Partnerships for protected area conservation in Rwanda

Transcription

1 The Geographical Journal, Vol. 172, No. 4, December 2006, pp Blackwell Publishing Ltd Partnerships for protected area conservation in Rwanda EUGENE RUTAGARAMA* AND ADRIAN MARTIN *International Gorilla Conservation Programme, c/o African Wildlife Foundation, British American Centre, Mara Ragati Road, PO Box 48177, Nairobi, Kenya School of Development Studies, University of East Anglia, Norwich NR4 7TJ This paper was accepted for publication in March 2006 This paper seeks to contribute to Rwanda s ongoing restructuring of protected area management. The early part of the paper draws on experiences elsewhere in Africa, as well as key contexts in Rwanda, to assert that conservation of national parks may best be served by flexible and inclusive partnerships that seek to integrate conservation activities across different agents and scales. This assertion is then explored more critically through empirical research that investigates the views of potential conservation partners. The findings suggest that attempts to develop partnerships that are built around national parks will face difficulties. Whilst there is a general willingness to be further involved in park management, this is complicated by cleavages in beliefs about how wider participation might be implemented. In particular, it is only international conservation NGOs that currently seem to be comfortable with the national park approach to conservation management: only they see themselves as having the expertise to be decision-making partners and only they would want their role in a partnership to be formalized. KEY WORDS: Rwanda, conservation management, stakeholder survey, protected areas, community conservation, partnerships Introduction The question of what constitutes best practice for biodiversity conservation in developing countries has generated vigorous debate in recent years. A spectrum of views abound, although it has become analytically convenient to think in terms of two main camps. Firstly, there is the preservationist camp which tends to favour centralized power in order to discourage resource use by human populations (an approach commonly referred to as fortress conservation ). This has to a great extent been supplanted by the community conservation camp which seeks to integrate conservation and development objectives by empowering local people to use their natural resources in sustainable ways. However, the question of whether state or community should be responsible for conservation is rather a crude one. This paper reviews the emergence of a less polarized narrative that looks beyond both the community and the state as the best locus for control. This body of thinking is really an extension of community conservation thinking that has emerged from field experience over the past two decades. It retains the commitment to integrating conservation and development objectives but more strongly emphasizes the role of partnerships and networks. We explore the prospects for multi-stakeholder, cross-scale conservation partnerships in Rwanda, both in terms of the pressures and opportunities that are driving policy in this direction, and in terms of the problems that will be encountered as policy moves towards implementation. The paper focuses on Rwanda s three national parks. Protected areas in Rwanda Rwanda s protected area network currently covers 8.4% of the total land area of km 2 (MINAGRI 2003). The bulk of this is made up of the three National Parks (Figure 1). The Akagera National Park (ANP, Ha) combines wetland and savanna habitats and has one of the most /06/ /$00.20/0

2 292 Partnerships for protected area conservation in Rwanda Figure 1 Study sites (map by José Kalpers) diverse avifaunas of the African continent, with over 500 bird species recorded (Kanyamibwa 1998; Lamprey 2002). The Nyungwe National Park (NNP, Ha) is one of the most biologically important montane rainforests in central Africa, with more than 260 species of trees and shrubs, 260 species of birds (Dowsett 1990), 100 species of orchid (Kanyamibwa 1998) and 13 species of primates. The area forms the watershed for 70% of Rwanda s streams and feeds both the Congo and Nile basins. The Volcanoes National Park (VNP, ha) hosts 245 plant species, 180 bird species and 115 mammal species, including the mountain gorilla Gorilla beringei beringei. The Nyungwe and Volcanoes parks form part of the Albertine Rift biodiversity hotspot (Myers et al. 2000) and have been ranked as high-priority conservation sites, both for their species richness and for endemic and IUCN listed threatened species; of 38 surveyed sites in the Rift, Volcanoes is ranked first and Nyungwe third (Plumptre et al. 2003). Management of conservation has not been easy in Rwanda. Rwanda has the highest population density in Africa (310 per km 2 ) and a population growth rate of 3.1% per annum (MINISANTE/ ONAPO 2003). Local dependence on natural resources is high with continuing trade-offs with conservation resulting from demands for farmland, fuelwood and bushmeat. Dependence is associated with poverty (Masozera and Alavalapeti 2004) and poverty levels are high. For example, in the provinces adjoining NNP, Gikongoro and Cyangugu, the proportions under the poverty line are 77.2% and 64.3%, respectively (MINICOFIN 2002). Whilst most rural households are landowning, farm sizes are small. For example, in Ruhengeri province (adjacent to VNP) 37.3% of farms are less than 0.2 ha, whilst in Gokongoro (adjacent to Nyungwe) the figure is 59.0% (MINECOFIN 2002b). At the same time, there remains little opportunity for diversification into off-farm sources of income. Urban migration is occurring quite rapidly but the lack of private sector enterprises (World Bank 2005) greatly restricts opportunities. The current status of protected areas represents a 65% reduction in size over the past 40 years

