Governance Lessons from Two Sumatran Integrated Conservation and Development Projects

Governance issues are at the heart of successful biodiversity conservation and sustainable development. This article examines two Integrated Conservation and Development Projects (ICDPs) conducted in parks on Sumatra, to better understand the foundations of effective biodiversity conservation programmes. The ICDP centred on a networked and multiscalar approach to governance issues seems to have had a longer-term positive impact on truly protecting biodiversity than the one that focused elsewhere. The ﬁ ndings from this research support the notion that an overarching spotlight on institutions and multilevel governance matters (ranging from spatial planning and policy making to arresting poachers to battling corruption) can help in addressing many conservation and development dilemmas. Grounded in ﬁ eld research, this paper calls for a model of biodiversity conservation based on multilayered, networked governance structures, proper law enforcement, and an emphasis on the development of institutional capacity, especially at the local level. These networks should be nurtured by long-term partnerships between governments, communities, and NGOs. Donors and planners should focus on these key areas in conservation design.


What have we learnt from Integrated Conservation and Development?
There is widespread recognition that conservationists must continue to work closely with a wide variety of actors in the fi eld, particularly communities in and around parks 1 (McShane and Wells 2004;Sunderland et al. 2008;McShane et al. 2011). What remains less clear is exactly how these partnerships should work, and how conservation and social needs should be linked.
This article aims to contribute to the discussion regarding what to emphasise in project design and implementation. Early Integrated Conservation and Development Projects (ICDPs) offer important lessons regarding the reconciliation of tensions between conservation goals and the aspirations of people in and around protected areas (PAs) (Wells and Brandon 1992). The ICDP model has evolved, and the general approach of integrating conservation and development in a project form remains widely used (McShane and Wells 2004;McShane et al. 2011).
It is now widely recognised that there are tradeoffs in the integration of conservation and development, rather than the original optimistic outlook that ICDPs present win-win situations (Brown 2004;Scherl et al. 2004;McShane et al. 2011;Salafsky 2011). One key to balancing these tradeoffs lies in the creation of inclusive, adaptive, and sustainable governance structures that build partnerships, empower people, foster dialog, and emphasise the use of the best science available (Scherl et al. 2004). Conservation institutions must be supported by an enforcement of rules by legitimate authority fi gures 2 and a bolstering of institutional capacity (Gibson et al. 2005).
Common pitfalls and problems of early ICDPs have been: •  a focus on project activities rather than biodiversity outcomes; • addressing local symptoms while ignoring macro-level problems and vice-versa; • lack of adaptive management; plans that dictate a timebound project cycle with externally imposed deadlines; • a failure to cede signifi cant decision-making powers to local communities, thereby preventing local ownership of project goals; • acting as if communities are homogeneous entities; • expectations of win-win scenarios and a failure to consider the potential tradeoffs (Wells and Brandon 1992;McShane and Wells 2004;Scherl et al. 2004: 30-31). Each of these problems shows up in the two case studies examined in this paper.

RESEARCH METHODS
The purpose of this study was to gain insights into the reconciliation of nature conservation with the improvement of human well-being through the examination of how selected ICDPs (see Table 1) were planned and implemented, and what their results have been, about a decade after the projects began. Four ICDP sites on Sumatra and Borneo were chosen as case studies (in addition to the two cases discussed in this article, I also studied two parks in East Kalimantan). Indonesia has the largest area of remaining rainforest in Southeast Asia and is rapidly developing. Field research was conducted through open-ended, semi-structured interviews, review of archival records, and direct observation of project sites (post hoc).
The process of implementation of each ICDP was traced and compared with the other case studies to evaluate the relative success of each project and gain insights from this. Project success was assessed based on overall sustainability, and lasting conservation and development impacts approximately a decade after initial implementation (see Table 2). I aimed to discover what did and did not work in ICDP design and implementation, and assess the involvement of international actors (conservation organisations, international organisations, donors, governments, etc.) in the socially just and effective integration of biodiversity conservation and sustainable development in Indonesia and beyond.
As a small-N case study comparison, this research aims to improve the generalisability of fi ndings from case studies by having more than one case, while retaining the ability to go into some qualitative depth by looking at a small number of case studies. In this way, the small-N case study comparison combines the advantages of single-case analysis with those of multi-case analysis, while attempting to avoid the disadvantages of each (Abbott 2004). Lin (1998) observed that a comparative case study offers the opportunity to combine the best features of positivist and interpretivist work, because case studies allow the researcher to see the phenomenon of interest in context, while the inclusion of several case studies forces the investigator to be more rigorous about defi ning relationships, and allows more generalisability.
The primary methods used in this qualitative study were interviews, observations, and archival records review. For each case study, I will briefl y list some of the key informants interviewed. For Kerinci Seblat, I interviewed all the locals I could fi nd who had been involved with the ICDP, which included several people in the head park offi ce (Balai Besar), individuals affi liated with local NGOs (such as LTA and WARSI) and individuals affi liated with global NGOs (such as FFI and WWF). I contacted several experts who had key positions in the design and implementation of the ICDP in Kerinci, including Matt Linkie, Sukianto Lusli, and Raleigh Blouch. I also interviewed key scholars such as Prof. Paul Jepson of Oxford University and Prof. Suraya Afiff of University of Indonesia, both of whom have studied the Kerinci ICDP. When visiting Kerinci, I was based in Sungai Penuh, and took trips to remote villages for interviews.
For Gunung Leuser, I interviewed Mike Griffi ths, one of the co-founders of Leuser International Foundation (LIF) and now a central actor in Badan Pengelolaan Kawasan Ekosistem Leuser (BPKEL). I also interviewed a number of the staff at LIF, former members of the Leuser Management Unit (LMU) including Yarrow Robertson, and staff of the local park offi ce. I was frequently in Medan for interviews, but also took several trips into and around Leuser, both in Aceh and the province of Northern Sumatra. At Manggala Wanabakti, headquarters of the Ministry of Forestry (PHKA) in Jakarta, I interviewed Ibu Listya (former head of KSNP) and Wiratno (former head of GLNP).
This study combines positivist and interpretivist approaches. Positivist aspects of a comparative case study include the observation of the specifi c details of cases-the facts that lead to similarities and differences between them-while the interpretive qualities exist in the examination of how these observed details are connected in each case and the researcher's comparison of the cases themselves (Lin 1998). Essentially it is the researcher analysing the facts and making sense of them, based upon the patterns that emerge. I utilised interpretive research tools such as coding of field notes, memoing, semiotic analyses and narrative analyses of interviews, fi nding story-lines, oppostitions and syllogisms (Feldman et al. 2004).
After examining the case studies through the processes of this research and evaluating and comparing the various projects, the key variables that continued to surface were institutions and governance, and these emerged as the overarching themes of focus, supported by much literature on the subject. The observations I make here apply to all four of my original case

