homogenized the problems, aspirations and commitments of villagers while people even in the same village had different risk perceptions, aspirations and capabilities. In these circumstances, the use of soft power and pseudo democratic hegemonic and governmentality controlling institutions promoted by the government did not escape resistance, either in subtle or open ways.

When the government officials used existing institutions to impose their political-cum-development ideology on the people, local communities responded with varied forms of resistance. In this regard, absenteeism during the watershed development campaign was one of the ways villagers demonstrated their resistance. The regional guideline on mobilization stated that 80–100% of the total working age population, age 15–65, should participate in the watershed development campaign work. During the research time, in study village one, the number of attendees was always higher than the number of absentees. Penalties on absentees also worked well. In the second study village, the turn out during the campaign was too low. During the research fieldwork, attendance during the campaign has never been more than 100 people compared to the expected 1400. There were even days where the campaign work was halted for a week due to lack of participation. In fact, there were entire sub-villages that refused to come to participate.

Even when people were out for the campaign work, they still showed their resistance in a number of other ways. One of the ways was by delivering fewer results compared to the norms set by experts. According to data obtained from Guba Lafto District, the labor efficiency of study village one was around 70% and village two was around 48%. The observation of the researcher in the field was that the labor efficiency is much lower than the official figures.

On top of absenteeism and labor inefficiency, people also showed their resistance by delivering poor quality work. There was a general consensus among experts and local communities alike that there was enough knowledge and skills on soil and water conservation measures for the work involved. For at least the last five years, all one-to-five and development team leaders were trained for three days every year on soil and water conservation measures. The development agents were also trained every year for at least a week. Apart from the training, farmers as well as development agents had sufficient practical experience. In spite of these, quality of the conservation work was one of the biggest concerns of both experts and local communities. In many cases, the physical structures especially on private farm plots were not well designed and were poorly constructed. Many of the respondents attribute this to ‘ignorance’, which essentially means carelessness as a way of showing resistance.

In more serious cases, people show their resistance through vandalism, destroying the physical soil and water conservation structures constructed and the fodder trees planted. There were no serious problems in the communal and public lands in both study villages. On private farms, however, villagers tended to actively destroy the structures. This happened mainly because the conservation structures took away a lot of space compared to the small size of the villagers’ land holding. Farmers complain about the land lose despite acknowledging the usefulness of the structures. Hence, rather than maintaining the conservation structures, they ploughed them, little by little, until the structures are destroyed. Such action might not be surprising given farmers had no say over what sort of structure should be made on their farm. One respondent from study village two complained

“Our land is now very fragmented, when we make the structures, two or three of it on one farm, farmers feel like it is taking too much space and they destroy it” (Individual interview at Laste Gerado village).

5 Discussion

From an institutional diagnostic point of view, few issues are worth pondering here. First, given the diverse set of actors and institutions involved, the process of watershed development is far from just technical prescriptions, as it is often portrayed in most adaptation and food security policies. Rather, it’s planning, implementation as well as evaluation needs to be seen as ‘performance’, which has to be enacted and managed as it unfolds under a particular socio-economic and political context (Crane et al., 2011; Mehta et al., 1999). Hence, institutions that regulate relationship among various social actors play a significant role in successful resource interventions. Often times, studies on institutions focus on the by-laws, directives, regulations or policies, but miss out on how a particular institutional arrangement is produced from a combination of various institutions (Mehta et al., 1999). By refocusing on the social and political processes that produce institutions, we better understand how institutional arrangements come into being and what they are capable of doing (Mehta et al., 1999). This also explains the performance difference in the two study villages. In study village one, where there was better legitimacy of the political party leaders, there was stronger will and commitment to work towards realization of the watershed development plans and hence the institutional arrangement functioned better. In the second study village, however, the local political leaders were at odds with local communities due to their own corrupt practices as well as resentment over the district government. As a result, there was poorly functional institutional arrangement, which led to low coverage and poor quality of the watershed development work as well as its maintenance.

