Bureaucratic Barriers Limit Local Participatory Governance in Protected Areas in Costa Rica

The importance of local participation in biodiversity governance was recently recognised by the International Union for Conservation of Nature (IUCN) through the incorporation of Indigenous Peoples’ and Local Community Conserved Territories and Areas (ICCAs) as a protected area category. This paper explores what barriers ICCAs might face in their successful implementation within already existing protected area systems. I look at this issue in the context of the decentralisation of biodiversity governance in Costa Rica and examine the internal makeup of four different conservation areas within the National System of Conservation Areas. My ﬁ ndings suggest that it is not enough to enact legal reforms allowing and encouraging local participation. Successfully involving local participation requires attention to the class-based relationships within the protected area bureaucracy that create incentives (or not) to link with the local rural citizenry affected by these areas. In three out of four conservation areas, the dominant social class and urban-rural dynamics combined with a lack of accountability mechanisms have discouraged any real rural involvement and empowerment for decision-making. The strategy of the one area that succeeded at sorting these obstacles to incorporate local participation is described in detail.


INTRODUCTION
For decades, tropical developing countries-which are the most biodiverse-have relied on centralised governance approaches, delegating little, if any, decision-making authority to rural inhabitants who are most directly affected by the establishment of protected areas, particularly national parks. As the limitation of centralised approaches to effective natural resources governance became better understood (McCay and Acheson 1987;Berkes 1989;Pinkerton 1989;Ostrom 1990; Baland and Platteau 1996;Hayes 2006;Ostrom and Nagendra 2006;Robbins et al. 2007), local participation gained importance as a factor that could facilitate more effective biodiversity conservation outcomes (Perry and Dixon 1986;Leader-Williams and Albon 1988;Wells et al. 1992;Ludwig et al. 1993;Lowry and Donahue 1994;Western et al. 1994;Robinson 1995;Brandon et al. 1998; Barrett et al. 2001;Brown and Kothari 2002;Berkes 2004;Chapin 2004;Southworth et al. 2006;West and Brockington 2006). Scholars argued that local participation offers better prospects of adequate or long-term use of biodiversity and other natural resources because local resource users 1) have higher stakes in the sustainable use of resources than do the state or distant corporate managers; 2) have more and better information about the intricacies of local ecological processes; and 3) can develop more effective means to manage available resources through local or traditionally accepted practices (Brosius et al. 1998;Brechin et al. 2002). Ribot et al. (2006) also pointed out that when local stakeholders are involved in governance processes, it is more likely that additional accountable and equitable policies will develop.
Since the 1980s, a discourse on decentralisation and local participation emerged among large international conservation organisations, helping to shift away from previous exclusionary approaches for biodiversity conservation (Campbell 2002a). By 2003, the discourse on local participation and decentralisation had gained such momentum that World Bank lending for projects that were community-based was estimated to have risen to USD 2 billion from USD 325 million in 1996 (Mansuri and Rao 2004). Scholars started to critically examine the emergence and effects that diverse discourses related to environment and sustainability have in the implementation of biodiversity conservation projects (Nygren 1998). In general, research has found that despite the dominant discourse on participation and devolution of power, the incorporation of local needs into a biodiversity conservation agenda has been very limited. For instance, Campbell (2002a,b) examined the said benefi ts to local people of the promotion in protected areas of non-consumptive (i.e., ecotourism) and consumptive (i.e., bioprospecting) uses; she found signifi cant problems in their implementation and questioned whether ongoing practices could bring signifi cant benefi ts to locals living around protected areas. Chape et al. (2003) criticised the process of listing and cataloguing of protected area types by the International Union for Conservation of Nature (IUCN), pointing out that while using a local participation discourse, the process forced national and local governments to fi t into internationally defi ned categories, with little participation from those affected by them.
Most recently, the IUCN modified its protected area categories to recognise the relevance and importance of Indigenous Peoples' and Local Community Conserved Territories and Areas (ICCAs), offering a promising form of incorporating local participation into biodiversity governance. In the context of previous research and fi ndings, it is of interest to better understand what are the barriers for ICCAs' ability to emerge as a mechanism of local participation within existing protected area systems. Or are ICCAs going to become part of the ongoing discourse on local participation and decentralisation? The goal of this paper is to contribute to a better understanding of whether current existing systems of local management are going to benefi t from the IUCN's change and how.
There are many dimensions to local participation, including issues related to how the creation of protected areas affects property rights and communities' ability to engage in successful collective action (Baland and Platteau 1996;Hayes 2006;Ostrom and Nagendra 2006), and the nature of the interaction with and the role of international, national, and local non-governmental organisations with protected areas (Sundberg 1998). This paper, however, takes a different angle. It focuses on the incentive structure within the bureaucracy formally empowered to open spaces for the implementation of ICCAs and other forms of local participation in biodiversity conservation governance processes. To do so, I rely on Andersson andvan Laerhoven's (2007: 1089) definition of participatory governance as the presence of institutional arrangements that facilitate the participation of ordinary citizens in the public policy process. Other studies using the same conceptual basis have covered issues like municipal service provision and enforcement (Andersson 2004;Andersson and van Laerhoven 2007); electoral and rural development (Ackerman 2004); participatory budgeting (Baiocchi 2001;Bräutigam 2004) and planning (Costa et al. 1997;Evans 2004); and the relationship between participation, accountability, and democratic local governance (Blair 2000), among others.
I explore these issues in the context of the Costa Rican reforms for the decentralisation of their protected area system. Since 1986, the Costa Rican government vowed to grant decision-making power to local inhabitants surrounding protected areas spanning more than 25% of the country's surface (SINAC-MINAE 1 2002). As part of such discourse, the government created the National Conservation Area System (Sistema Nacional de Areas de Conservación; SINAC). SINAC is integrated by eleven rurally based, so-called 'conservation areas', which were defi ned by a 1990 executive degree as composed of one or several contiguous (or not) protected areas of varying categories (loosely following the IUCN system). SINAC's goal is the conservation of biodiversity and the sustainable production of goods and services derived from the management of natural and cultural resources (MINEREM 1990). SINAC is defined according to its website as "a decentralised and participatory governance system that integrates forestry, wildlife, and national park protected area issues so that policy development, planning, and implementation can lead towards the sustainability of Costa Rican natural resources" (SINAC 2009). Interestingly, the development and operationalisation of local participation in protected area governance processes has been slow to emerge in most conservation areas except one, the Guanacaste Conservation Area (Área de Conservación Guanacaste; ACG). From the perspective of its ability to include local people in decision-making processes, the ACG constitutes a community conserved area (CCA) inside the SINAC. To illustrate what allowed the ACG to emerge as a CCA and what are the barriers in other areas of SINAC, I contrast the institutional arrangements prevalent in most conservation areas of SINAC and those developed by ACG. By doing so, I illustrate that the enacting of legal reforms allowing and encouraging local participation is only the fi rst of several steps into successful implementation. Whether participatory governance processes will emerge inside protected areas with biodiversity conservation objectives, depends on how much attention is paid to the nature of the bureaucratic structure of which local participation must become a part. As we will see, one of the main obstacles SINAC faced to incorporate local participation in decision-making processes was the incentive structure that class-based urban-rural dynamics created within the bureaucratic system, effectively precluding the involvement of rural citizenry.
In the next section, I describe the general research strategy including data collection methods and variable measures operationalisation. In the following section, I provide a brief account of the political, economic, and ecological situations prevalent in Costa Rica before the decentralisation reforms took place and that set the stage for the central Costa Rican government to push for a local participatory agenda for protected area governance. I then present and discuss the incentive structure within the SINAC bureaucracy and how the ACG differed from it, before concluding with some brief policy implications.