3 Partnerships for protected area conservation in Rwanda 293 (MINAGRI 2003) and there are important species and habitats not covered by the existing protected areas network (De Klerk et al. 2004). A decade of instability has eroded the institutional capacity for managing protected areas, temporarily militarized large areas of parks, forced huge movements of refugees and created demand for land for resettlement. Prior to the 1994 genocide, ha were covered by the protected areas network; this has been reduced to ha (Government of Rwanda 2004). The fate of Rwanda s forest reserves may be even worse, with an estimated 80% lost during the last 40 years (USAID 2004). The management of the remaining protected areas is made difficult by weaknesses in political and institutional frameworks, lack of human resources and funding, and widespread poverty. Akagera National Park has suffered more than others in terms of insecurity, lack of human and financial resources and conflicts of interest between local livelihoods and conservation. From 1990 to 1994 the park was a war zone between the government and former rebels of the Rwanda Patriotic Front. An aerial survey of the park showed that between 1994 and 2002, wildlife declined by 50 80% due to human activities, including cultivation, pastoralism and hunting (Lamprey 2002). In the last few years, additional pressures have been created by the settlement of people and cattle through the Ministry of Lands, Resettlement and Environmental Protection. Much of the decline in wildlife occurred during a period when the allocation of staff to the park was largely symbolic, with fewer than 20 people in the area. The government has had to consider de-gazetting the entire park and, as it is, has maintained only of the ha originally gazetted. Since 1999, support for managing this threatened protected area has come from the German aid agency GTZ, through their project Protection des Resources Naturelles (Lamprey 2002). However, this project faces complex problems and results to date have been mixed. The demand for land and grazing, together with conflicts between wildlife and agriculture, continue to undermine conservation efforts. The situation in Nyungwe is less complex but the park has also suffered from gaps in management. The area was gazetted in 1933 but there was no means to enforce this status until the 1980s, when a number of donors agreed on a management plan and committed funds. The projects initiated under these funds were stopped during the genocide in Only an NGO with limited funds (the Wildlife Conservation Society) and the responsible government agency (Office Rwandais du Tourisme et des Parcs Nationaux ORTPN) continued working. Again, ORTPN could only maintain a symbolic staff of 20 people to cover this large forest. Since this time, ORTPN and WCS have recruited and trained substantial numbers of new staff and management has improved considerably, boosted by its designation as a national park in The VNP, forming part of the Virunga Volcanoes Region, is very different to the other parks. Being host to the charismatic and endangered mountain gorilla has led to considerable attention and support from both government and international conservation organizations. The small size of the park also allows more comprehensive control. In the Virungas Volcanoes Region as a whole, approximately 5% of the 1989 population of Gorilla berengei berengei is known to have died due to military activity in the area. However, populations have grown despite this disturbance, a cause for some optimism about the commitment by staff, the resilience of the management systems in operation and the robustness of attitudes towards gorilla conservation amongst local populations (Kalpers et al. 2003). Why conservation partnerships? An approach based on conservation partnerships seeks to integrate biodiversity conservation and rural development objectives through a form of decentralization that progressively devolves power and responsibility to networks of partners. These networks are envisaged to be cross-sectoral and cross-scale in their composition. In this section, we provide arguments to justify our working assumption that conservation partnerships must play an important role in the management of Rwanda s protected areas. In the subsequent section, we examine this working assumption more critically, drawing on empirical findings. The need to integrate conservation and development objectives The problems faced in protected areas such as Akagera are representative of the difficulties that traditional preservationist approaches to conservation face in areas where society is highly dependent on natural resources. The principle that conservation and development objectives can and should be integrated is, in part, a response to such difficulties because it accepts the inseparability of social, economic and ecological objectives in populated areas, and especially in areas where poverty persists. There is a pragmatic side to this acceptance: the observation that biodiversity conservation will normally fail in the long term where it does not protect or enhance the social, economic and cultural basis of local lives (Child 2004). This pragmatic view is augmented by a strong moral case based

4 294 Partnerships for protected area conservation in Rwanda on a commitment to poverty alleviation; measured against this commitment, the outcomes of fortress conservation approaches often appear perverse. Throughout sub-saharan Africa, the principle that conservation and development objectives can and should be integrated has provided the theoretical backdrop for a variety of forms of community conservation in which local livelihoods are integral to conservation management (Adams 2001; Hulme and Murphree 2001; McNeely and Scherr 2003; Scherr et al. 2004; McShane and Wells 2004). This principle requires a shift away from a largely Western tradition of valuing Africa s protected areas in terms of their indirect use and non-use values, such as biodiversity conservation, carbon storage and aesthetics (Child 2004; World Bank 2004). Community conservation allows greater emphasis on local direct use values, recognizing the importance of natural capital to local livelihoods and sympathizing with institutional traditions that are more aligned with wise use than preservation (Berkes 2004). In Rwanda, as elsewhere, the commitment to integrate conservation and development objectives will be heavily dependent on finding ways in which local people can benefit economically and politically from managing relatively