Exploring decentralized forest management in Ethiopia using actor-power-accountability framework: case study in West Shoa zone

Environ Dev Sustain (2013) 15:807 825 DOI 10.1007/s10668-012-9407-z Exploring decentralized forest management in Ethiopia using actor-power-accountability framework: case study in West Shoa zone Abrar Juhar Mohammed Makoto Inoue Received: 10 July 2012 / Accepted: 18 October 2012 / Published online: 27 October 2012 Ó Springer Science+Business Media Dordrecht 2012 Abstract Decentralized forest management is a pivotal approach in Ethiopia for balancing biodiversity conservation with demand for economic development, and for improving forest-dependent local peoples livelihoods. With the aim of filling the literature gap on Ethiopia, this paper explores the dynamics of decentralization in the forestry sector using the actor-power-accountability framework. Generally, three forms of decentralization are practiced: deconcentration to government administrative branches, devolution of selected decision-making power to local people, and delegation to enterprises. Although transfer of meaningful discretionary power to local people or to downwardly accountable lower-tier governments is a precondition for achieving positive outcomes from decentralization, this prerequisite has been realized in none of the three forms decentralization. Overall, three important trends emerged from the latest decentralization reform, which was a switch from the conservation-oriented deconcentration form of decentralization to the income generation-oriented delegation form of decentralization. Those trends are as follows: monetary income generation for local people through enterprise, albeit with possible risk of being deprived of income and subsistence opportunities on which local people depend for their livelihoods; moving decision-making power away from the grassroots; and lack of incentive to manage natural forests, a major source of biodiversity. Keywords Actor Accountability Power Decentralization Forest Ethiopia 1 Introduction Balancing biodiversity conservation with demands for economic development and improvement in the livelihoods of local people who depend on these biodiversity resources is an imperative concern of the forest policy of Ethiopia, where both biodiversity loss and poverty are serious problems. Montane forests, one of the country s important homes of A. J. Mohammed (&) M. Inoue Laboratory of Global Forest Environmental Studies, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113-8657, Japan e-mail: aj_mhmmd@yahoo.com

808 A. J. Mohammed, M. Inoue biodiversity (Sayer et al. 2005), are located in areas inhabited by the majority of the Ethiopian population, and are estimated to be diminishing at a rate of 150,000 200,000 ha per year (Zewdu 2002). In recognition of the global significance of this ongoing degradation and deforestation, montane forest remnants have been incorporated into global biodiversity hotspots by Conservation International. With population growing at a rate of about 3.194 % (2011 estimate) and about 38.7 % living below the poverty level (FY05/06 estimate), it is a major challenge to find a system of management that will minimize further destruction of natural forests by balancing protection objectives and utilization interests of the state, non-state organizations, and local communities (Zewdu 2002). Decentralization of power has been a pivotal governance tool of Ethiopia since the latest constitution took effect in 1995. For example, Article 50(4) requires the states to grant adequate power to the lowest units of government and enable the people to participate directly in the administration of those units. The forest sector of the country is not exempt from this. In fact, decentralization has been an important policy tool in forest governance (Sayer et al. 2005; Jagger et al. 2005). Although decentralization in Ethiopia has attracted some research, less attention is given to decentralized environmental governance (Tegegne 2007). In addition, as mentioned by Agrawal and Ostrom 2001, studies on decentralization of natural resources largely failed to give sufficient attention to the nature of the property rights which have been decentralized. One example is the actor-power-accountability framework developed by Agrawal and Ribot 1999. Using a modified actor-poweraccountability framework that emphasized on a shift in property right due to decentralization, this paper explores the actors involved, the nature of property rights for which decision-making power has been decentralized, and the accountability relations among actors in the different forms of decentralized forest management that are identified in Ethiopia s Oromia region. It also investigates emerging trends that have implications for biodiversity conservation, for demands relating to the country s economic development, and for the livelihoods of local people due to the latest decentralization reforms. This paper is meant to fill the gap in the literature on Ethiopia, provide constructive policy proposals, and contribute to the already existing pool of literature on decentralization theory by positing a modified actor-power-accountability formwork for analyzing decentralization. The next section details the theoretical setting of the paper, which is followed by a section on research methods. The fourth section presents the results and discussion, and the paper is summed up with a conclusion and policy implications. 2 Theoretical framework In the context of natural resource governance, decentralization can be defined as any act by which a central government cedes decision-making rights over resources to lower-level actors and institutions (Agrawal and Ribot 1999) and in the process reshapes power and accountability relationships among involved actors (Devarajan et al. 2007; Yilmaz et al. 2010). It is a political and economic process that implies a redistribution of power and resources (Larson and Soto 2008) to different actors including community organizations (Chhatre 2008). Decentralization of natural resource management may be more suited to supporting better natural resource governance and to improving livelihoods of resource users in developing nations for these reasons: Unlike other sectors, such as health or education, which are sinks rather than sources of income, natural resources are significant sources of wealth and power (Ribot 2003); natural resources already exist physically within a specific