Data collection took place in Costa Rica between 2002 and
2006 for a total of 462 days. The fi rst stage of research allowed for identifying key informants and gaining access to a wide variety of archival records (i.e., unpublished reports and private records). The second involved conducting participatory observations, informal interviews, and 34 in-depth interviews of key informants using appropriate ethnographic techniques (Bernard 2006). All interviews were transcribed and archived. Interviews, participant observation, and archival research were conducted at SINAC's central offi ces and at four conservation areas: the Osa Conservation Area (Area de Conservacion Osa; ACOSA), the Tortugero Conservation Area (Área de Conservación Tortuguero; ACTo), the Arenal-Tempisque Conservation Area (Area de Conservación Arenal Tempisque; ACAT), and the previously mentioned ACG (Figure 1). These areas represent more than a third of the eleven conservation areas in the country, and their selection criteria was based on their relevance to biodiversity conservation purposes following the opinion of expert informants: Alvaro Ugalde, founder of the National Park Service; Daniel Janzen, senior tropical biologist in Costa Rica; and the published literature (Janzen 1983;Sanchez-Azofeifa et al. 2002;Bjorndal et al. 2005). Interviewees and key informants included conservation area park guards, programme coordinators and directors, previous and current directors of the national park system, senior advisors to the minister of the environment, and the minister of the environment himself, as well as directors or knowledgeable individuals who worked for non-governmental organisations that formally interacted with the conservation areas. Key informants are only identifi ed by number (C1, C2, C3, etc.) to preserve their anonymity. Only public fi gures cited elsewhere (Wallace 1992;Evans 1999;Allen 2001;Steinberg 2001), or those who have explicitly agreed to have their names recorded and published, are mentioned by name. The time period described in this study spans from the early 1970s to about 1996, unless otherwise stated. The reason is that the National Park Service was created in the 1970s and while the ACG became the fi rst conservation area to be formally established in the early 1990s, the formation of SINAC was not formally enacted into law until 1996. I measured the involvement of local rural citizenry in protected area governance in three ways: 1) through the presence or absence of local advisory conservation area boards where local rural citizenry participated in decision-making processes about protected area issues; 2) whether protected area staff could be held accountable for their actions through hiring and fi ring procedures; and 3) whether conservation areas had created spaces for local rural citizenry participation through direct employment into positions of decision-making responsibility within the protected area staff. This is a relevant measure in this context, given that before the decentralisation reform, there was an explicit policy in place not to hire local people under the argument that doing so could weaken the staff's ability to monitor and enforce locals' hunting and logging activities inside the protected areas.