intact ecosystems. This may rely on adding value to, and developing markets for, non-timber products, but may increasingly look towards mechanisms for local people to benefit from the values held by international actors. Limitations of community conservation The narrative and practice of community conservation has proved hugely persuasive to a wide range of actors, from international donors to communitybased organizations (CBOs). A survey by White and Martin (2002) found that more than a quarter of forests in developing countries are now under community control. However, community conservation approaches do not appear to offer a panacea for the complex difficulties faced in Rwanda. The track record elsewhere has been mixed, both in terms of meeting conservation objectives and in terms of securing development outcomes. Recent years have witnessed serious reflection over this record of achievement and a critique of community conservation is gathering momentum. Some authors characterize this emergent critique as a backlash against people-friendly conservation that is driven by a coalition of those who advocate the old narrative of fortress conservation (Ribot 2002; Wilhusen et al. 2002). Works by Oates (1995) and Struhsaker (1997) amongst others regret the dilution of conservation objectives with development objectives and are able to cite numerous conservation failures arising from this complex mixing of ambitions. There is also evidence that many park managers believe that a high density of fences and guards may be a more reliable route to conservation success than the participation of local people (Bruner et al. 2001). In Africa, the track record of community conservation is not that encouraging. Community conservation projects have only rarely rendered wildlife conservation economically acceptable to local communities (Emerton 2001). Projects are often excessively complex (Ferraro 2001), costly to implement and sustain (Adams and Infield 2001), and even when local incomes are improved, this only leads to conservation gains under particular sets of circumstances (Wunder 2000; McNealy and Scherr 2003; Salafsky and Wollenberg 2000; Wells et al. 1992; Ferraro 2001). Furthermore, the devolution of decision making can generate or intensify local resource use competition and conflict (Hulme and Murphree 2001), stimulate rent seeking and corruption, and enable resource capture by local elites. There are two main forms of response to the patchy performance of community conservation. The first is the backlash referred to above: that conservation objectives should not be diluted by development ones. In light of previous comments, we clearly do not consider this either pragmatic or morally appropriate in the case of Rwanda s national parks. The second form of response comes from those who continue to believe in the basic theory that biodiversity conservation and rural development can co-evolve. These writers are advocates and practitioners of community involvement in conservation but have also been contributing to the critique of community conservation (e.g. Wells et al. 1999; Salafsky and Wollenberg 2000; Hulme and Murphree 2001; Brown 2003; Scherr et al. 2004; Franks and Blomley 2004; Berkes 2004; Fabricius et al. 2004; Child 2004). We sympathize with this ongoing critique and suggest that the concept of conservation partnerships provides a crude but useful means of synthesizing some of the key lessons that are being learned. Lessons from community conservation experiments The success of the community conservation narrative opened up an enormous space for experimentation. A key lesson learned from this grand scale experiment is that the next generation of developmentoriented conservation management initiatives should focus on developing conservation capacity within networks of partnerships that include a range of stakeholders at different scales. This requires a blurring of the distinction between old and new approaches to conservation as we step beyond a

5 Partnerships for protected area conservation in Rwanda 295 polarized focus on either the state or communities. A network of partnerships is intended to develop over time in order to address the lack of institutional connectedness that has proved a common problem for project-focused community conservation initiatives (e.g. Barrett et al. 2001; Poteete and Ostrom 2002; Berkes 2002; Brown 2003). Lack of connectedness manifests itself in many ways and at many levels. Common problems include conflict between local and non-local values; lack of harmony between different government ministries; state legal institutions that constrain local institutional capacity; lack of high-level political support for local initiatives; lack of fit between local projects and landscape level objectives; and lack of involvement of certain stakeholders. On the other hand, strong linkages between partners potentially creates opportunities for synergies that will improve livelihoods (Borrini-Feyerabend 1997). Many of the (failed) community conservation initiatives of the past 20 years have been premised on the potential for local communities to accrue cash income from enterprises that are in some way linked to conservation, whether it be indirectly linked activities such as agricultural production in buffer zones or more directly linked activities such as ecotourism or payments for maintaining ecosystem services. And yet this potential has seldom turned into the desired reality of developing a comparative advantage in conservation oriented activities (Emerton 2001; Wunder 2000). One of the key reasons for this is likely to be a critical lack of partnerships or, indeed, the wrong partnerships (Scherr et al. 2004). With increasing emphasis on developing the market potential of Wild Resources (Wolmer and Ashley 2003) and ecotourism, and with new funding opportunities for providing environmental services such as watershed protection (Pagiola, Arcenas and Platais 2005) and carbon sequestration (Niesten et al. 2002), involving partners with expertise in market realities is becoming essential. The private sector has already been well integrated into conservation partnerships in, for example, South Africa (Fabricius et al. 2004; Child 2004), but