Exploring decentralized forest management 809 local area, with their own particular history and tradition of resource management, implying that certain decisions affecting those resources have probably been made by local people for centuries (Larson 2003); and the reality of local history and experience also means that local knowledge regarding the particular characteristics of a resource, such as a forest, and its management may have been developed over time (Ribot 2002; Larson 2003). 2.1 Forms of decentralization Decentralization reforms come in forms such as deconcentration, delegation, and devolution (Kinuthia-Njenga 2002; Larson 2003; Yilmaz et al. 2010). Deconcentration or administrative decentralization is when a central government redistributes decision-making power to its own representatives within defined geographic units. When decision-making power is transferred to lower-level elected or appointed leaders (Oyono 2004) or to communities (Larson and Soto 2008), it is called devolution. Delegation is the transfer of decision-making power to a semi-autonomous organization that is not a political and administrative extension of a government (Oyono 2004). The above forms of decentralization are not monolithic. To understand them, it is therefore useful to deconstruct them into their main elements (Smoke 2003). The three fundamental elements of any form of decentralization are actors involved, types of decision-making powers transferred, and accountability relations of actors (Agrawal and Ribot 1999). 2.2 Actors Decentralization of natural resource governance is a strategy to facilitate transfer of power among actors, bringing new actors into power in the process. Therefore, it is imperative to identify the actors who are involved in policy implementation (Agrawal and Ostrom 2001). Key actors involved in decentralized forest governance can be placed into one or more categories: citizens/grassroots, policymakers (politicians, lawmakers, regulators), or organizational and frontier providers (government ministries, private sector, cooperatives, foresters, extension workers, development agents, and others) (Devarajan et al. 2007; Yilmaz et al. 2010). Each of these actors typically has particular relations of accountability and has certain types of powers (Agrawal and Ribot 1999). 2.3 Decision-making power over property rights A property right as defined by Agrawal and Ostrom (2001) is a relationship among actors with respect to things like natural resources. It represents a claim to benefit streams with corresponding duties attached to them (Bromley 1992). In general, property rights can be of five types: access rights, management rights, use rights, exclusion rights, and alienation (transfer) rights (Schlager and Ostrom 1992). Of these, management, exclusion, and use rights are particularly important in natural resource governance (Agrawal and Ostrom 2001). The means by which power is distributed among actors in decentralization is decisionmaking power over these property rights that can be held by a wide variety of actors (Acheson 2006). These decision-making powers can be further unpacked into power to create rules or modify old ones, power to implement the rules, and the power to ensure compliance with the rules (Agrawal and Ribot 1999). The legal structure of decentralization, which is mostly inferred from official documents, reveals the type and extent of decentralization that is officially being implemented (Larson 2003). The legal structure

810 A. J. Mohammed, M. Inoue may not, however, tell whole history. Whatever the official central government policy may be, the actual power held by actors at different levels may be constrained because higherlevel government institutions and personnel may often stand in the way of implementing it (Larson 2003). 2.4 Accountability relation Accountability, a mechanism through which counter powers are exercised by those subject to actors holding decentralized power (Agrawal and Ribot 1999), is a principle element of decentralization that assures good governance (Ackerman 2004). Accountability comprises a set of mechanisms that can ensure that policy outcomes are consistent with actors needs and aspirations (Ribot 2004). This mechanism can be either vertical, resulting in downward or upward accountability, or horizontal, in which the power holder needs to report to other, same-level officials and agencies (Ackerman 2004). Whether vertical or horizontal, all modes of accountability are relational. To understand the nature of accountability, therefore, it is necessary to examine the actors between whom relations of accountability exist (Agrawal and Ribot 1999). The most commonly cited mechanism through which actors hold power holders accountable are the electoral process, a vertical accountability mechanism. Elections, however, occur once every few years, force an incredible diversity of opinions and evaluations together into a single ballot, and only hold elected officials accountable (Ackerman 2004). Nevertheless, it is certainly possible that in a particular decentralization effort, powers are transferred to actors other than elected officials (Agrawal and Ribot 1999). In such cases, vertical accountability mechanisms, such as elections, have to be complemented by horizontal accountability mechanisms (Ackerman 2004) or it is necessary to develop mechanisms in which these actors are held accountable to local elected officials (Ribot 2002). 2.5 Outputs and outcomes from decentralization At the center of the argument in favor of decentralizing, natural resource governance is the potential to bring decision making closer to grassroots actors (Larson 2005), reduce administrative and management transaction costs, and improve natural resource governance by using the proximity of local participants to decision-makers and access to local skills and information. (Agrawal and Ribot 1999). In addition, it is argued that having local people make local decisions results in more effective local investments and management, and ultimately in more socially and environmentally sustainable development (Ribot 2002). This, however, requires achieving democratic decentralization that is transfer of meaningful discretionary decision-making power for managing and using natural resources to either local communities (Dewees et al. 2010) or to lower-tier governments that are accountable to the populace (Agrawal and Ribot 1999; Ribot 2002). Prior to assessing the outcomes of decentralization on forests and forest-dependent people and gauging the efficacy of decentralization strategies, it is therefore useful to assess the output of decentralization, that is, extent of democratic decentralization that has actually occurred (Hajjar et al. 2012). Democratic decentralization is not an end by itself. Rather it is a means to desired outcome from decentralization. The central outcomes from decentralized forest governance are demand for positive social outcomes like improving livelihoods of the local households as well as improvement or conservation of natural capital, that is, forest condition (Larson and Soto 2008).