Second, the SRA perspective adds value to institutional diagnostics by revealing the structuration process that characterizes institutions (Jessop 2001). The institutional arrangement of the watershed development in the two study villages exhibits a relational view of structuration. On the one hand, the hegemonic developmental state political program of the state acts as a structure structuring village life in general and the watershed development work in particular. In this way, concealing itself with the term ‘public participation’, the state actors drive village actions in the direction that they desire. Looking closer, however, one can also see that the political program actually depends heavily on individual party members for its influence on villagers. Although the political program of the ruling regime structures everyday life in the study villages, it does not mean that villagers are passive victims. Rather, they use the same institutional arrangements promoted by the state to resist or avoid the state control over interventions that affect their life or take advantage of the interventions for their own good. For example, both party members and individual members in the study villages made use of the institutional arrangements such as prioritizing watershed campaign works to their locality or even their own farm. They also often site their constitutional right or the regional land use proclamation when dogging watershed campaign works. Such understanding of institutions adds value to institutional diagnostic as it reveals the ‘selective’ nature of institutions in terms of whom they favor and whom they discriminate, as opposed to the assumption that institutions enable and constrain all actors equally (Jessop, 2001).

Finally, a state-centric institutional diagnostic involves an interplay of the state, mechanisms of mobilizing a small portion of supporters and the use of these small groups of supporters, some call them ‘vanguards’, to influence the citizens at large. This requires understanding the institutional configuration that the state controls. In the case of this paper, it includes the national constitution, property rights over land, watershed work norms, political party related norms and local administrative control mechanisms.

6 Conclusions

This paper demonstrates the political nature of natural resource management interventions in Ethiopia, which contrast views that see them as a response to physical drivers such as land degradation or natural drivers such as drought (Nyssen et al., 2004; Crummey and Winter-Nelson, 2003). Resource interventions often involve institutions that bring rural communities and the state together for cooperation. However, this cooperation is not among equals. In terms of diagnosing what particular institutional settings are capable of doing and what their predicaments are within the Ethiopian context, the paper showed that institutions of natural resource management could be conceptualized as spaces of struggle between containment strategies of the state and counter containment strategies of local communities. In places where the institutional arrangements work well, it allowed the state and local communities to coordinate their action and rehabilitate degraded lands, conserve soil and water, and improve local livelihoods. In other places where the institutional arrangements face resistance from local communities, it results in failure to coordinate actions of state actors and local communities for the watershed development work even when the potential benefits were acknowledged.

The insights from the institutional diagnosis presented in this paper have practical implications for climate change adaptation and food security in Ethiopia and sub-Saharan Africa. They suggest that adaptation and food security requires proper understanding of the social and political processes of ‘production’ of institutions and ensuring that these processes are inclusive, deliberative and reflective. Combining the ‘politicized’ approach demonstrated in this paper with other approaches of policy analysis could provide a comprehensive understanding of adaptation and food security conditions in sub-Saharan Africa.

Acknowledgements

This paper is extracted from my PhD dissertation and modified to suite the special issue. The funding for my PhD study came from the German Academic Exchange Program (DAAD) and Bayreuth International Graduate School of African Studies (BIGSAS). I would like to thank Professor Detlef Mueller-Mahn for his constructive guidance during the PhD study. I also would like to thank the two anonymous reviewers for their constructive reviews and Frehiwot Tilahun for her assistance in proof reading the paper.

Notes

1 ‘Development’ is referred as ‘a normative process of becoming, a movement from poverty to wellbeing (Duffield, 1994). Developmentalism refers to a technologically modern policy choice that countries in the South take to catch up with countries in the North (Wallerstein, 2005). Developmental state on the other hard refers to ‘a coalition consisting of politicians and state bureaucrats that prioritizes economic growth ‘over all else” (Pereira, 2008: 1190). In Ethiopian case, the ‘developmental state’ ideology of the governing party EPRDF refers to the party-state with the exclusive responsibility of setting economic development priorities for the country (Bach, 2011).

2 Quotations are translated from Amharic to English by the author.