Social and political context of conservation and ICDPs in Indonesia
Both of the parks discussed in this article are on Sumatra, the westernmost island of the Indonesian archipelago. They are the two largest parks on the island, both mountainous and important habitat for the increasingly endangered Sumatran tiger and other rare species. Large contiguous habitats are more likely to have the ability to support a healthy population of high-trophic species, provided that the area contains the right mix of ecosystem types required by those fauna. Most PAs are in mountainous areas, which are more diffi cult to develop for human use, but are usually less biologically rich than lowland areas, although endemism may increase with elevation (Caldecott 1996). Both parks are threatened by illegal logging, hunting, and encroachment, but as Nellemann et al. (2007) show, Gunung Leuser was under greater threat than Kerinci Seblat was during the period of project implementation. In addition to being a high biodiversity country with numerous endemic and endangered species, Indonesia is the fourth most populous country in the world, and is rapidly developing despite the economic setbacks of the Asian fi nancial crisis. Indonesia was also one of the fi rst countries to implement ICDPs. Approximately 20 ICDPs were planned and implemented in Indonesia beginning in the early 1990s (Wells et al. 1999). During this time, effective conservation as well as social justice and human rights in Indonesia were often constrained by the authoritarian military government.
After three decades of rule by Soeharto's autocratic New Order regime , corruption had become deeply engrained in practice, and continues to hinder fair conduct in business and governance in Indonesia (Colfer and Resosudarmo 2002). In 1997, a series of crises destabilised the New Order and led to massive upheaval and reorganisation of Indonesian government and society. Drought, fi res, and famine across Sumatra and Kalimantan combined with the Asian economic crisis to fracture the power of the New Order regime, leading to chaos and transformation.
The fall of Soeharto in 1998 and the ensuing process of rapid democratisation and decentralisation (reformasi) to regional autonomy (otonomi daerah) had both positive and negative consequences for the practice of conservation, and the ICDPs in particular, in Indonesia. Negative consequences include the breakdown of the rule of law that prevailed in the years immediately following the economic crisis and collapse of the Soeharto regime (World Bank 2003).
Multiple Indonesians interviewed for this research noted the chaotic context of the late 1990s and the early reformasi era as major obstacles for governance in general and conservation in particular 3 . One informant described the situation as having been like "taking the lid off a boiling pot" 4 and another described the changes as having been like "a pendulum swinging from strict authoritarianism to anarchic chaos in a very short period of time" 5 . Central government authorities went from having absolute power to having relatively little power, especially in remote areas.
With their new authority to plan development locally, and the political climate of chaos and disorganisation at the national level, district governments (pemerintah Kabupaten) began to plan roads and hand out permits regardless of the federal status of the land. Regional planning and coordination was poor, even with neighboring Kabupaten. This was further exacerbated by the proliferation of Kabupaten, as local elites were quick to realise that Bupati (the chief governors of Kabupaten) were suddenly the most powerful people in the country. Kabupaten rapidly became smaller and more numerous in the decade after decentralisation, as new dividing lines were made, complicating regional planning and coordination.
In the desperate economic situation and disjointed political climate of the late 1990s, many Indonesians saw that there would not likely be any punishment for certain unlawful actions in parks (McCarthy 2006), especially from a central government agency like the Directorate General of Forest Protection and Nature Conservation (Directorat Jenderal Perlindungan Hutan dan Konservasi Alam, PHKA), the subsection of the Ministry of Forestry which manages parks and PAs. Illegal activities in conservation areas increased, including farming, settlements, poaching, and illegal logging, and in many conservation areas, PHKA offi cials felt powerless to take action on these violations for a variety of reasons 6 .
Firstly, it does little good to arrest people who will likely not be processed through the local legal system due to corruption and/or sympathy on the part of local justice offi cials, especially when it would take a signifi cant portion of the park budget to initiate this legal process 7 . Secondly, local PHKA offi cials (park rangers) often have very little incentive to go after law breakers. Not only is it unlikely to further their career (because promotions are usually based on seniority more than merit), but they also have to live in the same community with the families and larger social networks of the offenders (McCarthy 2006). Thirdly, when large numbers of people are breaking laws, the park rangers are out numbered 8 . In several Indonesian parks, there were instances of violence and retaliation against park rangers (Nellemann et al. 2007) for attempting to arrest perpetrators of illegal activities during this time period 9 . Finally, due to the fi nancial crisis, many rangers felt pity for some lawbreakers, who may have been forced to seek new ways to survive 10 .
For all of the above reasons, the implementation of ICDPs was hampered in Indonesia during the late 1990s and early part of this century, which is the time frame during which most of them began and ended. However, the wide variation in ICDP outcomes shows that the turbulent political climate is not an adequate explanation for project failure, as so many were implemented in roughly the same time period and yet had very different results.

State management of Indonesian parks
In the late 1990s, a majority of PAs in Indonesia were facing serious threats, but had minimal management infrastructure (IUCN 1999). These fi ndings are supported by more recent research conducted during IUCN-UNESCO World Heritage (WH) monitoring missions to Sumatra, regarding the status of the Tropical Rainforest Heritage of Sumatra (TRHS), which is a World Heritage Site comprising the three largest national parks (Taman Nasional) on Sumatra: Kerinci Seblat, Gunung Leuser, and Bukit Barisan Selatan (Hitchcock and Meyers 2006). The following selection from the WH mission report explains their recommendation for the inscription of the TRHS on the Danger List: The capacity of management to effectively respond to and resolve critical situations has failed to keep pace with the mounting threats due to a range of institutional constraints, including funding constraints; inadequate cooperation and support from local, provincial and central government agencies, including in some cases law enforcement agencies; confusion over the rights of local government within national parks; and bureaucratic procedural constraints and ineffi ciencies. In addition, local communities and local government remain largely uninformed about the importance of and threats to WH property, and are therefore often antagonistic. (Hitchcock and Meyers 2006: 1) The above description of some of the institutional problems facing parks on Sumatra offers a preface to the descriptions of the ICDPs carried out in Kerinci Seblat (Kerinci) and Gunung Leuser (Leuser). In addition, there also exist less visible and more complex "webs of power and interest" that perpetuate illegal activities in PAs; McCarthy (2006: 142) described logging networks involving investors (cukong), "bosses" known as tauke, and corrupt "rogue offi cials" (oknum). The fi eld research I conducted in 2007 and 2008 indicated that all of the problems listed above remain serious despite some improvements.