Exploring decentralized forest management 811 2.6 A modified actor-power-accountability framework for analyzing decentralization Previously used frameworks functioned by categorizing decentralization as political, fiscal, or administrative, and although they are critical for statecraft, they do not hold up as analytical categories. The actor-power-accountability framework has been the most influential framework to analyze decentralized forest governance (Coleman and Fleischman 2011). It provides a tool to understand the principles that make decentralization effective (Larson and Ribot 2004). The framework encompasses the three vital dimensions that underlie all acts of decentralization: actors, powers, and accountability. Without an understanding of the powers of various actors, the domains in which they exercise their powers, and to whom and how they are accountable, it is impossible to analyze the extent to which meaningful decentralization has taken place (Agrawal and Ribot 1999). However, the framework has failed to give emphasis on silent aspect of forest governance, that is, the shifts in property rights due to decentralization policies (Agrawal and Ostrom 2001; Larson and Soto 2008). In addition, it has largely neglected the impact of decentralization policies on local-level environmental and social outcomes (Andersson and Gibson 2006; Coleman and Fleischman 2011). In this paper, we introduced a modified actor-power-accountability framework (Fig. 1) that atones for the aforementioned limitations. 3 Research method 3.1 Study site selection Ethiopia s 1995 constitution established a federal government structure comprising nine national-regional state governments plus Dire Dawa and Addis Ababa City Administration. Regions are further divided into zones which are intermediate administrative units located Fig. 1 A modified actor-power-accountability framework

812 A. J. Mohammed, M. Inoue between regional governments and Woreda governments. A woreda is a government unit composed of six to 10 peasant associations (kebeles) and corresponds to districts or counties in other countries. This research was conducted in West Shoa zone of Oromia regional state. Oromia region was selected because of its vital position in terms of both forests and population. It is Ethiopia s largest by both population and area. It also has about 70 % of the country s natural high forest. Deforestation is serious in this regional state. It is believed that some 60 years ago, Oromia montane forests were estimated at about 14.7 million ha, which covered about 40 % of Oromia at that time. These forests were located in the central, western, southern, eastern, and southeastern mountainous areas. Currently, the remaining montane forest of the region is estimated to be only 7 % of Oromia (Feyissa 2007). From Oromia region, West Shoa zone was purposively chosen for this case study because it represents diverse forms of decentralized forest management. The Chilimo participatory forest management (PFM) project, which is one of the pioneer projects that have implemented the devolution form of decentralization in the country, is located in West Shoa zone. In this project, local people have formed forest user groups (FUGs) that later became cooperatives. The FUGs signed agreements with woreda government representatives to protect, manage, and use Chilimo forest. In addition, the zone covers Jibat forest, which is managed through a delegation type of decentralized forest governance. Both of these forests were managed by their respective woreda agriculture and rural development offices. Chilimo forest and Jibat forest are among the very few remnants of dry Afromontane forest in the country. Chilimo forest covers about 5,000 ha, while the size of Jibat forest is about 38,000 ha. Both forests are composed of natural forest containing endemic and endangered tree species like Juniperus procera, Podocarpus falcatus, Prunus africanum, and Hagenia abyssinica, as well as plantations which are mostly dominated by Eucalyptus species. Most of the natural forests are managed for conservation purposes while the plantations are for production. 3.2 Data collection and analysis Document analysis and open-ended interviews were employed to collect data. About 15 documents 1 comprising policies, proclamations, regulations, and action plans that have implications for forest governance were identified and reviewed to explore the de jure framework of decentralization. We also collected general data on income opportunities created for local people by forest enterprises, and data on forests recently sold by forest cooperatives, from the respective enterprise and cooperative offices. In addition to 1 The documents include Ethiopia policy and strategy on forest development, conservation, and utilization; federal proclamation to pronounce the coming into effect of the constitution of the federal democratic republic of Ethiopia; federal environmental impact assessment proclamation; federal environmental protection organs establishment proclamation; federal land administration and land use proclamation; federal proclamation to provide for the development, conservation, and utilization of forests; growth and transformation plan (GTP) 2010/2011 2014/2015 (draft); PASDEP: a plan for accelerated and sustained development to end poverty (PASDEP) (2005/2006 2009/2010); Oromia revised regional proclamation; forest proclamation of Oromia region; a proclamation for the reorganization and redefinition of the power and duties of the executive organs of Oromia regional state; a proclamation to provide for the establishment of Oromia bureau of land and environmental protection; a regulation to provide for the establishment of Oromia regional state forest and wildlife enterprise; an agreement signed between FUGs (cooperatives) and government representatives; and the internal bylaws of FUGs.