The institutional setting
The National Park Service (NPS) was first created as a centralised agency within the Ministry of Agriculture and Livestock ( Figure 2). The NPS director dictated policy to national park administrators and their teams of park guards, who were in charge of implementation. In 1979, the National Parks Foundation (Fundacion Parques Nacionales; FPN) was created with the goal of serving as a unifi ed fundraising agency for the Park Service. As a mixed private-public entity, it could receive private donations, and three out of fi ve board members had to come from the central government (Costa Rican Foundations Law # 5338) (Barnard 1982).
Most national parks in Costa Rica were declared in the 1970s and 1980s (Boza 1993;Arguedas and Rodríguez 2003). Despite the grave national economic crisis of the 1980s (Molina and Palmer 2004), the government continued taking land off the market, away from rural inhabitants and withholding compensation. Thus, increasing numbers of landless, unemployed, low-income farmers-many exemployees from banana plantations-saw national parks as an unfair and unproductive governmental land allocation policy. The epitome of such tensions would be symbolised by the well-documented miners' invasion of Corcovado National Park in 1985 (Christen 1994;Evans 1999: 144). The incoming Oscar Arias administration (1986)(1987)(1988)(1989)(1990) vowed to reform natural resources governance in the country. On one hand, national parks spanning 11% of the national territory were under increasing pressure to show that they could be active contributors to regional development (Jones 1992). On the other, Costa Rica had one of the highest national deforestation rates in the continent, at 2.9% annually (Lehmann 1992). Logging was widespread outside of national parks, including on land under protection status such as forestry and wildlife reserves that accounted for 14% of the country (Brockett and Gottfried 2002). Arias responded by consolidating decisionmaking power about protected areas under national park, forestry, and wildlife jurisdiction into a new ministry of the environment. This administration also proposed to develop a new decentralised conservation area system where integrated management of all protected areas under different categories could take place with the participatory governance of rural inhabitants (Durán and Sánchez 1989;Umaña and Brandon 1992: 89).
The creation of SINAC faced significant challenges. Integrating land management under different property-rights regimes, i.e., national parks, forestry reserves, etc., required the design of new institutional arrangements and bringing together personnel who perceived the goals and purpose of natural resources management very differently. Parquistas (parksmen) blamed forestales (foresters) for the deforestation crisis affecting Costa Rica, and forestales viewed parquistas as hindrances to national economic development.
SINAC was now the jurisdiction of the new Ministry of the Environment. The system was composed of 11 administrative units or conservation areas ( Figure 1) under the coordination of a central offi ce and general director based in San José. Each regionally based conservation area director and staff had the mandate to involve local stakeholders on implementation decisions regarding park, forestry, and wildlife management activities related to each of the conservation areas' territorial jurisdiction ( Figure 3). Such a mandate would not be formalised until 1998 in the biodiversity law (No. 7788, Articles 29 and 30), establishing that each of the 11 conservation areas must form a local board of directors.

Barriers at SINAC for local participation
In three of the four conservation areas studied (ACOSA, ACTo, and ACAT), local rural citizenry participation in decision-making processes had not been incorporated through the formation of local boards of directors, to whom the conservation area director could be held accountable 2 . The ACG did succeed in forming a local board and holding its actions accountable at the local level.
Several key informants explained that the reorganisation of the protected area system into SINAC did not allow for any signifi cant transfer of decision-making power to local areasmostly because most of the previously existent class-based centralised bureaucratic structure from the country's capital was reallocated into the rurally based conservation areas and few mechanisms were in place to include locals into their decision-making processes (C1, C6, C15).

