Social forestry in Southeast Asia: Evolving interests, discourses and the many notions of equity

Southeast Asia has long promoted social forestry (SF) in conservation areas, fallow forests, tree plantations, areas in timber concessions and locally managed agro-forest systems, with the engagement of diverse actors and objectives. SF has evolved from early aims of empowerment and devolution of rights advocated by global reform movements, and is now reframed in the market ideal as a win–win–win endeavor for sustainable forest management, climate change mitigation and robust entrepreneurial livelihoods. Southeast Asian states have formulated numerous standardized SF programs and policies that are often linked to broader development goals and priorities, but which have not always been a ‘win’ for local communities in falling short to provide full tenure rights. Civil society organizations that have provided grounded perspectives on environmental justice and rights have also converged with states on entrepreneurship and market-based solutions. Meanwhile, the private sector actor that is seen as key to these solutions is conspicuously absent within the SF policy space. Within this space of diverse and at times contradictory objectives, whose interests do SF policies serve? We examine the social forestry assemblage to investigate the different discourses, interests and agendas in the implementation of SF schemes in Indonesia, Vietnam, and the Malaysian state of Sabah. The formal SF schemes involve shifting or reinforcing old discourses around forest problems and possible solutions, territorialization processes that can lead to inequities in the exclusion of rights, participation and access, and risks exacerbating contestations and inequities in claims to forest land and resources.


Introduction
Southeast Asia is currently pushing forward a revived social forestry 1 agenda. Indonesia has set a target of 12.7 million hectares (mil ha) of forest, approximately 10% of the state forest estate, to be managed under various formal social forestry schemes by 2019, and official statistics suggest that the Indonesian Ministry of Environment and Forestry (MOEF) has allocated permits for 3.4 mil ha as of September 2019. Vietnam set a target of 4 mil ha for social forestry which it met in 2016, four years ahead of its target date (RECOFTC 2017) and of which, most are being managed by commune people's committees (MARD 2017). In Malaysia, Sabah's Forest Policy 2018 emphasizes sustainable forest management (SFM) principles to generate diverse revenue streams for the state and benefits local communities (Sabah Forestry Department 2018), with social wellbeing objectives that aim "to strengthen the participation local communities in forest management activities" and "to encourage Native [people's] participation in forest industries" (ibid, p. 31).
Social forestry (SF) is broadly defined as "initiatives, sciences, policies, institutions, and processes that are intended to increase the role of local people in governing and managing forest resources" (RECOFTC 2014), and includes formalized customary and indigenous governance as well as government-led initiatives. However, the basic tenets of social forestry schemes are the devolution of rights to local communities, support to livelihoods and alleviation of poverty and conservation or sustainable management of forests (Gilmour 2016, RECOFTC 2014. While earlier framings of social forestry policies and initiatives in the region emphasized devolution of tenure rights and local empowerment in forest management and conservation, these have evolved into SF as a model for agrarian transformations, and as a foundation to support neoliberal solutions as reflected in the trend towards community forest entrepreneurship, commercialization of forest products and private sector investments (MacQueen 2013, Moeliono et al. 2015). More recently, social forestry initiatives are being reframed as a solution to support climate change mitigation and ecosystem service provision (Agrawal andAngelsen 2009, Hajjar et al. 2016).
Formal social forestry schemes with their specific objectives and implementation models are being implemented in rural forested regions where demography and livelihoods are rapidly transforming with increased mobility and migration (Kelley et al. 2019;Peluso and Purwanto 2018), increased cultivation of cash crops and rising non-farm incomes (Rigg 2003), changing perceptions and expectations on communal versus individual tenure (Fisher et al. 2018, Moeliono et al. 2016) and changing perceptions of the roles and values of ecosystem services from forests (Pham et al. 2012). Against this backdrop of change, we draw on the notion of assemblage that builds on framings in forestry (Li 2007, Nel 2017, Astuti and McGregor 2017 to tease out the different actors and their interests, and the ways in how the material and discursive shape their interactions in case studies from Indonesia, Vietnam and the Malaysian state of Sabah. We argue that current assemblages of social forestry have created more losers than winners. The dominant interests and materialities that emerge from messy interactions in the social forestry space have been translated as institutionalization of restrictive rules around access and rights, and discourses of market-based approaches that limit local agency and imaginations of sustainable futures. As formal social forestry schemes unfold within the specific social, ecological and political contexts, we examine how discourses in forestrythe framing of problems and their solutionsshift, accompanied by changes in the design of social forestry policies and their notions of equity and empowerment.