ICDP context and plan
Located in southwestern Sumatra, the Kerinci Seblat National Park (Kerinci) is one of the largest PAs in Southeast Asia, with an area of approximately 1.4 million ha. The park straddles four Provinces (Jambi, West Sumatra, Bengkulu, and South Sumatra) and thirteen districts (Kabupaten) 11 , making coordination with local governments regarding park management administratively complex.
The park includes a wide range of habitats, supporting rich biodiversity, including 85 of the 199 Sumatran mammal species. As of 2002, 24% of the species-rich lowland forests within the park had already been destroyed (Component A 2002), in addition to the lowland forests surrounding the park that have been lost to logging concessions. Threats to the park include road construction, wildlife poaching, agricultural encroachment, in-migration, illegal logging, mining, collection of non-timber forest products, and issues regarding boundary overlap (Wells et al. 1999;Component A 2002;Hitchcock and Meyers 2006).
The ICDP in Kerinci was planned and fi nanced by the World Bank, implemented and managed largely by various foreign consultants hired by the World Bank, in partnership with WWF Indonesia and the Sumatran Conservation NGO Network Warung Konservasi Indonesia (WARSI). The six year project was originally envisioned as just the introductory period of a much longer programme required to fully protect the park and integrate its proper management with regional development. However, there was never any commitment from the World Bank to a multi-phased project such as might have been achieved through an Adaptable Program Loan, and no commitment from the Government of Indonesia (GoI) to continue funding after the ICDP. It was imagined that the Kerinci ICDP would be a model for reconciling conservation and development throughout Indonesia and Asia (World Bank 2003). Instead, it stands out as a notorious ICDP failure, partly due to its high profi le and large investment from the World Bank, combined with the unsatisfactory results and cancellation (Linkie et al. 2008).
A World Bank sponsored assessment of the Kerinci ICDP conducted between September 1996 and April 1997 (Wells et al. 1999) was pessimistic about its potential for success. This is particularly notable for two reasons: fi rst, the ICDP in Kerinci had only just begun, and second, the World Bank Task Team Leader of the Kerinci ICDP (Asmeen Kahn) was one of the authors of the report. The World Bank's fi nal project evaluation (2003) rated the ICDP in Kerinci as unsatisfactory overall. Most of the comments explaining the unsatisfactory evaluation are related to problems regarding governance, bureaucracy, and lack of institutional capacity and coordination on the part of PHKA and local governments. Although the ICDP contained several elements that aimed to address such problems, the scope of the issues was not likely to be resolved in the six years of project operation (especially the politically tumultuous time period of [1996][1997][1998][1999][2000][2001][2002]. This highlights the need for long-term partnerships and the establishment of governance networks that transcend the time scale of a project. The final ICDP plan consisted of four components: (A) Park Management; (B) Area and Village Development; (C) Integrating Biodiversity in Forest Concession Management; and (D) Monitoring and Evaluation. The overall objective of the project was "… to secure the biodiversity of KSNP and stop further habitat fragmentation… The project design proposes to meet this objective by institutional strengthening in the areas of integrated planning, coordinated implementation and regular monitoring and enforcement at provincial and local levels; building institutional capacity through increased staffi ng and in-service training; and improving livelihoods through improved resource management and services delivery." (World Bank 1996).
The fi rst goal listed here (integrated planning) was not actually emphasised in implementation (see Wells et al. (1999: 21); if it had been, perhaps outcomes would have been better. The second goal listed (building institutional capacity) was relatively successful, while the third was not directly addressed (improving livelihoods through improved resource management and services delivery).