Exploring decentralized forest management 813 document analysis, open-ended discussions were conducted with a total of 44 actors. 2 Purposive sampling was used to select the organizations and the interviewees from the organizations. Owing to the diversity of offices that are policy makers, implementers, advocators, or clients of ongoing decentralized forest governance, and to the division of labor within organizations, it was necessary to interview a relatively large number of samples (44 samples). The interviews were undertaken in the interviewees offices (for those employed by government or businesses) or their homes (for local people). Each interview lasted 1 2 h. Data collection was undertaken from November through December 2010, September through October 2011, and January through February 2012. To analyze outputs from decentralization, pattern-matching technique is carried out. Pattern matching is one of the most desirable techniques in case study which compares an empirically based pattern with that of theoretically designed or predicted (Yin 2008). In our study, it is used to check whether the empirical data (field reality) on extent of decentralization fit with the theoretical democratic decentralization. For this purpose, analysis of what is decentralized to whom and with what accountability mechanism and direction is done. Actors with decision-making power on rule making and/or implementing and/or execution of management, exclusion, and withdraw right on the forest are classified into three levels based on their proximity to grass root populace. Forest user groups/forest cooperative leaders, peasant association leaders, and forest guards are identified as close actors, while woreda government and woreda administrative offices as well as district level of Oromia forest and wildlife enterprise (OFWE) are taken as intermediate actors. Finally, all the others including regional government, branch and head office of OFWE are taken as distant actors. 4 Results and discussion Overall three forms of decentralization have been practiced since decentralization was formally inaugurated by the 1995 proclamation of the nation. These are the deconcentration form of decentralization until 2004 in Chilimo forest and until 2007 in Jibat forest, the devolution form of decentralization from 2004 to the present in Chilimo forest, and the delegation form of decentralization from 2007 to the present in Jibat forest. 4.1 Deconcentration of forest governance to lower administrative branches This form of decentralization involves the distribution of power from the central government to its own representatives within defined administrative and geographic units who are accountable to their superiors in a hierarchy (Agrawal and Ribot 1999). 2 The interviewees were from these organizations (numbers of interviewees are in brackets): Ministry of Agriculture and Rural Development (MARD) [3]; environmental protection authority (EPA)[1]; Oromia agriculture and rural development bureau (OARDB), natural resource core process [3]; Oromia bureau of land and environmental protection (OBLEP), environmental protection core process [1]; OBLEP, land use planning core process [1]; Oromia forest and wildlife enterprise (OFWE), head office [3]; OFWE, branch office [2]; OFWE, district office [4]; zone administrative office [1]; Woreda government [1]; Woreda judiciary [3]; Woreda agriculture and rural development office (WARDO) [4]; Woreda land and environmental protection office (WLEPO), Environmental protection work process [2]; WLEPO land use planning work process [1]; peasant associations (PAs) [3]; Melka Mahber NGO [1]; Oromia water work design and supervision enterprise [1]; and local people [upper class 3, lower class 6].

814 A. J. Mohammed, M. Inoue 4.1.1 Actors involved The actors involved in making or implementing policy, or involved as clients in this form of decentralization in the regional and woreda governments, are the regional bureau of agriculture and rural development (RBARD), the woreda agriculture and rural development (ARD) offices, woreda courts, peasant associations (PAs), forest guards, and local people. The regional government has both the legislative branch that is formed by elected members and the executive branch formed from heads of administrative bureaus of the region. One of the administrative branches is RBARD. The bureau has different departments including a bureau that is responsible for drafting forest-related rules and implementing forest policy. The head of the region is elected from the region s council, which comprises elected members from woredas in the region (Federal Proclamation No. 1/95; Oromia Proclamation No. 72/2003; Oromia Proclamation No. 87/2004). Like regions, woredas also have legislative bodies called woreda councils and executive bodies called woreda administrative councils. Woreda ARD offices are one of the woreda administrative branches. Woreda councils are formed by elected members from peasant associations inside each woreda. Each PA is represented in the council by three individuals. The other important actor, which is independent from the woreda government and which has a judiciary function, is the woreda court (Oromia Proclamation No. 46/2001). The lowest government unit is the PA. PAs too are required to have their own legislative, judiciary, and executive branches (Oromia Proclamations No. 87/2004 and No. 46/2001), but probably due to the lack of incentive, the only functional branch is the judiciary branch, which is composed of the PA head and secretary, and one member. In general, woreda governments and PAs can be understood as local governments owing to their proximity to the populace. 4.1.2 Type and nature of powers transferred The natural resource branch of RBARD is responsible for developing public instruments such as the forest policy and proclamations of the region. The power to make final decisions is held by the council of the region, which approves these public instruments before implementation (Federal Proclamation No. 1/95; Oromia Proclamation No. 46/2001). Although the region does not have a formal written forest policy, the 2003 forest proclamation of Oromia was prepared by the natural resource branch of the agriculture bureau of the region. The region s public instruments must be in accord with federal public instruments (Federal Proclamation No. 1/95; Federal Proclamation No. 542/2007). However, this requirement was found to be nominal. Detailed operational decisions such as what to plant, when and where, how much to harvest, when and who should harvest, and how it should be harvested are made at the woreda level. Woreda ARD offices, through their natural resource branches, assess the forest and provide woreda heads with plans for their final decisions. These decisions, however, must jibe with regional and federal forest policy. After getting approval from the woreda government, the Natural Resource Bureau executes the plan. The forest is protected by guards hired by the woreda ARD office. Guards are mostly hired from the peasant association military. Rule breakers, that is, illegal loggers or charcoal makers, are taken to the woreda court, which determines their punishment in accordance with the federal forest proclamation. Woreda governments have relative autonomy in budget allocations to their executive bureaus. They receive block budgets from regional governments, which likewise receive their block budgets from the federal government.