The role of class-based, urban-rural, foresters-parksmen relations
Visits to ACOSA, ACAT, and ACTo suggested that their bureaucratic structure mirrored the class-based relationships prevalent inside the central bureaucracy, where rural inhabitants usually occupy the lowest paying positions. In general, my observations indicate that individuals in directorship positions belonged to the upper-middle class and had long-standing connections in central government politics and with members of the upper class. Most were of urban origin and a few had postgraduate degrees. Middle-management personnel included park administrators, department managers, or other personnel with administrative functions. Some of them had university degrees and were generally from urban lower-middle class backgrounds. On rare occasions, middle-level positions were occupied by personnel who had started as park guards and climbed through the ranks by completing a technical secondary education degree while on the job. Most park guards came from rural areas, had formal education up to sixth grade on average, and had salaries comparable to janitorial positions in urban centers. From an urban perspective, park guards constituted part of the working class (Biesanz et al. 1982(Biesanz et al. , 1999. It is important to note that although belonging to a rural background, administrators or park guards did not necessarily belong to the local population surrounding the protected areas, and thus, could rarely represent or voice the interests or needs of the particular local rural population where the conservation area was located. Although nowadays this policy is not followed, by 2006 I had estimated that a minimum of 60% of all employees of the conservation areas still did not belong to the region, and this is consistent with the fi ndings of Arguedas (2002).
Class differences and the disregard of members of the bureaucracy to the importance of class-based effects seem to be an important difference between the ACG and other conservation areas. One ACG informant said that parkguards in other areas felt forgotten while their bosses lived the "good life". He continued: …Is this the new SINAC system?... the idea about SINAC was good, but was poorly implemented… So why [did] it work here [at the ACG]? Because we were very careful, people did not seduce us to buy cars… or land [for ourselves]… I could have made myself a millionaire [buying land for the ACG], but those were not the rules of the game. I did not [learn] that at home, I did not see other colleagues, nor Daniel [Janzen] making himself rich from it… we were very conscious about those sorts of things… (C15).
The internal conformation of SINAC also shifted the locus of power of the new agency from parksmen to forestry personnel. The fi rst director of SINAC was a forester, and all but two of the 11 conservation area directors were foresters as well (García and Ortíz 1991). Key informants mentioned that the allocation of high positions within the new hierarchy, which still depended on civil service-based rules, was based on academic education credentials. As a consequence of having foresters at the top of SINAC, there was a shift of focus to govern forestry-based protected areas, arguably with less biodiversity conservation value, at the cost of working to involve local participation in those areas devoted to biodiversity conservation (C8, C15, C20). One informant captured appropriately the dominant vision among most conservation area directors in SINAC: "… we have protection in place for national parks, we are done, now we need to work in the areas outside [national parks]" (C28). In sum, the conformation of SINAC's bureaucratic structure and the locus of decision-making power, made it very diffi cult for local participation to infl uence SINAC's bureaucratic agenda and practices.

Accountability mechanisms inside SINAC
At ACOSA, ACTo, and ACAT, most hiring and fi ring power was held by the director of SINAC in San José, and local rural citizens had no infl uence on the hiring and fi ring of conservation area personnel and could not hold them accountable for their actions in the region. Formally, hiring and fi ring procedures within SINAC are governed by the statutes of the Costa Rican civil service and the Ministry of the Environment, written in the hiring and job categories handbook (Manual Institutional de Clases) (MINAE 2004), which rewards experience and academic qualifi cations. However, according to interviewed Costa Rican senior policy-makers and other key informants, there is little enforcement on this set of rules-in-form and instead rules-in-use apply: the director of SINAC is usually a political appointee of the Ministry of the Environment and the directors of the conservation areas are political appointees of the director of SINAC. As a political appointee, the director's survival is based on his or her ability to be loyal to the director of SINAC. Once appointed, the director of the conservation area has little autonomy to choose, hire, or fi re his personnel given that most of them have civil service protection obtained from the National Park, Forestry, or Wildlife Services when they existed as separate agencies. The Costa Rican civil service offers career bureaucrats a certain level of benefi ts such as medical insurance, yearly bonuses, and bonuses for working away from home, among other perks. Although fi ring somebody from the civil service is technically possible, in practice it rarely happens. Opening an administrative fi le to fi re a civil servant entails engaging in a long and confl ictive public process. It seems to be customary among middle-and upper-level Costa Rican bureaucrats not to leave a paper trail in a colleague's administrative fi le of any of his/her wrongdoings, complaints, or reprimands by their superiors. As one director put it: "…the government changes every four years and you do not know where you might end [up] next or which of your colleagues might be your boss during the next presidential administration…." (C19).

ACG's approach to develop local participation
The ACG took a different path than the rest of SINAC and has achieved very different outcomes. The ACG engaged in a more transparent land acquisition process that contrasted with the local expropriation practices prevalent elsewhere. At the same time, it found a way to form a local advisory board that held the conservation area director and his staff accountable to local stakeholders. All these changes eventually allowed the incorporation of rural citizens with little formal educational background into decision-making positions. Finally, through administrative and fi nancial autonomy, the ACG was able to shift signifi cant decision-making power from the urban center to the rural locality where it is located. For the most part, the section described below is illustrative of a period in the ACG history encompassing [1985][1986][1987][1988][1989][1990], which is when most of the changes took place at the ACG.