Kerinci ICDP implementation and outcomes
According to the World Bank, as the project planning unfolded, the "emphasis evolved away from bioregional integrated landscape and development planning to a focus on small-scale village development based on the (fl awed) assumption that poverty and the lack of alternative livelihoods were driving deforestation and agricultural encroachment into the park (they might have been a contributing factor, but were not the main one)" (World Bank 2003: 1). This stark analysis by the World Bank of their own project supports the thesis that it would have been more fruitful to focus on larger issues of governance (such as eliminating corruption and large-scale illegal logging), including planning (such as coordination and harmonisation of plans among various levels of government) rather than focusing on village development projects.
Component A appears to have been the most successful part of the Kerinci ICDP, and these activities were all related to formal park governance. The major conservation achievements of the project were part of Component A, such as achieving the formal gazetting of Kerinci as a national park in 2000 after a complex, 11-step process including demarcation on the ground using wooden and concrete markers. Kerinci was the fi rst national park in Indonesia to be legally gazetted. Despite boundary disputes with local communities and companies holding adjacent logging concessions, the park would not have had proper management (pengelolaan) without the ICDP 12 .
Also as a part of component A, a management plan was produced, which continues to be used as the basis for annual KSNP work plans. A baseline survey of forest cover was made, to be used for landscape monitoring, with a goal of identifying 'hot spots' of encroachment. Three park staff were trained in GIS, and they continue to use it for some landscape monitoring, although the system is incomplete and can only be used for printing basic maps for patrol activities (World Bank 2003). Each year, the Park GIS unit buys NASA LandSat data for translation into GIS maps for tracking changes in the forest cover of the park over time and conducting patrolling activities 13 . In addition, the project stimulated the addition of park staff, an increase from 71 to 180, including numerous university graduates (fi rst time graduates had been recruited as fi eld staff). These new staff were trained and taken on study tours to parks in Indonesia and Malaysia as a part of Component A. Also, a new interpretation centre with a library was established at the park headquarters and a park website was developed (World Bank 2003).
Component B-Area and Village Development-originally intended to work with 134 of the 383 14 of villages along the park boundary (the number of villages included in the ICDP was later reduced to 75 15 ). The idea was to provide development assistance to villages surrounding the park, intending to take pressure away from park edges. Agricultural encroachment by local people was perceived as a principal direct threat to the park (Barber et al. 1995). However, a pre-implementation analysis by Barber et al. (1995: 26) notes that efforts "to change government policies on regional economic planning and investment" may be more critical than the ICDP's emphasis on boundary village investments. Similarly, Wells et al. (1999: 22-23) found that logging and road construction posed more serious threats than villages. Component B was planned and implemented by WWF Indonesia and WARSI in concert with World Bank consultants. The strategy used involved a contractual commitment from each village to undertake certain conservation measures (Village Conservation Agreement, or VCA) in exchange for a development grant. The Village Conservation Grants (VCGs) from the World Bank of USD 50,000 per village (given in two installments several years apart to ensure cooperation) were for infrastructure or economic development such as revolving funds or agricultural inputs 16 .
Several locals who were Indonesian collaborators with the project reported that WWF Indonesia and WARSI (coordinating village development projects) were shocked and saddened by the unsatisfactory rating by the World Bank 17 . They told me that many locals working with these organisations had hoped that the project would be extended, and that more villages could be incorporated. From their point of view, things were just getting started when the project abruptly ended. They were especially disappointed that they were left with the responsibility of conveying the disappointing news to all of the villagers with whom they had been working for years on this project. The sentiment was that the World Bank decision-makers should have told the villagers that the project was over and why 18 .
Component B appears to have been the most controversial, complex, and problematic part of the ICDP. Despite confl icting viewpoints about the relative success of the outcomes of Component B, there seems to be agreement that there was very little, if any, connection or linkage between the VCAs and the small development grants (VCGs) given to the villages when it comes to the success or failure of either, even though the main point of the grants was to provide an incentive for conservation-and this was meant to be the main linkage point between conservation and development in the ICDP. This is the classic lack of linkage between conservation and development activities bemoaned by so many reviews of ICDPs (see Wells and Brandon 1992;Wells et al. 1999;Salafsky and Wollenberg 2000;Brown 2004. More recent study results have revealed some interesting patterns regarding this conservation-development project linkage, or lack thereof. Linkie et al. (2008) published a quantitative analysis of the relationship between forest cover changes and inclusion of villages in the Kerinci ICDP. They found that a village's participation in the project had no effect on local deforestation rates compared to non-participating villages. They conclude that strengthening law enforcement and local property rights would be far more effective than the ICDP strategy of providing economic development assistance, which had no effect on forest conservation. However, they do note that the overall rate of deforestation in the Kerinci Seblat region, at less than 1% per year during the ICDP, is lower than the 1.69% per year recorded between 1972 and 2002 in the comparably-sized region containing the Bukit Barisan Selatan National Park, and much lower than the annual deforestation rate of 5.9% recorded in unprotected areas of Sumatra. Within Kerinci itself the deforestation rate was just 0.28% annually, which the authors believe may be related to the accessibility of forests around the park, currently acting as a buffer, but being diminished annually (Linkie et al. 2008). This data suggests that the ICDP was relatively effective at protecting the park from encroachment and logging, although not because of the village development or conservation agreements.
An in-depth qualitative study by Syaf et al. (2008) also found that there was almost no connection between grants received and whether or not the village conservation agreements were respected. However, they also found much more nuanced information related to both the VCAs and the VCGs than was studied or reported by the World Bank. The main lessons to be learned from the study by Syaf et al. (2008) are (1) that the VCAs were surprisingly effective, despite having little connection to the success of the VCG; (2) economic development assistance would need more supervision and consultation for success, particularly for revolving funds, and any connection to conservation success is unclear; (3) traditional rules are relatively effective in supporting conservation when there are no confl icts with the park authorities; (4) while traditional rules may not be suffi cient to protect the park, they are very helpful, along with social capital, good relations with the park, and law enforcement, to create good governance; (5) investments in infrastructure were far more sustainable and much more highly valued by villages than economic development projects (or tree-planting schemes).
These findings support the notion that respect for traditional land management regimes and attempts at adaptive co-management (as well as small infrastructure and human development activities) may be more fruitful than economic development schemes purporting to incentivise conservation efforts. Co-management plans for special and traditional use zones for villages within the park were drafted during the last year of the ICDP, but never signed. Although these plans seem to have been well designed, they were undermined by larger governance problems, especially lack of law enforcement and support from district governments (Blouch 2010).
Component C of the Kerinci ICDP was largely futile. Biological surveys showed that several concessionaires were logging within the park, and that the concessions given to these companies were incredibly species-rich areas that deserved repatriation to the park, but the Directorate General of Forest Management took no action. This evidence could have been used to revoke the concessions and return the remaining highly biodiverse areas to the park, but no action was taken (World Bank 2003), which suggests flagrant corruption within the Ministry of Forestry, and underscores the key role of governance in conservation and the importance of the ability of outside organisations to effectively work with government institutions to reach conservation goals.
Component D, Monitoring and Evaluation, supported landscape monitoring activities, including field transect walks, fi xed-point photography, socio-economic surveys, hydrological studies, data collection on illegal logging activities, and GIS surveys.

Analysis of results: Lessons from Kerinci Seblat
The ICDP in Kerinci is widely regarded as a failure, despite quiet successes discussed by informants 19 and by Linkie et al. (2008) and Syaf et al. (2008). Aspects of the ICDP were successful; particularly those related to park governance in Component A 20 . However, the 'development' part of the project, which became a central feature, was a lot of work for very little conservation payoff. Here, I will highlight three design weaknesses related to planning, assumptions, and focus of the project. First, the ICDP was a short-term project and was also abruptly discontinued. This lack of long-term commitment on the part of the donor organisation was a major destabilising force, combined with the sudden deleveraging, was a main contributor to project failure 21 , as well as economic and social shock in the region. The ICDP was an unsustainable investment, financially and institutionally (Wells et al. 1999: 69).
The project had a very complex design that took quite a long time to organise before implementation began (partly due to the chaotic Indonesian political context during the beginning of the project 22 ). By the time that the project was actually up and running, it was discontinued by the World Bank, causing shock on the part of local teams 23 and undermining the trust of communities that had been engaged by the project 24 . The sudden massive investment and abrupt deleveraging by the World Bank meant that the implementation teams were actually trying to "spend too much money too quickly" in the words of the team leader for Component A (Blouch 2002: 2). This caused its own set of problems, both for project managers and regionally, as people from elsewhere began arriving to the area with the idea that they might be able to obtain some of the benefi ts of the World Bank funding for projects in the region 25 . The context proved to be more complex than the World Bank's short-term commitment could address. The understanding of the local social and political context could have been improved through a thorough institutional analysis, which was not conducted (World Bank 2003).
Second, the ICDP was predicated on the dual inaccurate assumptions that local communities and their encroachment into the park were the main problem facing park conservation 26 , and that short-term village development activities would be suffi cient and effective in curbing this encroachment. Large-scale projects sponsored by government agencies at multiple scales (road building, forest conversion, estate development, etc.) posed a greater threat to biodiversity in the long run (Wells et al. 1999: 69), along with corruption, clientelist networks, and lack of institutional capacity. These assumptions are emblematic of the design of early ICDPs, and point to the need for better coordination with governments.
A third issue was the lack of sustained focus on multilayered governance, especially supporting effective law enforcement activities. Despite signifi cant capacity building of the park offi ce done by Component A, park guards remained incapable of effective law enforcement to control encroachment, which undermined the co-management attempts of the ICDP (Blouch 2010). There was also local resentment that 25% of ICDP resources went to consultants, apparently "to compensate for the absence of local institutional capacity" (Wells et al. 1999: 69).
There was inadequate attention to the creation of a lasting, multiscalar conservation governance communication network aimed at coordinating activities and rooting out corruption. Although the ICDP made promising attempts to coordinate all of the local and regional governments with park authorities regarding harmonisation of conservation and development resulting in a moratorium on road-building into the park, these interprovincial spatial planning coordination meetings with the park authorities did not continue post-project.