remains marginalized in Rwanda, with the exception of gorilla tourism operations. Networks of partnerships, involving central and local government agencies, communities, NGOs and the private sector may offer the best institutional framework for enabling flexible, adaptive management based on social learning (Wells et al. 1999; Sanjayan et al. 1997; Hulme and Murphree 2001; McNeely and Scherr 2003). Ultimately, such an approach is expected to bring greater resilience to livelihoods as partners develop the capacity to respond in a timely and locally appropriate fashion to emergent threats and opportunities. The political context in Rwanda The need to pursue biodiversity conservation in tandem with efforts to involve the rural poor, to alleviate poverty, and to avoid conflict is in keeping with the broader political priorities faced by Rwanda today. A new constitution was enacted in 2003 and multi-party elections took place in August and September 2003, returning President Kagame and the Rwandan Patriotic Front led coalition in a landslide victory. Rwanda is now seeking a political process, and governance structures, that foster peace. This will involve allowing for wide participation in governance, an emphasis on consensus building and the avoidance of divisive politics. This desire is reflected in the current commitment to processes of governance reform, as highlighted in its programme of decentralization (MINALOC 2000; Government of Rwanda 2004), and including the recently launched Decentralization and Environmental Management Project. The drive towards good governance is encouraged by international donors such as the World Bank and is central to, for example, USAID s program in Rwanda (USAID 2004). This commitment to governance reform is evidenced in recent policy documents that underline the need to combine strong protection of the environment with the further incorporation of local government and communities into environmental decision making (Table 1). The challenge in Rwanda is to move towards implementation of this raft of strategic policies and action plans by generating partnerships amongst a wide range of actors. Overall management responsibility for protected areas rests with the Rwanda Office of Tourism and National Parks (ORTPN) attached to the Ministry of Commerce, Industry and Tourism (MINICOM). At policy level, ORTPN is also supported by the Ministry of Resources and Environmental Protection whilst funding passes through the Ministry of Finance and Economic Planning. ORTPN is supported by a number of international conservation organizations but there are no national NGOs of any significance involved in protected area management. Local government is currently only consulted to deal with problems when they occur, whilst community-based organizations (CBOs) have few responsibilities and are yet to develop the institutional capacity for involvement. Sensitization activities take place around the parks but this is done without a strategic view to widening participation in conservation. At national and field levels, the institutional framework is characterized by a lack of effective coordination, weakness of governmental departments and agencies, overlapping, duplicative and contradictory interventions by government departments

6 296 Partnerships for protected area conservation in Rwanda Table 1 Policy context for conservation in Rwanda Policy document Ministry Key proposals National strategy and action plan for biodiversity conservation (MINITERE 2000) National land policy (MINITERE 2002a) National policy of environment (MINITERE 2002b) National decentralization policy (MINALOC 2000) 2020 Vision (MINECOFIN 2002a) National forestry policy (MINAGRI 2003) National population policy for development (MINISANTE/ONAPO 2003) Ministry of Resources and Environmental Protection Ministry of Resources and Environmental Protection Ministry of Resources and Environmental Protection Ministry of Local Government and Social Affairs Ministry of Finance and Economic Planning Ministry of Agriculture, Livestock and Forestry Ministry of Health/National Office for Family Planning Develop a management plan for each protected area. Involve local people in protected area management. Institutional capacity building. Strengthen partnerships and networks of stakeholders. Regional and international cooperation. Inventory and improved boundaries for protected areas. Legal framework for sustainable use of resources in protected areas. Structures for common management of protected areas. Coordination and dialogue amongst stakeholders. Involvement of local communities and local government. Sensitization of local people to environmental and family planning issues. Promotion of job opportunities. Tourism and environmental protection to be included in district level functions. Sustainable management of natural resources. Participation of local communities in management of resources. Education programme to promote importance of forest conservation. Sensitization programme for environmental protection. System of indicators to monitor local use of environmental resources. and NGOs, and a lack of effective dialogue amongst stakeholders (MINITERE 2000). Furthermore, the existing institutional architecture tends to obstruct the development of incentives for local people to pursue conservation oriented behaviour, making it very difficult to integrate conservation and development through conservation activities that deliver income. What National Park managers say As a preliminary stage of this research, we undertook an evaluation of current protected area management using the Scorecards method developed by The Nature Conservancy (Courrau 1999). A workshop was held in Kigali with 12 managers from the three national parks. Participants evaluated the performance of each park in respect to 43 management factors across five categories: social, administrative, planning, natural resource management, and political legal financial. We do not present the full results of this survey in this paper, but it is useful to summarize some key findings here as they add weight to the view that the involvement of stakeholders as partners is relatively undeveloped, and identified as a priority. The most satisfactory elements of management were considered to be: the stability and satisfaction of park staff; the legal status of the parks; and clearly defined park boundaries. The least satisfactory elements were: lack of participation of interest groups; the lack of schemes to enable interest groups to benefit from parks; the negative impact of parks on local communities (especially crop-raiding); and lack of environmental education. Similar methodologies have been applied in evaluations in the Congo Basin in Africa, in Brazil and in Costa Rica, with broadly similar findings (Dudley et al. 1999; Cifuentes and Izurieta 1999). Research aims and method Rwanda is in the process of a planned restructuring of conservation management. So far, this paper has