Local demand for participatory governance
The deforestation and protected area crisis of the mid-1980s that prompted a decentralisation discourse by the Arias administration also awakened the need for change among local stakeholders in Santa Rosa National Park in Guanacaste. Daniel H. Janzen and Winnie Hallwachs-US biologists and long-time Guanacaste residents at Santa Rosa National Park-wrote a proposal to the Arias administration offering to operationalise the environmental reform advocated by the President. 3 Santa Rosa National Park (Figure 4) was created to protect seasonal tropical dry forests thought to be the most endangered of all types of tropical forests (Sanchez-Azofeifa et al. 2005). Most of the dry neotropical forests have been converted to pasturelands (Janzen 1986), and burned every year to increase the productivity of a fi re-loving, introduced species of African grass (Hyparrhenia rufa), and fi res have frequently spilled to the forests of Santa Rosa National Park. The Janzen-Hallwachs team proposed to buy and restore degraded pastureland by stopping the fi res and allowing forest regrowth under national park status. At the time, land was for sale due to the economic crisis in the cattle ranching industry (Myers 1981). Janzen-Hallwachs saw in the Arias administration's willingness toward reforming natural resources management the opportunity to put forward their proposal. In their view, national parks within the NPS needed to be able to make decisions at the local level in order to adequately address biodiversity conservation contextspecifi c issues, e.g., anthropogenic fi res.
The restoration plan found strong support among the Costa Rican academic class, and as part of the country's intellectual elite and oligarchy, they used their long-term connections to introduce Janzen to senior political circles. Rodrigo Gámez, a childhood friend of the President, introduced Janzen and Wallachs to Arias with the purpose of briefi ng him on the proposed new initiative. The President offered political support but did not promise any government funding (Allen 2001). Funding would enter via the National Parks Foundation, which would then disburse funds to hire staff, and pay for land purchases, equipment, and infrastructure. At the time, the governing board of the National Parks Foundation was formed by Dr Pedro León, Alvaro Ugalde (Director of the NPS), and Mario Boza (Vice-minister of the Environment). All of them shared roughly the same political agenda to reform biodiversity conservation governance, understood the Janzen-Hallwachs initiative, and trusted Janzen's motives.
In 1987, as a result of lobbying efforts by Costa Rican scientists and key bureaucrats, President Arias provided the Guanacaste initiative with its fi rst strong public political show of support. In a public appearance in Guanacaste, Arias gave zona protectora (protected zone) status to lands that Janzen and Hallwachs' initiative had targeted for purchase. According to Janzen, much of the land was owned by large and small cattle ranching farms willing to sell, and under the new property-rights regime, private ranch owners could continue to work their land but were forbidden to perform any environmentally damaging activities such as hunting, fi shing, logging, or burning. This regulation was aimed at discouraging potential competing land buyers who would develop the land for commercial purposes. 4 As soon as fund-raising made it possible, Janzen and his Costa Rican allies assembled a land-buying team, hiring personnel through the National Parks Foundation, and creating the legal basis for a local board of directors that would eventually govern the ACG. For the fi rst time in the history of the protected area system in Costa Rica, a local board of directors would have input on decision-making processes inside a conservation area.
The local board of directors was formed in 1989, through the convocation of about 19 different national, regional, and local institutions, including the province's government, bankers, central government agencies in the region (planning, agriculture, cattle ranching, education, fi sheries, and public works), regional commerce chambers (cattle ranching, agriculture, tourism, and forestry), local municipalities of Liberia and La Cruz, and local communal development associations of the towns surrounding the ACG: Santa Cecilia, Cuajiniquil, and Dos Ríos. Through this meeting, seven board members were elected. They included the governor of the Guanacaste province, the president of the cattle ranchers association, the president of the regional agricultural centre, a member of a local cultural association, the president of the regional university campus, the president of the La Cruz municipality, and the president of the local communal association of the neighbouring community of Santa Cecilia. The ACG director did not have a vote in the board but was part of it. The board's main responsibilities were to 1) understand the structure and functions of the ACG in order to be able to make responsible decisions about its activities; 2) approve the annual budget; 3) approve the activities report; 4) evaluate the director's performance and assign salary changes; 5) open to public contest the directorship position every fi ve years (in which the previous director could participate); 6) assess the annual operative activities of the trustee of ACG's funds; and 7) generate proposals for changes of uses of ACG funds and activities. During 1989, the board met 16 times (13 ordinary and 3 extraordinary meetings) (MINEREM-ACG 1990).