ICDP context and plan
Located in the provinces of Aceh and Northern Sumatra, the Gunung Leuser National Park (Leuser) forms the mountainous centre of the larger Leuser Ecosystem. The Leuser Ecosystem constitutes an area of 2.6 million ha, containing a wealth of different forest types, including dipterocarp lowland rainforest. It is the only area on the island of Sumatra where viable populations of all four megafauna species co-exist and remain: the Sumatran orangutan, the Sumatran rhinoceros (the last viable population), the Asian elephant and the Sumatran tiger (MacKinnon et al. 2004). Over 100 mammal species in the ecosystem represent 60% of the Sumatran total (Wells et al. 1999).
The mountainous park itself does not contain suffi cient lowland forests to support the increasingly rare megafauna species and other biodiversity. Therefore, the Leuser Ecosystem, which is a very wide buffer zone of mostly intact rainforest surrounding the park, was established as a part of the ICDP (see Figure 1), to protect the areas of high biodiversity value that surround the park boundary (Robertson 2004).
The ICDP, which was called the Leuser Development Programme (LDP), began in late 1995 and was intended to end in late 2002, but was extended for two years through 2004. The original vision for the LDP was described in the LDP Masterplan, written by the leaders of the Leuser International Foundation (LIF)-an NGO based in Medan and Banda Aceh that was created to conserve the greater ecosystem and work towards its permanent protection (Rijksen and Griffi ths 1995). The fi nal project was designed and planned in partnership with the main funding organisation, the European Commission (EC). The EC committed 31 million EUR and the GoI contributed six million EUR to fi nance the ICDP (Leuser Management Unit 2002).
The keystone of the LDP conservation strategy was the establishment and legal protection of the Leuser Ecosystem (Kawasan Ekosistem Leuser). This was done through the designation of the Leuser Ecosystem as a conservation concession 27 to the LIF. This unprecedented strategy was made possible largely through the high-level political connections of the LIF founders (McCarthy 2006), essentially expanding the national park to cover the areas of rich biological diversity and intact ecosystems, especially lowlands. This innovative institutional design was one of the fi rst examples in Indonesia

Map of Leuser Ecosystem and Gunung Leuser park boundaries
of the move to a landscape-planning approach to conservation planning (Wells et al. 1999) based on ecological data and an integrated, multiscalar approach to governance.
The legal designation of management rights to the locally-based NGO was intended to provide a defence against corruption as well as combating the common conception of state-owned forest as 'empty land' (tanah kosong) open for exploitation. It was also meant to give the LIF the ability to help to protect the area by working with local communities, developing the potential for nature tourism, and conducting ecological research and monitoring in the region. The production forests and wilderness areas designated for nature tourism that surround the edges of the conservation concession act as a buffer zone, creating one of the few examples of a functional buffer zone around a national park, which was particularly rare in Indonesia (Rijksen and Griffi ths 1995).
The team at the centre of the LDP was the Leuser Management Unit (LMU), which was to provide technical expertise and coordinate with stakeholders at multiple scales. These included the LIF, the Leuser Steering Committee (made up of representatives from the key agencies of the national government), regional government agencies, communities, and local NGOs, research institutions, and private organisations. The LMU team was led by foreign experts as well as senior Indonesian staff, headquartered in Medan, with several fi eld offi ces. The LMU was to transfer the project activities to the Leuser International Foundation upon completion, in the name of project sustainability (which it did). The focus areas of the LMU were park management, buffer zone spatial planning and rural development, communications, research and monitoring (Monk 2001).
The ICDP included a component on education and local public awareness regarding the valuable ecosystem services provided by the Leuser Ecosystem to the entire region (Rijksen and Griffi ths 1995: 199). These include the regulation of water supply for over four million people, fl ood and landslide control, fi re prevention, and carbon fi xation. The total value of these services was calculated to be approximately 300 million EUR per year (MacKinnon et al. 2004).
The major threats to the ecosystem during the time of the ICDP were large scale clearance of forests for plantations, the allocation of logging concessions, the poor performance of many concessionaires as well as illegal logging, transmigration and settlement into the last remaining tracts of lowland forests in the Leuser Ecosystem, and planned infrastructure development such as large roads and dams (MacKinnon et al. 2004). Additionally, the Final Report of the ICDP noted the hunting and capture of threatened species such as the tiger, elephant, rhino, and orangutans for medicinal and pet trades. The report particularly emphasised the problem of "rampant illegal logging among local villagers, government offi cials, timber companies, and powerful individuals" who use "threats, bribes, and terror tactics to continue and extend their activities" which "local law enforcement agencies remain powerless to halt" (MacKinnon et al. 2004: 1). The logging networks described here appear to be the same that Dauvergne (1997) and McCarthy (2006) discuss.
In addition to the hidden webs of corruption left from past clientelist regimes, the new, fragmented politics of regional autonomy also shifted the balance of power in the nation. Land use designations determined at the central level were frequently being rejected locally. Nowhere was the rejection of control by Jakarta more visible than in Aceh. Violence between government forces and the independence fi ghters in Aceh (Gerakan Aceh Merdeka) presented an additional complication for the LDP (MacKinnon et al. 2004). Despite these signifi cant threats, the project was able to realise many of its planned goals, including coordinating better protection of the entire conservation concession, and improving the oversight of conservation management and buffer zone development activities (Rijksen and Griffi ths 1995: 164-165).

LDP implementation and outcomes
This ICDP had high-level political support, a strong institutional framework, and well-equipped staff in a centralised headquarters (Wells et al. 1999). Effective project efforts protected the park from two planned roads and various swamp-forest drainage projects, four transmigration schemes, and halted plans for six new logging concessions and fi ve large oil palm plantation permits (Griffi ths et al. 2002) within the Leuser Ecosystem (see Figure 2).
Biological studies were carried out to identify the most important conservation areas of the park, and a corridor was established to connect the biodiverse Singkil Swamp area to the Leuser Ecosystem (MacKinnon et al. 2004). The boundaries of the Leuser Ecosystem were demarcated in the fi eld and incorporated into the spatial plans of most existing (and newly established) regencies (Kecematan) and districts (Kabupaten), as well as at the provincial and national levels (Robertson 2004).
Project managers created an anti-poaching unit of 45 members, which conducted monthly patrols in remote locations to protect rhinoceros, an elephant patrol unit, mobile patrol units that documented illegal logging activities, and a hidden camera programme that provided photographic evidence of the large terrestrial mammals. During the programme extension of two years (2002)(2003)(2004), an airstrip was built in Kutacane and handed over to the government of Southeast Aceh, and the construction of the Leuser Conservation Centre on the campus of University of North Sumatra in Medan was completed and turned over to the LIF (Robertson 2004). The LIF continues to operate this offi ce and another in Banda Aceh, along with fi eld offi ces.
A total of 613 small-scale village development projects were conducted in 11 districts in Aceh and Northern Sumatra between 1995 and 2001, at a total cost of over twenty-two billion IDR. These projects were delivered to villages depending upon their needs and requests, and the ability of the LDP to provide such projects. These projects included a wide range of activities including providing seedlings for nutmeg and mace gardens, animals for raising, canoes or motors for boats, educational activities, family planning clinics, lodges for ecotourism, small grants for startup businesses such as embroidering men's hats or processing grains, building of bridges and irrigation canals, and many other types of projects (Unit Manajemen Leuser 2001a;Unit Manajemen Leuser 2001b).
Like the ICDP in Kerinci, development was designed to be linked to conservation through the establishment of "quid-pro-quo agreements" with village leaders agreeing to cease logging and clearing of forests in the Leuser Ecosystem in exchange for micro-projects (McCarthy 2006). In the end,