7 Partnerships for protected area conservation in Rwanda 297 presented a case that supports the working assumption that it makes sense for this restructuring to include a strong commitment to the development of multi-stakeholder, cross-scale conservation partnerships. The aim of the main phase of our field research was to explore this assumption more critically in terms of the potential for realizing effective partnerships within the context of Rwanda. We wanted to know what potential partners perceived their own capabilities to be, how they envisaged their involvement as partners, and what benefits they would expect to gain in return for investing their efforts in conservation-oriented activities. A total of 217 individuals were consulted through 48 group interviews during April June Interviews were conducted by the first author, a native Rwandan, using both French and Kinyarwanda. Participants came from a range of backgrounds: senior staff in government ministries, academics, NGO leaders, tour operators, hotel managers, local government, CBOs, local business people, teachers, religious leaders and indigenous Batwa ( pygmy ) people. The selection of participants was purposive, with selection criteria based on the aim to generate interview groups that represented a range of different stakeholders for each of the parks. Groups were therefore constituted and categorized in two ways. Firstly, five categories of groups were identified according to professional status; secondly, these groups were further categorized according to location (see Table 2). When selecting individuals for inclusion in groups such as local traders Bisate, the total population from which representation is drawn is not well known and the lack of a sample frame precludes random sampling. Partly for this reason, and partly because the numbers of possible participants were generally low, sampling was based on an attempt to maximize representation from each stakeholder group. Relevant individuals were identified (e.g. through consultation with key informants and through snowballing), and those who were available and willing became part of the sample. Whilst this approach involves a certain amount of self-selection and is therefore prone to sampling bias, this problem is minimized by the size of the sample. For example, when sampling the primary teachers in a particular location, the total population is very small and the resultant sample size (the smallest of the 47 groups had three participants) amounts to a large proportion of that total. Each group was asked to discuss a series of questions. Following discussion, a set of options was presented and the group was asked to select the one or more of these that best described their conclusions. So, for example, after discussing a question about the constraints on participating in protected area management, groups were then asked to identify the one or more constraints that were relevant to their group, selecting from a list of lack of awareness, lack of interest, lack of skills, and so on. Thus, open-ended questions and discussion were followed up with closed questions, generating qualitative data, but also facilitating some statistical testing for association. Responses to the closed questions were tested for associations between groupings using chi square and Cramer s V tests. The latter gives an indication of the strength of association with values between zero (weak association) and one (Kent 2001). The groups were relatively homogenous in terms of the status of individuals. For example, a group was composed entirely of primary school teachers or entirely of goat keepers (Table 2). This separation of stakeholders into different interview groups had the advantage of minimizing problems arising from power relations within the groups. There are some limitations to this strategy, such as the lack of opportunity to generate deliberation between different stakeholders. However, the ability to establish the independent views of individual groups of stakeholders was the primary objective here 1. Results and discussion Current involvement in protected area management Conservation NGOs are currently involved in a wide range of activities, including provision of expertise, advocacy, funding and even law enforcement. Local administrations are involved in law enforcement and advocacy. National organizations and local teachers and religious leaders see their roles as restricted to advocacy. Most CBOs and local businesses do not see themselves as having current involvement, though a few groups did claim to play a role in advocacy and the provision of expertise. There is a marked distinction between groups views of how their involvement is framed. All focus groups representing conservation NGOs considered that their involvement took place within some form of framework in which roles were defined (although they also wanted to improve these frameworks). On the other hand, no group from any other sector considered their work to be defined in this way. This includes government ministries within the national organizations category of groups. Constraints on involvement The most commonly cited constraints on involvement in managing protected areas were lack of a framework for involvement (65% of groups); lack