Creating local trust through land-buying processes
Those most involved in the land purchasing team like Janzen, affi rmed that as much as possible the process was guided by the explicit goal of making the ACG a desirable neighbour to have by the resident local population. The project differentiated from the standard land-purchasing procedure in the creation of national parks: expropriate land and pay for it at an undetermined date and at 'offi cially' determined prices. Instead, landowners had the opportunity to negotiate the selling price of their land and obtain prompt cash payment. At other conservation areas, in 1998, the central government still owed locals about 100 million dollars from land expropriations (Castro and Arias 1998: 5). According to my own estimates, by 2005 about 50% of the expropriated lands had still not been paid for. According to Janzen, the ACG's land-buying process "…did not guarantee owners that they would obtain better prices for their land than if they had been expropriated, but guaranteed that a payment would indeed come, and allowed people to be part of the negotiation process…" The intricate land-buying process involved a gamut of actors from the local to the international level, and it would not have been possible without the decided support of key central government offi cials including the minister of the Treasury among others (see Allen 2001 for details).
Given that some of the large landowners who sold land were among the oldest families in the region (Edelman 1992), people watched carefully. Perhaps a measure of local satisfaction with the land-buying process was that some of them would agree to become part of the local board of directors with governing power over the ACG.
Interviews with those involved in the negotiations indicate that people were willing to sell their land because: "…the cattle [business] was dead and the land had very little value… [The owner] wanted to be living in the San José world of good medicine and good hospitals for his old age…" (C6). Those close to the negotiations, like Janzen and Randall Garcia, affi rmed that sometimes the same piece of land had to be bought twice, once from the rightful owner and then from squatters who were occupying it. Some large landowners (owners of a thousand hectares or more) sold their land as a way to leave a legacy to the region. Other cattle ranchers also wanted to sell, but afraid of being isolated by the ranching community, asked the government to make it look like they had been expropriated. Finally, a small number demanded exorbitant prices for their properties but agreed to negotiate under the possibility of expropriation. A few became confl ictive neighbours, for e.g., by setting off fi res in order to pressure the ACG to buy them off quicker.

Accountability at the local level
The ACG broke with the central government's standard hiring practices when it declared its intentions to hire its fi rst director through an open and public process. This was possible due to the broad support from senior members of the central government and the political and academic elite. At the end of the process there were two fi nalists, one favoured by those closer to the interests of the central government in San José, and the other favoured by the supporters of the ACG and closest to the interests of rural Guanacaste. In the end, the Guanacaste-favoured candidate prevailed on due to Janzen's great leverage as the main fundraiser. The director's contract stipulated that the Guanacaste-favoured candidate was hired, and was thus accountable, to the local board of directors and not to the central government. Effectively, the director was no longer under the control of the NPS, the central government's bureaucracy, or its urban-based oligarchy, but under the control of a rural Guanacastecan oligarchy. Informants noted that under the prior scheme, a park directorship was considered to be a prized position reserved for political appointees, loyal and only accountable to the central government. Appointing a local candidate was not a welcomed stance, and cost Guanacastecans signifi cant support from key central government players (C6, C8, C22).

Creating a new bureaucratic culture
Administratively, the ACG director had the administrative support of two adjunct directors. One was an experienced parkguard, trusted by the NPS director, and in charge of coordinating all NPS personnel deployed at Santa Rosa National Park before the reform began. The other was in charge of coordinating all new personnel hired through the National Parks Foundation, and was a key negotiator of land purchases. In sum, the new emerging institution (i.e., ACG) was populated by two broadly different types of stakeholders (who did not like each other) and governed by different property-rights regimes. The NPS employees had a long history of presence in Santa Rosa, but were unmotivated and underequipped, and saw with jealousy and mistrust the motives of an increasingly large group of new non-governmental employees, buying and managing private lands. Two events were noted as 'key' in the history of the formation of the ACG for bringing together both antagonistic groups. As described by a former parkguard: …There was a large fi re in "Cerro El Hacha" [newly bought land of high archeological value outside national park boundaries]... At that point we did not want anything to do with the [National Parks] Foundation, its people, or their [private] land… We fi nally went against our will… the fi re united us because it was a common challenge to put it off. It became irrelevant who was who, that was the element that decreased the tension between parksmen and the [National Parks Foundation] people, and we started seeing [us] all as one group… [In addition] we were labeled as traitors by the [NP] System… Before, I would go to San José to receive my monthly pay, and I knew I would bump-at the bar or the market-into other parksmen stationed elsewhere…and we would talk. After we started integration with the ACG project we would perceive a negative, cold atmosphere towards us, specially from those at the NP central offi ces…they thought we were expecting to be paid in dollars… there was a campaign organized by the National Parks Union saying "no to privatization!" [of the national park]… The negative atmosphere in San José had the effect of making us more united over here, and we started working more with a regional vision… to the point that we understood that the [new] project could bring benefi ts to [Santa Rosa] national park, and that this was not a crazy selfi sh idea from a gringo or that the intention was to bring more gringos, nor to privatize, but really improve conservation and all… (C8).
Eventually, the newly bought lands were also given national park status and three large interconnected protected areas -Santa Rosa National Park, Guanacaste National Park, and Rincon de la Vieja National Park-came to be included under ACG (158,000 ha; see Figure 4). Altogether, the ACG is considered the largest biodiversity restoration project in the neotropics (Perrow and Davy 2002); all of it is managed by a single team of more than 100 people under a single land tenure regime (the National Parks Law), and as will be described in the next section, with much more active decision-making and participation of the rural surrounding population than at other conservation areas of SINAC. The ACG was able to make the surrounding rural citizenry the main recipient of the direct and indirect employment, educational, and ecological services (Allen 2001;Blanco 2002Blanco , 2004Daily and Ellison 2002). Nationally, the ACG would eventually be viewed as a success for the parksmen's cause because most of its personnel were devoted to biodiversity conservation goals. Internally, the ACG had moved beyond the parksmen-versus-forestry dichotomy that prevailed elsewhere as the ACG had developed its own identity and bureaucratic culture.