Figure 2 Inappropriate plans within Leuser Ecosystem defeated (and addition of the Singkil Swamp Corridor) due to LDP (MacKinnon et al. 2004; Annex 12.5)
the progenitors of the LDP found that there was not much of a relationship between the village development activities and the conservation goals of the project. They concluded that seeing local communities and individuals as the problem was a "misidentifi cation of the threat to conservation" 28 and instead focused on addressing larger governance issues such as state policy, regional spatial planning, public investment decisions, and development coordination (McCarthy 2006), as well as corruption within the institutions meant to protect biodiversity. Following the LDP, the LIF continued to work with communities in and around the Park regarding conservation, but replaced development projects with meetings and consultations 29 .
Much of the strategy of the ICDP was focused on improving conservation governance through coordination among various stakeholders to resolve confl icting plans of different constituencies, including central and provincial governments, universities and schools, the military, the private sector, NGOs, and villagers, as well as donor agencies and international parties (Monk 2001). One major avenue for doing this was through strengthening communications between conservation organisations, and regional planning and development [Downloaded free from http://www.conservationandsociety.org on Friday, January 10, 2014, IP: 129.79.203.216] || Click here to download free Android application this journal agencies (Bappeda) and improving the integration of ecological awareness into development processes (Rijksen and Griffi ths 1995). It was this central focus on governance, communication, and integration (rather than on village development projects) that lead to signifi cant conservation achievements in Leuser. However, the LDP could potentially have been more effective with proper support from government, although this may have further undermined its legitimacy in the eyes of villagers.
Through the seven-year conservation concession from the Minister of Forestry, beginning in 1995, the LIF was responsible for the management of the conservation area 30 . However, this was really only possible through coordination with governmental institutions at multiple scales, since the NGO was not able to conduct law enforcement activities, and the LIF never had suffi cient support from the bureaucracy 31 . This left the LMU in the diffi cult position of being very constrained in its ability to employ authority tools, while witnessing the failure of incentive tools to gain the support of villagers. This failure stemmed both from the inadequacy of the incentives as well as the association of the LDP with the corrupt and locally unpopular state park management bureaucracy. The reliance on state authority for law enforcement undermined the legitimacy of the LDP for many locals (McCarthy 2006: 227).
This weakness in the design of the LDP and its conservation concession model illustrates a larger quandary for conservationists operating in places with authoritarian or dishonest governments. Conservation projects often aim to protect PAs which have been created under colonial or repressive regimes, and may currently be managed by an oppressive government. But if a conservation organisation wishes to protect the biodiversity there, they must work with the existing regime to do so. The original plan for the LDP, as conceived by the LIF, wished to actually replace the state management structures through the concession model. Instead, the LMU and the LIF were forced to co-exist with, and rely on, those institutions, despite their lack of support for the LDP, which undermined both optimal law enforcement levels and LMU relations with villages (McCarthy 2006: 225). However, the LIF continues to thrive and advocate for regional conservation, despite political struggles. They have provided continuity to the LDP goals with new funding from other large donor agencies, continued to recruit interns globally, send units into the fi eld, map the region with a well-funded GIS offi ce, and negotiate with communities and governments.
The GIS offi ce of the LIF (in collaboration with other organisations such as BPKEL) creates maps based on satellite imagery of the Leuser Ecosystem, and calculates the area of forest lost over time (See Figure 3). According to this data, the cumulative percentage lost in Leuser between 1973 and 2010 was approximately 13%. When compared with data for Sumatra as a whole, or even just Aceh (see Pravettoni 2011), the charts show that deforestation rates in Leuser have been less severe than elsewhere in Aceh and the rest of Sumatra. Although the percentage of Aceh deforested between 1985 and 2007 is less than half the Sumatran average (see Pravettoni 2011), it is still roughly 150% of the total for the Leuser Ecosystem within Aceh 32 . While deforestation trends in Leuser have increased over time, they are lower than areas outside of the Leuser ecosystem, even within Aceh, suggesting that protection efforts over the years in Leuser have been relatively effective. However, this does not account for potentially confounding variables, such as the presence of rebels in the forests during part of the time periods studied.
Based on my own experience in the fi eld, the LDP does seem to have facilitated a lasting conservation consciousness in Aceh, which is now upheld by other organisations, and even championed by the government and politicians in Aceh 33 . The idea of Ekosistem Leuser and the PA as the "lungs of the world" was disseminated into the general public 34 . This suggests that environmental education has been a useful tool in informing many communities in the region about the importance of ecological integrity.

Analysis of results: Lessons from Gunung Leuser
I draw three main conclusions about the accomplishments, strengths, and weaknesses of the ICDP in Gunung Leuser. Firstly, this case study shows the potential value of prioritising the creation of multilscalar governance networks for conservation and capacity building. The innovative institutional experimentation of the network implementing the project resulted in a unique and contextually appropriate conservation legacy that has endured beyond the end of the project, despite lack of support from local government agencies. Secondly, these networked actors were able to identify larger threats to conservation than local communities posed, and push for a mixture of authority and incentive-based policy tools (Schneider and Ingram 1990;McCarthy 2006) to address these threats. Thirdly, the project shows the importance of adaptive management and planning for continuity.
The LDP addressed conservation governance by aiming to establish a context better suited to conservation. Bold actions contributing to this included: confronting law enforcement problems and corruption within various levels of government, establishing the Leuser Ecosystem surrounding the park and

Percentage of total forest cover lost within the portion of the Leuser
Ecosystem that is in Aceh integrating its new boundaries into regional and national spatial plans 35 . The design and implementation of the LDP is an example of an early attempt to create a multi-layered conservation governance network within an ICDP context. The LDP aimed to integrate multiple scales of governancefrom the local to the global-and include a diverse network of organisations in policy making to improve transparency, oversight, institutional accountability, and outcomes. It was both locally tailored and globally connected. LDP leaders identifi ed larger threats to conservation than local communities posed, and pushed for a mixture of authority and incentive-based policy tools (Schneider and Ingram 1990) to address these threats. The fi nal project evaluation observed that "many objectives could never be achieved without a higher level of law enforcement than has proved possible", and noted the lack of success of the provision of development assistance to win the support of local communities (MacKinnon et al. 2004: 1). The role of the incentives in the form of development projects was not perceived to be signifi cant to conservation outcomes, although environmental education seems to have been relatively useful.
Interestingly, LDP designers consciously aimed for continuity between preparation and implementation phases, and rejected blueprint planning in favour of an adaptive management approach. The implementation of the ICDP was more 'top-down' than project designers had originally intended. Both the chaotic political context and the complex logging networks linking communities to corrupt offi cials and unscrupulous investors proved to complicate the concept of working with communities more than originally envisioned. The accomplishments of the LDP are all the more impressive given the institutional and contextual obstacles, as many of these are the same problems that plagued the Kerinci ICDP and led to project cancellation there.