Exploring decentralized forest management 815 Local people are required to participate when labor mobilization is needed for forest activities. They may take forest litter for subsistence fuelwood. However, local people have no decision-making power in legislation, the judiciary, or forest policy implementation. In sum, in this form of decentralization, the management, exclusion, and use right decision-making power to make and implement rules as well as to adjudicate is transferred either to regional governments or to administrative branches of the state. As mentioned by Agrawal and Ribot (1999), it is certainly possible in a particular decentralization effort to transfer powers to administrative branches of the state rather than to representative local governments at the same level. Whether such transfer of power leads to effective decentralization, however, depends on the nature of accountability relations among the actors (Agrawal and Ribot 1999). 4.1.3 Accountability relations Local people are held accountable for their actions in forests by forest guards, who can sue them in woreda courts. The guards are held accountable by woreda ARD offices through their power to hire and fire the guards. The woreda ARD offices in turn are upwardly accountable to woreda governments and to Zone and Regional ARD bureaus. The office must make reports for zonal ARD bureaus, which are administrative extensions of the regional ARD bureaus that supervise and determine whether woredas are executing regional policies. Because woreda ARD offices budgets are allocated by woreda governments, the offices are also held accountable to woreda governments through their salaries and other benefits. The last important accountability relation is between local people and two important governments: the regional and woreda governments. The only mechanism to keep these governments actions downwardly accountable to local people is elections. However, elections are a very weak accountability enforcement mechanism (Ackerman 2004) for three reasons. First, local people do not elect the woreda head, and they elect council members. As such, they lack direct influence on who will be the head. Second, elections are held once in 5 years. Even if local people were to elect the head directly, the forest dwellers who are directly affected by forest governance represent a small portion of the total population of even the woreda, let alone the region. That means it is next to impossible to enforce their demands. 4.2 Devolution of power to local people The second form of decentralization is the transfer of power to local people (devolution) that is practiced through participatory forest management (PFM) policy. A well-known project using this scheme is the Chilimo PFM project. In this project, local people in and around Chilimo forest signed an agreement with the woreda ARD office to protect, manage, and utilize Chilimo forest. 4.2.1 Actors involved Actors that have considerable role in this form of decentralization are local people, leaders of newly formed cooperatives or forest user groups (FUGs), woreda courts, woreda ARD offices, woreda and regional governments, and NGOs (agreement document). Communities in and around Chilimo forest signed the first agreement with the woreda ARD office in