Broadening local benefi ts and participation of rural neighbours
Programmes like fi re fi ghting, biodiversity protection, park maintenance, ecotourism, research, biological education, biological monitoring, and land tenure, among others, constituted the main operational units of the ACG. The ACG directors, supported by Janzen's technical advice, placed significant effort in identifying and mentoring suitable personnel-regardless of their academic credentials. When interviewed, one director-Johnny Rosales-spoke passionately, convinced that breaking with long-standing class-based and urban relationships versus rural dominant relationships and patterns was key for local rural involvement and more effective biodiversity conservation outcomes. Sigifredo Marin, also an ex-director like Janzen, also talked about the importance of hiring staff from the rural resident population and largely training them on-the-job. As described by Janzen (2004), the ACG offered honourable and desirable jobs to low-income farmers by on-the-job training by park guards to become teachers, wildfire-fighters, police, and parataxonomists. Interviews with many of these employees showed that most had on average a second-grade formal education, and were trained for specifi c occupations that they could be proud of. While diffi cult to measure, it is likely that the ACG reshaped the local employment landscape for the surrounding rural populace with a low formal education background, given that traditionally such a population could only gain access to seasonal janitorial-level positions at farms and small towns. The ACG provided its employees with job security and a variety of learning opportunities. For instance, the parataxonomists programme trained on-the-job low-income rural farmers in all aspects of how to conduct a biological inventory of Lepidoptera (i.e., butterfl ies and moths). Parataxonomists eventually became responsible for all aspects of the programme. Moreover, the nature of the job demanded that parataxonomists keep a busy schedule yearround, hike the forests, commute between their hometowns and remote rearing stations, and constantly interface in two worlds: rural Guanacaste and the international biodiversity scene with scientists like Janzen. One of the most easily noticeable results of these interactions was that while many of the parataxonomists seemed to have adopted the work ethic of North American academicians (see Janzen et al. 1993), they still stayed deeply rooted in their hometown traditions and families (Basurto 2007). As a result of settling in the same village where some parataxonomists lived, I observed that some of them had gained enough self-confi dence to take on leadership roles in their home communities (i.e., local school board associations or as part of the local governance council), serving as role models for other community members who were exposed to an alternative model of rural lifestyle (Basurto 2007).
ACG administrators were also quick to point out that hiring local rural people with little formal education did not come without the costs of training and learning (for e.g., cars did not last long and a fi re truck was burned in the process of learning to fi ght fi res). However, the ACG was able to withstand such costs because it was explicitly committed to employing local rural people to conduct activities to which they usually had no access, like driving or organising a wildfi re-fi ghting operation (Janzen 2004). Also, through the support of foreign researchers and other international contacts, the ACG had access to fi nancial support with which it could buffer the economic and political costs of supporting members of the lower and rural class. 5 According to the ACG directors, the returns on such investment seemed to pay off well as it created a very loyal employee body, a regional reputation as a supporter of local employment, and a good employer, given that it broke with the normal practice of fi ring or punishing their employees at the fi rst gross mistake they made at work, a signifi cant inhibitor of employees' capability to learn and adapt to on-the-job training.
The ACG policy towards purchases and resources acquisition also emphasised to buy locally as much as possible. As one informant put it: "…we wanted to create a mutually dependent relationship with the localities surrounding the ACG…" (C15). Vehicles and supplies were bought regionally instead of at the country's capital, San José, even though purchasing supplies locally was more expensive than if it was done at an urban centre where resources could be bought in bulk and cheaper. One informant suggested that members of the central government who benefi ted from these practices resented these measures, as kickbacks and the power that came from being in charge of authorising purchases disappeared for them, effectively changing hands from the central to the local level (C22). Another informant pointed out that at a local level, ACG personnel were perceived as privileged, and people wanted to be associated with it (C8). ACG personnel would pay with cash at the gas station, instead of using government coupons. Although on occasion money was lost and those practices eventually had to change, examples like this helped to create a local image that the ACG was a local economic force of which rural, instead of urban, people could play a part as well as reap some of the benefi ts. 6