Biodiversity conservation governance lessons from ICDPs
The lessons learnt from these case studies led me to focus on the role of governance in biodiversity conservation planning, because it emerged from my research as a key factor. Tsing et al. (2005: 31) have observed that "all conservation programs are necessarily projects in politics and governance." However, not all conservationists have fully come to terms with this, preferring to utilise simpler, more seemingly generalisable terms of incentives or communities. Incentive tools will generally not be adequate to support conservation (Gibson and Marks 1995;Newmark and Hough 2000) or sustainable development-especially when these incentives are aimed only at local villagers, who will maximise net benefi ts in an absence of rule enforcement rather than limiting themselves to a tradeoff (Gibson et al. 2005).
Early ICDPs were frequently based upon the assumption that local communities were the biggest problem facing parks.
While this may be true in some places, it is certainly not true for all parks (Wells et al. 1999). Conservation plans usually do not conduct adequate research on social contexts, which is particularly problematic in the case of ICDPs (Newmark and Hough 2000: 589), since development can only take place within a social context. Attention to the socio-political (as well as the ecological) landscape seems to be foundational to the viability of conservation programs (and sustainable development efforts). Dietz et al. (2003Dietz et al. ( : 1907 have astutely recommended four promising strategies for creating ideal governance conditions: ongoing dialogue between all parties; "complex, redundant and layered institutions; a mix of institutional types;" and adaptive management. I posit that the key element associated with conservation effi cacy is a focus on governance and institutions at multiple scales, particularly law enforcement, capacity building, and long-term partnerships within the context of adaptive, context-specifi c conservation networks. Below I refl ect on these observations. Barrett et al. (2005: 195) observed that "it matters less which rules a community or country adopts than how well they monitor and enforce the rules they set." Whether rules are formal (laws) or informal (community-based), those who enforce them must be perceived as effective and legitimate (Dietz et al. 2003). These (and other) Indonesian case studies illustrate why law enforcement and some other important elements of state governance and rule of law are required for park protection 36 , such as co-operation among government agencies, functioning checks and balances, and judicial oversight. A necessary foundation for effective law enforcement is inter-agency cooperation, such as between local judicial systems and law enforcement agencies. In Indonesia, the pendulum of decentralisation has begun to come back to a point of equilibrium from the chaos of the early reformasi era, but there is still much work to be done to coordinate the central and regional governments, and reign in corruption.

Law enforcement and rule of law
Knowing that belief in panaceas is problematic (Ostrom et al. 2007), Gibson et al. (2005: 273) show that at least one factor-regular monitoring and "rule enforcement-is a necessary condition for successful resource management." These rules need not be imposed on people by governments in a hierarchical fashion. In fact, resource conservation is far more successful when communities have a leading role in setting and enforcing the rules governing resource use (Hayes and Ostrom 2005). This is what the ICDP designers in Kerinci were trying to achieve. However, there are specifi c characteristics shown to be associated with communities that can successfully manage their resources (Ostrom 1999;Dietz et al. 2003), which many of the Sumatran communities around PAs do not currently possess (although it may be possible to foster them).

Institutional capacity building
Conservation of global biodiversity will require strong and resilient institutions. There is an urgent need for institutional strengthening and capacity building for PA management. Agrawal and Gibson (1999) recommend that conservationists focus more on the role of institutions than communities. But the imposition of externally conceived plans and rules by outsiders will be less likely to lead to a sustainable conservation endeavour than local ownership of conservation goals (Hayes and Ostrom 2005;Ostrom and Nagendra 2006), provided that there is adequate rule monitoring and enforcement (Gibson et al. 2005).
This need for institutional capacity building offers a welcome entry point for expertise and fi nancing from global organisations and agencies for international assistance. This could be in the form of educational and experiential training and consulting, or perhaps the provision of supplies or direct funding (such as long-term salary supplements). Strengthening local institutions should involve supporting coordination between NGOs, communities, park offi ces, and various parts of local government (like planning and development agencies), to facilitate landscape planning that prioritises conservation needs while respecting the needs of local people and any effective existing sets of rules used by these communities. Institutional capacity building should be viewed as development, as it is an essential component of sustainable development.
Institution building for successful conservation governance requires time and commitment. While the ICDP in Kerinci did some important and valuable capacity building with the park staff and headquarters, this was partly undermined by the sudden deleveraging of the project. In Gunung Leuser, the capacity of the LIF was strengthened by the LDP despite constraints, such as the short-term usage of foreign experts 37 to manage the project. Most ICDPs aim to leave behind a positive conservation legacy, yet often have not invested enough in building local conservation institutions or improving overall governance, as well as communication and collaboration among key local and regional actors. These projects have often failed to create any real lasting change.

Long-term partnerships
Continuity and commitment are key factors in conservation. The establishment of long-term partnerships are therefore likely to be more fruitful than 'projects' as a model for conservation efforts 38 . Projects are unsustainable, by defi nition. Conservation is a long-term process, not a product that can be delivered by a project, because conservation is never fi nished. It seems short-sighted and even ironic to attempt to create something sustainable with a project format. The normal project cycle (3-5 years) is particularly unsuited to ICDPs (Wells et al. 1999;Newmark and Hough 2000).
Rather than fi nancing projects, donors should focus on nurturing "deliberative global policy networks" (Duffy 2007), building local capacities through global information fl ows and provision of expertise and training, and strengthening ongoing relationships. Commitment to particular places is important because every place is unique, with complex social and historical contexts. Often there are subtle forces at play that are not visible to outsiders. The LIF offers an example of long-term commitment and partnership with local communities, state and regional government institutions, and global organisations 39 .
The message here for donors and NGOs is the importance of long-term commitment rather than short-term projects. There is an urgent need in conservation for ongoing partnerships with communities over the long term. This style of implementation requires adaptive management, which is increasingly being utilised.