816 A. J. Mohammed, M. Inoue 2004. An international NGO called Farm Africa has played a pivotal role in advocating both sides to bring them to the table and sign an agreement. Local people were organized into 12 FUGs, of which eight later became cooperatives. The NGO also was the major player in covering the transaction cost of the decentralization process as well as helping the community in training and livelihood activities. 4.2.2 Type and nature of powers transferred In this scheme of decentralization, decision-making power on forest management and on excluding unlawful forest users is fully devolved from the woreda ARD office to the local people. Local people and the leaders of cooperatives and FUGs are empowered to make decisions including what to plant, where, when and by whom, who should guard the forest, and when. Cooperative/FUG leaders also have the authority to decide on who should be allowed to engage in minor income-generating activities such as selling fuelwood in Ginchi, the nearby town. They are also authorized to make final decisions on who should get logs for subsistence uses such as building or repairing houses. In addition, they have the power to resolve conflicts arising from rule infractions as long as the offender abides by punishments prescribed by leaders. If a wrongdoer refuses to accept leader-prescribed punishment, the case is passed to the local woreda court. Cooperatives/FUGs had to seek permission to harvest wood for income from the woreda ARD office until 2007, and since then from the district-level office of the Oromia forest and wildlife enterprise (OFWE). Before 2007, the woredas ARD offices assessed forest resources to determine whether forests that cooperatives asked to harvest had reached harvestable size and also whether the requested area is sustainable, before submitting the report to the woreda head for final approval. In line with the 2007 region-level power transition from Regional ARD bureaus to the OFWE, experts from the OFWE district office would assess the condition of forest either on their own initiative or when local people ask to harvest. The district staff sends the assessment report to the OFWE branch and head offices to get their final approval. The other important actor is NGOs. Although NGOs do not have any formal decisionmaking power, their role as advocates and as financing sources for the decentralization process is significant, albeit their functioning has been limited by the much-debated charity and society proclamation, Federal Proclamation No. 621/2009. 4.2.3 Accountability relations Four accountability relations are worth noting in this devolution form of decentralization, which was implemented until 2007. The first is accountability of local people to cooperative/fug leaders, who are authorized to enforce exclusion and management rules and to punish rule breakers. The second is downward accountability of leaders to cooperative/ FUG members. The third accountability relation is that local people to the woreda ARD office. Local people can vote leaders out office, but for this to be effective the woreda ARD office must grant its final approval. This, however, was found to be a loophole to divert the balance of leaders accountability toward the woreda administration instead of toward local people. In two cooperatives, local people were dissatisfied with the service they obtained from their leaders. However, they could not vote out the leaders because the woreda administration resisted the change. The reasons that woreda administrations resist power change are the belief that the existing leadership is doing well from the viewpoint of conserving the forest, and patronage networks which some woreda staff members have

Exploring decentralized forest management 817 with the leaderships of the cooperatives (Mohammed and Inoue 2011; Mohammed and Inoue 2012). The last accountability relation is the accountability of all cooperatives/fugs to the woreda ARD office. According to the agreement signed between the woreda ARD office and the cooperatives/fugs, the ARD can take forests away from cooperatives/fugs if they do not properly manage the forests. The direction and mechanisms of these four accountability relations accountability of local people to cooperative/fug leaders, accountability of leaders to FUG/cooperative members, accountability of cooperative/fug leaders to district office of the OFWE, and accountability of the whole cooperative/fug to district office of the OFWE after 2007 are similar to those before 2007, as described above, except that the OFWE district office replaces the woreda ARD office. Two additional vital accountability relations are the accountability of the whole enterprise to the regional government, and the upward accountability of the hierarchy of offices within the OFWE (discussed in detail under Delegation to Oromia forest and wildlife enterprise ). 4.3 Delegation to Oromia forest and wildlife enterprise As of late 2007, the governance of the forest priority areas has been delegated to the then Oromia regional forest enterprise supervising agency (ORFESA). The agency comprises eight enterprises. The major reason for the creation of the agency is the demand by forest professionals to institutionally and financially empower the forestry sector of the country by establishing an autonomous organization that can sustain itself and contribute to the country s GDP (Oromia Regulation No. 122/2009). Previously, less attention was given to the forestry sector mainly because of the claim that it contributes little to GDP. The proportion of the budget which the forestry sector obtains, compared with that of agriculture, is much lower even than that in other African countries, which have likewise been said to give minimum attention to the forestry sector (see Dewees et. al 2010 for a review of comparisons of some African countries in terms of agriculture and forestry budget allocations). For the same reason, the natural resource section of the agriculture and rural development Bureau, which is the agency responsible for governing forests, used to get a tiny amount of the total budget of the regional and woreda governments and small fraction of the budget allocated to the Bureau of agriculture and rural development (Table 1). After the adoption of the 2009 Oromia Regional Forest Regulation, the agency s name was changed to Oromia forest and wildlife enterprise (OFWE), while the previous enterprises that constituted the agency were converted to branch offices. Currently, all the forest priority areas in Oromia region are under the OFWE. The current structure of the OFWE has three levels: OFWE head office, mid-level branch offices, and low-level district offices. 4.3.1 Actors involved Important actors in this scheme of decentralization are the regional government, the OFWE, its branch and district offices, the woreda land and environmental protection (LEP) office, local people, and the woreda court (Oromia Proclamation No. 147/2009; Oromia Regulation No. 122/2009).