The paper is structured as follows: Section 2 provides a historical context to the development of social forestry in Southeast Asia and the theoretical frame on assemblage. After a brief discussion on data and methods in Section 3, Section 4 presents the three case studies highlighting how discursive problematizations of forest management over time in each of the regions leads to changing policies and how social forestry is posited within the set of 'solutions'. Section 5 discusses the SF assemblage and how specific agendas and interests are served by the formal social forestry models in the case studies, and how processes of territorialization affects notions of equity. Section 6 concludes with some perspectives for future SF research in Southeast Asia.

State and forestry in Southeast Asia
Forests is a contested resource and political space in Southeast Asia with many having staked claims on the resource across its history. Forest dwellers and smallholders have long lived on goods and services from forests. The colonial urge to control and produce valuable commodities from forests are still dominant in Southeast Asian states' land and forest use policies (Peluso, Vandergeest and Potter 1995). The Dutch colonial government established teak plantations that cover about 8% of Java island's land area and nearly a quarter of the island is state controlled forest of one kind or another (Peluso 1992). The Dutch also initiated the establishment of the 'Forest Estate" (Bosch Wezen) with enforced boundaries and exclusion policies. State control over forests has now expanded to almost two thirds of Indonesia land, or 126 mil ha, under the control of the Ministry of Forestry in the form of State Forest Estate (Fisher et al. 2018, Siscawati et al. 2017. In its centralized role, the Indonesian state has identified and classified forests and land based on technical, bureaucratic and institutional practices of forestry (Peluso and Vandergeest 2001).
After independence from French colonial powers in 1945 which had established instutions for exploitation of forest resources and development of a plantation economy (Cleary 2005), Vietnam established a socialist form of state forestry with centralized state management and creation of state forestry enterprises for the continued 'rational' exploitation of forest (Dang, Turnhout and Arts 2012). In the Malaysian state of Sabah (formerly British North Borneo), the North Borneo Chartered Company claimed ownership of all land, forests and resources in order to generate revenue for British shareholders during its reign from 1877 to 1946 2 (Doolittle 2007). The colonial legislation was adopted as part of Sabah's independence package and in the formation of Malaysia together with Sarawak and the Federation of Malaya (Lunkapis 2013).
Thus in Indonesia, Vietnam and Sabah, post-independence forest policies were implemented with the same rationale of economic production and little consideration of the diverse social-cultural and economic needs and practices of local, forest-dependent communities. Local forest practices and land uses were depicted in outdated colonial ideologies as "@" backward, lazy and responsible for the past and current state of deforestation and resource degradation" (Doolittle 2007, p. 476). Shifting cultivation practices in particular were considered as a threat to both, forest conservation and production, (Thu et al., in press) to be banned and replaced with more productiveand profitableforms of agriculture (Cleary 2005, Peluso, Vandergeest andPotter 1995). The path dependency of such policies and ideologies has had long term consequences for how social forestry has been conceived, with tangled legacies and persisting contestations over customary rights. In Indonesia an estimated 37-48 million people in over 25,000 villages live within or on the border of Forest Estate lands (MOEF 2018) and have variably claimed customary or individual rights over forests, used forest resources for livelihood and social needs, and converted forests to agriculture. In remote areas where the Vietnamese state has less control, ethnic minority groups circumvent the rules and continue to manage forest resources and practice a range of swidden agriculture systems (McElwee 2011, Thu et al., in press). The colonial government in Sabah has tended to treat local communities and their property regimes as homogenous and rigid institutions (Doolittle 2001), despite studies demonstrating a long history of diverse and nuanced forms of rights to forest products, trees or water (Peluso 1996), a notion repeated in recent implementation of policies on native land titling. Complexities around the dynamism and pluralism of social and political institutions, interests and power relations governing forest access and control (Leach 2002) are typically buried and simplified under layers of bureaucratic governance and expectations for the potential of social forestry.