Shifting decision-making power through administrative and fi nancial autonomy
The ACG's ability to gain administrative autonomy took place when it was able to return to the central government all civil service positions belonging to personnel working at Santa Rosa National Park. This was only possible after Sigifredo Marin-who had become ACG director in 1992 and was viewed as a leader by the rest of the staff-returned his civil service position and became staff of the National Parks Foundation (the trustee of ACG's funds): "…I told the people that it was a very good opportunity… [and] that here would only stay those that worked [hard]… I signed fi rst and everyone followed." All but two staff opted to give up their civil service positions, even though they only enjoyed a slightly higher salary and the benefi t package was not much more attractive than that from the government. For the ACG as a whole, this meant that personnel could now be hired and fi red more easily and be made more accountable for their performance to local-level actors (i.e., the local board and the ACG director).
In exchange for the civil service positions-a highly valuable commodity to the central government-the central government granted the ACG with offi cial fi scal recognition as an autonomous administrative agency within the Ministry of the Environment. As an administratively autonomous entity, the ACG gained signifi cantly more control over its own administrative and budgetary issues than any other conservation area would be able to.
The ACG's fi nancial autonomy resulted from Janzen and Hallwachs' fund-raising capacity. In total, between 1986 and 1989, Janzen and Hallwachs raised about USD 50 million, of which USD 12 million constituted the organisation's endowment. The funds came from more than 1,000 different small and large donors, with the largest amounts donated by the Swedish and Norwegian governments, and helped make the ACG the fi rst endowed, conserved area in the world (Allen 2001). At the time, these funds constituted most of the ACG's operating budget and were administered by the National Parks Foundation. The management of the endowment is a constant source of tension between the Foundation and the ACG. The central government has often tried to meddle in the ACG's local biodiversity conservation agenda through intervening in the Foundation's management of the endowment. 7 It has often been said that the ACG's ability to develop differently from the rest of the conservation areas was due to unprecedented support from the international community on account of Janzen's fund-raising abilities. The evidence does not seem to support these claims. My own interviews and an internal SINAC-MINAE (1996) report of the status of all conservation areas show that 1) in 1996, ten of the eleven conservation areas had seed money to build on their own endowments, but no sustained effort to build on those initial funds took place afterwards; 2) furthermore, other conservation areas have received much higher levels of external funding from the international community than the ACG during the last twenty years, but this funding was not invested in biodiversity conservation activities that increased the participation of locals in decision-making processes (Basurto 2007).

CONCLUSIONS
The Costa Rican case provides useful lessons to other efforts around the world seeking to better understand how ICCAs will be able to become part of already-established conservation area systems. First, the Costa Rican experience shows that it is not enough to enact legal reforms allowing and encouraging local participation. Successfully involving local participation requires attention to the class-based relationships within the protected area bureaucracy that create incentives (or not) to link with the local rural citizenry affected by these areas. At SINAC, the dominant social class and urban-rural dynamics, combined with a lack of accountability mechanisms, have discouraged any real rural involvement and empowerment for decision-making. Second, this case study also illustrates that decided central government support is a necessary but not suffi cient condition for the eventual emergence of local rural citizenry participation in protected area governance. The ACG would not have been able to develop without decisive political support from the central government, nor without the capacity to locally organise and implement change. These two ingredients seem to be lacking in the effort undertaken by SINAC. Third, in contrast to SINAC, the ACG also benefi ted from strong demand for local participatory governance, which came from biologists-non-destructive users of 'biodiversity'-who had a clear stake in the benefi ts that local governance of biodiversity conservation could bring. The history of the ACG makes it clear that the leadership role of biologists Janzen and Hallwachs was fundamental in the development of the ACG, especially at the beginning of the process. However, it would be naïve to conclude that creating participatory governance processes depends on a few key individuals for its success. The more important question is: what kind of leadership do such individuals need to provide to the overall process of local involvement? What seems to have made Janzen and Hallwachs' role particularly important was their ability to develop creative ways in which to broaden the benefi ts of the formation of the ACG to a diverse set of local, national, and international players and at varying scales, so that the ACG eventually could become its own community of stakeholders and a new locally based bureaucratic culture. As one informant put it when Janzen approached him to be part of the ACG initiative: What he [Janzen] proposes to me is a challenge, the challenge of fi nding a new way of doing things. That is what motivated me above all [to work at the ACG]. The challenge of being able to accept more responsibilities than the ones that are normally taken in the realm of public administration… (C23). As a policy system, SINAC is still in its infancy and experimentation with different institutional arrangements is ongoing. In 1998, the new biodiversity law (No. 7788, Articles 29 and 30) established that each of the eleven conservation areas must form a local board of directors to which the conservation area director has to respond. The ACG experience served as an inspiration for the inclusion of this participatory governance process in the law. However, most conservation areas have been slow to develop their own boards due to the lack of internal incentives within SINAC to share power and accountability, as well as the lack of a well-defi ned enforcement mechanism to do so. Despite this, some preliminary interviews conducted after 2006 indicate that congressmen, and to a lesser extent the municipalities, are emerging as stakeholders and have gained some power over the decision-making processes taking place in their respective conservation areas. Whether the type of power that emerging stakeholders are gaining will result in support to the biodiversity conservation or forestry objectives of conservation areas is an entirely different question.