818 A. J. Mohammed, M. Inoue Table 1 Budget allocated for Agriculture and Natural Resource Bureau from total regional and woreda government budgets over the last five years (currency ETB) Government Fiscal year Total budget Agriculture (% of total budget) Natural resource (% of total budget) Region 2005/2006 3,337,171,264 428,984,774 (12.85) 2006/2007 6,259,292,648 414,085,868 (6.62) 2007/2008 8,331,060,054 610,090,176 (7.32) 2008/2009 10,672,557,811 884,957,442 (8.29) 7,399,698 (0.07) 2009/2010 8,935,238,289 534,172,818 (5.98) 4,060,486 (0.045) Woreda 2006/2007 8,719,974 1,207,020 (13.84) 142,535.8 (1.631) 2007/2008 13,807,461 1,889,272 (13.7) 179,450 (1.3) 2008/2009 18,328,228 2,587,513 (14.1) 245,749 (1.34) 2009/2010 22,033,170 2,525,551 (11.4) 121,556 (0.55) 2010/2011 25,617,542 1,911,611 (7.46) 145,489 (0.58) Source: Data from regional bureaus and woreda offices 4.3.2 Type and nature of powers transferred The regional government has the final decision-making power for major issues such as regional proclamations, regulations, projects, and action plans from executive offices, as well as for the delegation of authority over the forest of Oromia region to OFWE and other semi-autonomous organizations (Federal Proclamation No. 1/95; Oromia Proclamation No. 46/2001). The OFWE head office has the supreme decision-making power over the OFWE. It also makes final decisions on high-value sales from the forest and large expenditures by the OFWE, developing general guidelines for forest management, and hiring the heads and management members of branches and district offices of the OFWE. The role of a branch office is more or less to manage the district forest under it. It also develops inventory forms and other standards, based on specifications from the head office, to be used by district office experts. The district offices make decisions on overall operations including guarding, planting, and harvesting. Forest guards and temporary and permanent workers at nurseries and for other forest activities are also hired by district offices. Local people participate as hired laborers in forest management and protection. 4.3.3 Accountability relations Five important accountability relations emerge from this form of decentralization. First, the OFWE is accountable to the regional government as per the agreement they have signed, and the regional government is empowered to withdraw delegated authority from the OFWE if its management is unsatisfactory. Second is the upward accountability of the lower levels of the OFWE to the higher levels, that is, the district to the branch office and the head office, and the branch office to the head office of the OFWE. This accountability is based on the power of the head office to assign, hire, and dismiss management personnel of the branch and the district. The third accountability relation is that of local people and guards. Guards have power to take the rule breakers to woreda court. The fourth accountability relation is that of guards to the district office, which can hire and fire them. The fifth accountability relation is that of the regional government to the local people

Exploring decentralized forest management 819 through elections. As discussed under the deconcentration form of decentralization, this is the weakest form of the accountability mechanism. 4.4 Outputs from the three forms of decentralization In deconcentration form of decentralization, decision-making power on rule making for management, use and exclusion right is held at regional and Woreda level which are distant and middle-level actors in terms of proximity to local people. The adjudication power for all the key property right is also held by the middle-level woreda court. The only decisionmaking power that is very close to local people is implantation of exclusion right which is held by guards selected from local people (Table 2). However, guards are upwardly accountable to Woreda. The downward accountability of woreda and regional government to local people through election is also very weak. The outputs from the devolution form of decentralization before 2007 and during transition from the ARD office to the OFWE can be characterized as a central government extending its centralized authority under the guise of decentralization policy (e.g., Gibson 1999; Smoke 2003; Mohammed and Inoue 2012) and a central government forcing costly policies onto local people (Agrawal and Ribot 1999; Gibson 1999; Blair 2000, Mohammed and Inoue 2012). Only decisions on forest protection and management have been fully decentralized for local people, while ARD offices placed limits on requests by local people to harvest plantation for income (Table 2). However, in recent years, the OFWE has allowed local people to harvest and sell forest products which are discussed under outcome from decentralization. Local people can also keep leaders of forest cooperative downwardly accountable through periodic evaluation and election. This accountability mechanism, however, is also structured to be utilized as a loophole by higher-level actors. The case for delegation is also similar to deconcentration form of decentralization. Except for the guards that are involved in the implementation of the exclusion responsibility, none of the actors that are closer to grass root have discretionary decision-making power on the forest. The decision-making power for rule making, implementing and adjudication of use right and management right are held by either the head office of OFWE and branch office of the OFWE, whom both are distant actor from local people, or the middle-level district office of OFWE and Woreda court (Table 2). Such outputs from the three forms of decentralization, particularly that of deconcentration and delegation form of decentralization, were found to be far from major objective of decentralized governance in the country which is achieving democratic decentralization. Such discrepancy was found to be a result of manipulation of deceiving terminologies by different actors to maintain their own power. For example, the term local participation has gained currency recently in Ethiopian discourse; including that on forest policies, proclamations, and statements by policy makers and implementers. The meaning of local participation as used by that currently in power, however, was found to be nothing more than controlling local people. According to interviewees from the head office, branch office, and district office of the OFWE, local participation means creating awareness among local people about the importance of the forest, and then getting them to participate in forest guarding and development activities, and sharing the benefits with them. The reason given for local participation in operational-level activities only but not in the decision-making process is the property-right dilemma of local people. Specifically, local people lack an awareness that the forest is and will be there for them and that lack makes them act opportunistically, resulting in deforestation. This lack of property rights awareness was also cited as a reason for the OFWE s plan to promote Joint Forest Management