8 Government administration Table 2 Interview groups National/international Akagera National Park Nyungwe National Park Volcanoes National Park 1. Ministry of Commerce, Industry and Tourism 2. Ministry of Resources and Protection of Environment 3. Ministry of Agriculture, Livestock and Forestry Research 4. National University of Rwanda 5. Research Institute of Sciences and Technology NGOs 6. REASON (national environmental NGO) 25. Projet Resource Naturelles (PRORENA/GTZ) 7. African Indigenous and Minority People s Organization 8. Association of Rwandan Ecologists 9. Association for the Conservation of Nature and the Environment in Rwanda Business and CBOs Teachers/ religious leaders 15. Cyarubare District 13. Gatare District 12. Kinigi District 16. Rukara District 14. Mudasomwa District 18. Bisate Sector 17. Gabiro District 20. Bweyeye Sector 19. Rwinzovu Sector 21. Cyabajwa Sector 23. Army 22. Gishanda Sector 28. Wildlife Conservation Society 24. International Gorilla Conservation Programme 26. and 27. DFGF-Karisoke Research Centre 10. Hotel Ninzi 39. Local traders, Nyankora 32. Twizamure-Handcraft, 29. Bee keepers, Mutobo Banda 11. Primate Safari and Swift Tours 40. Local traders, Cyarubare 33. Goat keepers, Bweyeye 30. Local African Indigenous & Minority People s Organization 37. Local traders, Gisakura 31. Abarwanashyaka 38. Local traders, Banda 34. Gorilla Nest; Muhabara; Guest House, Kinigi 35. Local traders, Bisate 36. Local traders, Nyagahinga 45. Primary teachers, Gishanda 43. Primary teachers, Gisakura 41. Primary teachers, Bisate 46. Primary teachers, Kiburara 44. Primary teachers, Kiyabo-Bweyeye 42. Primary teachers, Kabwende 47. and 48. Religious leaders 298 Partnerships for protected area conservation in Rwanda

9 Partnerships for protected area conservation in Rwanda 299 of awareness (50%); and lack of interest owing to the costs of participation. These reasons dominated the responses of those groups who also expressed the desire to be more involved than they currently are (local administrations, CBOs, local businesses, teachers, religious leaders). National organizations were unanimous in citing the lack of an adequate framework for participation, whilst conservation NGOs identified this with poor coordination from ORTPN, the agency in charge of National Parks. Preferred involvement in protected area management Ninety-two percent of groups interviewed want to be involved as partners in the planning of protected area management, reflecting the aspirations raised by Rwanda s ongoing programme of decentralization. Slightly fewer (88%) want to be involved in the implementation of plans and respondents were more specific about how they could be involved in implementation. For some groups, preferred roles closely tied with existing roles. This is the case for groups within the national organization, local administration and conservation NGO categories. However, there is an untapped desire for involvement from CBOs, local businesses, teachers and religious leaders who would prefer to have a greater involvement in advocacy and sensitization and, to a lesser extent, information sharing. Only three out of eight groups of teachers and religious leaders concluded that they would want to play a role in information sharing. During the discussions underpinning these decisions, they associated information sharing with informing protected area managers about illegal use of resources. Such a perception is common to communities living in and around protected areas in developing countries where requests for local knowledge have often been restricted to this type of information. Understandably, many teachers and religious leaders would not want to undermine the respect and moral authority that they command within their communities. Whilst this response might seem predictable, it is nevertheless important. There is very strong evidence that rule enforcement is an essential requirement for successful local resource management institutions (Gibson et al. 2005) and too often there is an assumption that forms of participatory management will result in local engagement in rule enforcement. There are some differences between groups in different locations. In particular, the desire to share information is associated with location (χ 2 = 14.8, n = 48, df = 3, p = 0.002, V = 0.556), as is provision of expertise (χ 2 = 13.9, n = 48, df = 3, p = 0.003, V = 0.539). Most notably, more groups operating in the location of the VNP see themselves as having a role to play in the provision of expertise, an outcome that partly reflects the fact that the sample contained more environmental NGOs in this location. One of the most challenging findings, in terms of how Rwanda can move towards developing conservation partnerships, is that, with the exception of conservation NGOs, groups operating at local levels do not consider themselves to have expertise to contribute to protected area management. Of the 32 focus groups covering the categories of local administration, CBOs, local businesses, teachers and religious leaders, only one group expressed a capacity to contribute in this way. This group of indigenous Batwa people (group 30, Table 2) have a high level of knowledge about traditional medicine and this is the expertise that they argued they could provide. This is a rather surprising finding and provides the first evidence that national parks are quite alien to local people, including local elites. In addition to questions about types of role, we also asked groups to discuss and define their preferred level of involvement in terms of whether they wanted to be informed, consulted or involved in decision making (Figures 2 and 3). The findings show significantly different preferences across the professional status categories of groups (χ 2 = 25.7, n = 48, df = 8, p = 0.001, V = 0.52), as well as across locational categories of groups (χ 2 = 20.5, n = 48, df = 6, p = 0.002, V = 0.26). Only 10 of the 48 groups expressed a preference to be involved at a decision-making level, with half of these being conservation NGOs. Given that all conservation NGOs considered themselves as potential experts, it is perhaps not surprising that they also see themselves as decision makers. The locational difference suggests a contrast between groups located in the three National Parks, with more groups in the Volcanoes locale wanting involvement at the decision-making level. Again, this can partly be explained by the sample, as there are more environmental NGOs operating in this location. Various authors have identified a spectrum of participation with relatively weak forms of participation characterized by stakeholders merely being informed or consulted, as opposed to stronger forms of participation in which actors are empowered to take decisions and to embrace adaptive forms of self-management (e.g. Freire 1975; Oakley 1991). With the exception of the international conservation organizations, all groups operating at a local level consider that they would best be involved at a weak level of participation, being kept informed about decisions or being consulted by formal agencies as part of a process in which responsibility for decision making remained with an external agency. Unfortunately, our data do not