Multilayered governance and innovative, tailored, adaptive institutions
A combination of the elements described above would yield long-term global conservation partnerships supporting adaptive co-management and effective law enforcement while building the capacity of locally situated institutions. Multiscalar approaches are more likely to be successful in such an endeavour (Barrett et al. 2005). The design model that emerges from this is a heterogeneous network of agents representing communities, governments, international and local NGOs, multilateral organisations, and academia all working together to negotiate consensus and to reach common goals. The design of LDP offers an example of an early attempt to create this type of network that can coordinate amongst scales and diverse actors to support lasting and accountable local conservation institutions. Truly integrated multiscalar conservation networks would ideally develop innovative, tailored and transparent institutions to combat corruption, coercion and complacency.

Case study discussion
In both case studies, some aspects of the project were more effective than others in reaching project goals. In Kerinci, the ICDP was effective at strengthening the local park offi ce and supporting the establishment of, and planning for, park management. In Leuser, their fresh, innovative approach to park governance was locally rooted but well-connected and adaptive. In neither case did the village development projects contribute meaningfully to conservation. The ICDP that focused on multiscalar governance structures and processes was more effective, while the one that did not place enough emphasis on creating lasting institutions for sustained coordination with the larger context was largely a failure. Table 2 breaks down the criteria for success on which I have evaluated these two case studies, based upon project goals, designs, and outcomes. Nellemann et al. (2007) assessed the threats of illegal logging and poaching to several important parks on Sumatra and Borneo, and found that Leuser faced higher levels of threats than Kerinci. If their assessment was accurate, then higher levels of threats were not a predictor of project failure.
It is also interesting to compare deforestation rates for the two parks. Looking at the period between 1973 and 2010 (see Figure 3), the average annual deforestation rate for the Leuser Ecosystem within Aceh was approximately 0.35%. Although this is slightly higher than the rate of 0.28% reported for Kerinci between 1972 and 2002 by Linkie et al. (2008), several factors must be taken into account for such a comparison. First, the Kerinci fi gures do not include the period between 2002 and 2010. This period of time saw some of the highest rates of deforestation in Leuser according to the data in Figure 3, so it would be important to look at that period as well for Kerinci, even though the ICDP ended in 2002 there. Secondly, the rates for Leuser include the Leuser Ecosystem, which includes the massive buffer zone around the park created and institutionalised by the LDP. In contrast, Linkie et al. (2008) note that the lack of inclusion of the "buffer" of rapidly disappearing forest around Kerinci is likely to be a key reason why deforestation levels in the Kerinci region were so low compared with the rates of deforestation in unprotected areas of Sumatra. Analysis of a slightly longer time frame and larger area bordering Kerinci would likely show higher rates of loss.
The Kerinci ICDP is widely perceived as a high-profi le failure, despite some quiet successes upon closer inspection. The project was relatively weak in all four of the governance areas I emphasise. There should have been a more enduring commitment from the World Bank and more emphasis on networking between governance scales (such as multilayered law enforcement coordination). Both might have contributed to the creation of lasting institutions and bolstered the capacity of these to be more effective. The poor communication between ICDP components exemplifi es the failure of project managers to properly emphasise coordination among parties. It also required better social and anthropological research to truly understand the macro context.
The ICDP in Gunung Leuser stands out as a resilient ICDP, winning against the odds. Despite a turbulent political context and lack of support from local governments and park management agencies, the determined leaders of the LIF and the LDP were able to greatly enhance park governance anyway. Project strengths included their steadfast attention to multiscalar networking and capacity building, as well as enhancement of law enforcement and longer-term planning for conservation of the Leuser Ecosystem.
These two case studies show some challenges of designing and creating ideal institutions to govern PAs, especially in turbulent contexts. The global conservation pendulum swung from a 'fines and fences' style of governance embodied by simple state authority to a model focused on virtuous communities and incentives. Now that we can see the fl aws of both extremes, it is essential to "take a more encompassing view" of the ways in which state and community actors "interact with emergent clientelist socio-legal orders" that shape resource access (McCarthy 2006: 241).
The next challenge is to create hybrid, multiscalar, adaptive institutions that have the capacity to respect people's needs and foster the creation of effective conservation rules, despite the weakness of the existing institutions (including states, communities, and markets) in so many of the developing states of the tropics (Barrett et al. 2001). The role of conservation donors in "the necessary rehabilitation of institutions in tropical countries and of effective coordination among them" is essential, if not cheap or simple (Barett et al. 2001: 497).

Conservation planning in Indonesia
Based on research sponsored by the World Conservation Union, the World Bank, and the UNESCO (IUCN and WCPA 1999;Wells et al. 1999;UNESCO 2007), as well as the fi ndings of this research, the GoI is not currently able to provide the level of protection necessary to the globally important PAs within their jurisdiction. Involvement of international organisations is, and has been, necessary to help address this crisis of governability faced by the GoI. Democratisation, decentralisation, and the emergence of civil society organisations in Indonesia has allowed for the involvement of communities in the offi cial governance and management of PAs-a new chapter in Indonesian history. ICDPs were fi rst implemented on the cusp of this transition, and improvements seem likely to continue, especially with ongoing partnerships.
Poor park law enforcement was a key obstacle in both case studies, and these law enforcement problems continue to hamper biodiversity conservation throughout Indonesia. Political will (kebijakan politik) is a major prerequisite for adequate law enforcement 40 , just as adequate rule enforcement is a major prerequisite for successful conservation efforts. Illustrating this, when President Yuhoyono issued a Presidential Instruction 41 in 2005, which specifi cally directed 18 Ministries in the central government (including the national police forces and national army, the Attorney-General, and the head of national intelligence) 42 to crack down on illegal logging, park law enforcement was markedly improved 43 in many places (UNESCO 2007).
There are also some overall problems with conservation planning in Indonesia 44 . These issues could potentially be addressed by multiscalar governance networks, especially through continuing partnerships (as in the Leuser Ecosystem). First, a recognition that roads are a primary threat to PAs, a strong national commitment to control road-building, and the creation of roadless areas would be very positive steps in the direction of conserving areas 45 . Second, the jurisdictional overlap between Kabupaten and the Park is problematic from both institutional and spatial perspectives, and might be addressed by excising parks from district boundaries.
Third, the PHKA is within the Ministry of Forestry, which controls logging and palm oil concessions as well as forest conservation. These goals are confl icting, and conservation often loses. A powerful new Ministry of Conservation would show that Indonesia is serious about forest protection. Finally, the corruption, clientelism, and bureaucratic inertia leftover from decades of authoritarian rule negatively impact the possibilities for conservation of the Indonesian rainforests. A continued focus on eliminating corruption related to illegal logging networks should be a top priority for the Indonesian government. two anonymous reviewers on an earlier draft of the manuscript that greatly improved the fi nal document.