820 A. J. Mohammed, M. Inoue Table 2 Synthesis of actors and their decision-making powers in the three forms of decentralized forest governance Decisionmaking power Nature of decisionmaking power Deconcentration Devolution Delegation Management Exclusion Use Rule making Regional governments, regional and Woreda ARD bureaus Leaders and members of cooperatives/fugs Execution Woreda ARD bureau Leaders and members of cooperatives/fugs Judiciary Woreda court Leaders of cooperatives/fugs, and/or woreda court Rule making Execution Regional government, regional and woreda ARD bureaus Woreda ARD bureau and forest guard Leaders and members of cooperatives/fugs Leaders and members of cooperatives/fugs Judiciary Woreda court Leaders of cooperatives/fugs, and/or woreda court Rule making Regional and woreda governments, regional and Woreda ARD bureaus Woreda ARD (district office of enterprise), leaders and members of cooperatives/ FUGs Execution Woreda ARD bureau Woreda ARD (district office of enterprise), leaders and members of cooperatives/ FUGs Judiciary Woreda court Leaders of cooperatives/fugs, and/or woreda court Head, branch, and district offices of enterprise District office of enterprise Woreda court Head, branch, and district offices of enterprise District office of enterprise and forest guards Woreda court Head, branch, and district offices of enterprise District office of enterprise Woreda court Source: Document analysis and interviews (JFM) instead of the existing PFM, thereby limiting local people s decision-making power on the forest. In the case of Chilimo forest, so far, the OFWE is abiding by the previous agreement signed between local people and the previous Natural Resource Bureau, under which local people get 70 % of the benefits from the sale of forest products, and also have decision-making power over operational activities, but the transition to JFM will eliminate this decision-making power and the benefits that accrue to local people. The formal documents that empower local people to participate in decision making are the environmental policy of the government and the environmental proclamation. All other government and non-government agencies are required to check their activities against the guidelines in these documents, which include the condition that agencies shall support community and individual resource users to sustainably manage their own environments and resources; ensure uninterrupted and continuing access to the same land and resources; ensure social equity, particularly in resource use; and ensure acquisition of power by communities to make their own decisions on matters that affect their lives and environments. The offices responsible for enforcing this policy are the environmental protection authority at the federal level and the regional and woreda land and environmental

Exploring decentralized forest management 821 protection bureaus at the regional and woreda levels, respectively. But in fact, the implementing offices at various levels from federal to woreda were found to be debilitated, even though they were aware of their responsibilities. Sometimes, they are called antidevelopment when they try to interfere in development activities which they believe have negative environmental or social consequences. Currently, it can be said that the offices are paralyzed. For example, OFWE interviewees said that they have never had any communication with the land and environmental protection office, except during re-demarcation of Jibat forest. The only activity undertaken by this office is the inconsistent regulation of whether products on the market are from legally harvested forest products. 4.5 Important outcomes from ongoing decentralized forest governance Two important outcomes are worth noting with regard to the implications of recent decentralization reform for balancing the conservation, economic development, and social outcomes from ongoing forest governance. 1. Income generation for local people albeit with possible risk of being deprived of income and subsistence opportunities which local people used to depend upon for their survival: The transition from conservation policy toward forest policy oriented toward generating income has been a blessing for local people under the devolution form of decentralization, from the perspective of cash income. For example, the Natural Resource Bureau has forbidden the sale of matured plantation by communities in Chilimo forest. However, since the OFWE took over the decision-making power of the woreda ARD office in 2007, restrictions on selling have become very liberal. Before 2007, the communities earned about 392,500 Ethiopian birr (ETB) by selling plantation. However, since 2007, the total amount of income generated to 5,086,391 ETB. This income is 70 % of the total income from the sold plantation. In accordance with the agreement, 30 % of the income goes to the OFWE. As mentioned by Acheson (2006), the difficulty of managing resources like forests is that two different kinds of rules must be put in place to solve two different kinds of problems. First, property rights must be defined and devised for identified groups. Second, those permitted to exploit the resource have to agree to establish rules curbing their exploitation rate (i.e., management mechanisms). Doing one without the other will not achieve good resource governance. Establishing FUGs/cooperatives to manage and utilize the forest has achieved the first requirement. However, before 2007, those involved in benefiting from a resource, that is, forest user groups and the district Natural Resource Bureau, have underutilized the resource because of the natural resource office s conservation objective. Since the OFWE took over, however, the resource is being overexploited because the OFWE s objective is generating cash income. The most reasonable policy and legal document for making the system accountable for both social and environmental outcomes are the federal forest policy and federal forest proclamation, respectively. However, the agencies which implement the environmental policy and proclamation the environmental protection authority, OBLEP, and WLEPO are currently paralyzed. In the delegation type of decentralization, about 997,109 ETB and 1,149,430 ETB were paid in wages and salaries to totals of 1,705 and 3,566 local people in 2009/10 and 2010/11, respectively (Table 3). However, this income seemed very small considering that local people are restricted from any de facto use right they previously had on the forest. In addition, the number of jobs created per peasant association is very small. For example, only five people were hired as guards from Tutu.