10 300 Partnerships for protected area conservation in Rwanda Figure 2 Proportion of respondents wishing to engage in different levels of involvement in protected area management in relation to professional status Figure 3 Number of respondents wishing to engage in different levels of involvement in protected area management in relation to location sites enable us to gauge the reasons behind these preferences with any confidence. We know that these groups believe that they are poorly informed, lack relevant expertise and also lack financial motivation. We also know that they consider involvement with conservation management to entail costs and we know that previous regimes have generated few benefits for local people. We can also begin to feel

11 Partnerships for protected area conservation in Rwanda 301 more confident that the specific objectives and regimes associated with national parks are somewhat alien to all but the most obvious advocates: the environmental NGOs and ORTPN. National parks, as constituted, may therefore prove a difficult platform for the development of strong partnerships. We do not know whether different regime scenarios would generate different expectations about levels of participation. It may be, for example, that if a group was asked to discuss involvement under hypothetical conditions of community tenure, they would seek deeper levels of participation, feel they had greater expertise and be more willing to share information on rule breaking. What we should take from this discussion is that partnerships, as we tend to think of them (i.e. equal partnerships founded on principles of joint decision making), cannot simply be imposed as a blanket strategy in Rwanda. The narrative of community conservation can lead planners to assume that all communities are ready and willing to share responsibility. This poses a tricky dilemma for Rwandan authorities. On the one hand, the views expressed to us by stakeholders suggest the need for a slow development of partnerships in which the capacities of some participants are progressively enhanced and the level of participation is only deepened if and when these partners feel able and inclined to seize the opportunity. On the other hand, such a gradual approach to growing partnerships leaves the door open to the criticism of tokenism and may prove out of step with the community conservation narrative permeating international donors. However, we would argue that it is more progressive to take a small step with which all partners are comfortable, than to impose a large step which will quickly lead to a sense of failure. Decentralization through partnerships requires root and branch changes in institutional capabilities, cultures and working practice. We need to be realistic about what can be achieved where, and about what constitutes a sensible sequencing for institutional restructuring. What kind of framework is wanted? The lack of a framework for enabling and coordinating conservation partnerships is the most commonly expressed reason for lack of involvement. However, when asked to discuss and express an opinion on the nature of framework that is desired, highly differentiated responses were given. The association between the preferred participatory framework and professional status was significant (χ 2 = 23.3, n = 48, df = 6, p = 0.003, V = 0.49), as was the association between preferred participatory framework and location (χ 2 = 24.3, n = 48, df = 6, p = 0.001, V = 0.50). The difference between professional groupings again pointed to a chasm between the objectives of conservation NGOs and other categories of groups. Conservation NGOs unanimously want their involvement to be framed in a formal agreement, whilst local administrations, CBOs, local business people, teachers and religious groups want either informal mechanisms or no framework at all. The main locational difference is between groups operating in the ANP, who want no framework, and groups in the NNP and VNP who do want either a formal or, more often, an informal framework. The views expressed about potential partnership frameworks appear to be contradictory. On the one hand, the majority of respondents cited the lack of a framework for involvement as a constraint; on the other hand, they mostly rejected the idea of a comprehensive formal agreement in favour of either informal guidelines or no framework at all. One of the problems arising here is the difficulty of categorizing responses in order to establish a clear distinction between formal and informal mechanisms. The definition of formal framework that emerged from interviews was of sophisticated protocols for formal agreements enshrined in a Memorandum of Understanding. Informal frameworks were understood as guidelines that fell short of formal agreements. The desire for formal frameworks was largely restricted to conservation NGOs. It is relatively easy to understand this perspective. Firstly, there are no organized national conservation organizations working at protected area level. Therefore, the five group interviews with conservation NGOs were all with international NGOs. These groups require formal agreements with the Government of Rwanda in order to secure their presence. Many have already signed Memorandums of Understanding (MOUs) with the Wildlife Authority and would want to see these amended to reflect a change in working practices and partners. Other groups were similarly influenced by the particular contexts of their work, though in most cases these contexts rendered formal agreements undesirable. One of the reasons for this is the fear that formal partnerships offer the state de facto access to powers that might otherwise be exercised locally. This apparently paradoxical concern about the potential impacts of formalizing participation has been identified by Edmunds and Wollenberg (2003) in their study of the impact of devolution policies on the control of forest resources in Asia. The potential siphoning off of control from local agencies is most easily identified by the case of district governments in Rwanda. Under the National Decentralization Policy (MINALOC 2000), there is a proposal for districts to be mandated responsibility

12 302 Partnerships for protected area conservation in Rwanda for tourism and environmental protection. The intention to implement this decentralization of responsibility was confirmed during interviews with senior staff in the Ministry of Resources and Protection of the Environment, the Ministry of Tourism, Industry and Commerce and the Ministry of Agriculture, Forests and Livestock. This goes some way to explaining why only two of the 12 groups of local administrators in this sample concluded in favour of framing partnerships within agreements that would formally dilute their putative responsibilities. The views of local administrators are coloured by their experience of the unequal exchanges between partners in the past. On the one hand, they report that the ORTPN only approaches local government when there is a problem that they need them to deal with. This is an accurate perception that arises from ORTPN s limited mandate for law enforcement. ORTPN has to rely on local governments to deal with infractions that require judicial response. At the same time, ORTPN are seen to reap the financial benefits of protected area management; it is they alone who receive income from tourism operations 2. Thus, there are issues of power and money that make local administrative bodies wary of formalizing protected area management partnerships. It is clear that space for dialogue between these sets of actors needs to be opened up if partnerships are to develop. For local businesses and CBOs, one of the fears of entering formal partnerships relates to concern about the costs of conservation. These groups find it relatively easy to identify some of the economic costs of conservation (such as time commitment) but have little experience of the benefits that may offset these. Such benefits have in most places remained undefined and untested. In Akagera National Park, for example, most respondents argued that they wanted neither formal nor informal agreements to underpin conservation partnerships, despite the fact that they also identified the lack of such agreements as a key obstacle to participation. This park is characterized by greater dependence on illicit use of park resources, for example through pastoral activities that rely on water resources in the park. There is therefore a greater fear of loss associated with conservation. This contrasts somewhat with the experience of the VNP, where there is more of a history of cooperation and experience of income derived from gorilla tourism. The case of NNP is different again due to the importance of tea production to the local economy. A recent study found that communities bordering tea plantations are less dependent on the national park because of the income they receive from this employment (Masozera 2002). There are correspondingly lower costs associated with conservation and a greater willingness to introduce some kind of framework for developing conservation partnerships. Economic incentives for entering conservation partnerships As identified in our preliminary, evaluative work with protected area managers, the negative impacts of parks on local people, together with the lack of initiatives for enabling potential local partners to benefit economically from parks, were two of the key weaknesses identified. The main survey also revealed that many local partners consider their lack of involvement to be at least partly connected to the costs and trade-offs that participation incurs. Group discussions about the types of incentives that were required in order to offset these costs led to the ranking of social benefits highest (covering service provision and infrastructure development), followed by financial benefits (grants, credits), job opportunities, capacity building and access to natural resources. Local administrations, conservation organizations, CBOs and local business people generally felt that financial incentives would enhance positive attitudes towards protected areas. Local administrations favoured infrastructure development due to the potential impact on their constituencies; teachers highlighted the need for schools as some teach classes under trees. People around Akagera prioritized more water holes for their cattle, a reflection of the importance of this activity to their livelihoods and the dry climate. By contrast, people around Nyungwe prioritized job opportunities over service and infrastructure provision. This is not because the area is well provided for; indeed its needs are enormous. However, there are reasons for this difference in priority. Firstly, they do not keep cattle as a major livelihood strategy and they farm relatively poor and heavily exploited lands. Secondly, diversification of livelihoods into non-farm income is further advanced. Around the NNP there are six tea plantations and factories employing around people, with wages totalling almost US$3 million annually (Dusabemariya 2003). A private investor plans to establish new plantations and factories generating a further 5000 jobs. This company aims to be connected to electric power and also to rehabilitate roads leading to major towns such as Cyangugu and Kibuye (Karyabwite Pierre, personal communication). People around the park are aware of these developments. Again, the findings of this survey are that local context is critical, not only in determining how stakeholders envisage potential involvement in partnerships, but also in terms of defining some of the expectations they may bring to these partnerships.