The social forestry assemblage
The global push towards community-based, and participatory approaches to conservation and sustainable forest management in the early 1990s advocated for devolution of rights to local communities for the responsibility to manage natural resources held as commons (De Jong, 2012, Moeliono et al., 2017. More recently, new global-local and public-private alliances are shaping the negotiation of agendas and interests, with forests being seen as stores of global biodiversity and carbon wealth, sources of ecosystem services or timber value, and as resources to support local livelihoods or cultural concerns. In exploring the interplay between these agendas, the old discourses, in which outdated knowledge is entwined with power and institutional interests, has emerged to bring back into play older attitudes around around rights, shifting cultivation and community institutions (Skutsch and Turnhout 2018).
We put forward an assemblage approach as it allows for an understanding how the diverse interests, agendas and discourses around social forestry can cohere in a particular place. We do so by building on Li's (2007) definition of the social forestry assemblage as "things (trees, logs, non-timber forest products, tools, documents), socially situated subjects (villagers, labourers, entrepreneurs, officials, activists, aid donors, scientists), objectives (profit, pay, livelihoods, control, property, efficiency, sustainability, conservation) and an array of knowledges, discourses, institutions, laws and regulatory regimes" (Li 2007, p. 266). More recently, these objectives have included climate mitigation and adaptation and rural development. The assemblage approach highlights possibilities of emergence, cohesion, shifts, dispersion, and multiplicity (Anderson andMcFarlane 2011, Cons andEilenberg 2019). Seen as sprawling collections of messy interactions between socio-material elements, assemblage is a useful approach for connecting the complexities of the political nature of state-local relations in social forests.
Assemblages are also characterised by processes of territorialisation; and forest boundaries are delineated through colonial and contemporary policies of gazettement for production or conservation, formation of social identity or claims, or arrangement of jurisdictional and administrative borders (Cleary 2005, Peluso andVandergeest 2001). Territorialisation can signify exclusion such as in evictions from contested protected areas under new carbon forestry projects (Nel 2017) or omissions of certain ethnicities in a process of producing homogeneity to which particular community comes to declare itself as a specific Indigenous entity (Astuti and McGregor 2017). Similarly, social forestry with claims to "fix" forestry governance problems or counter exclusions of locals from forests and land, come with formal boundaries and territorialization processes that will inevitably exclude those that cannot perform the type of stewardship required or do not fit within predefined notions of "community".
The social forestry assemblage varies with context but share commonalities in how disparate objectives and problem framing is pulled together in relation to "diagnoses of deficiencies and promises of improvement" (Li 2007, p. 264). The discursive problematization of forestry and forest management is not static but changes over time (Moeliono et al. 2017, De Jong 2012, Peluso, Vandergeest and Potter 1995 and along with these changes, are shifting promises of improvement or proposed solutions. The central proposition of social forestry is that local people living in and close to forests can best manage forests effectively over the long-term. This 'solution' does not however come from a unified specification of a problem. Local people were once considered to be the culprits in driving deforestation and degradation but are now seen to be the solution to effective forest conservation (Gilmour 2016, Peluso, Vandergeest andPotter 1995). They have shifted from being the subjects of sustainable forest management to agents for implementing various sustainability schemes under the umbrella of social forestry (Skutsch and Turnhout 2018).
We lean on assemblage theory to tease out the different actors and their negotiations and adaptations of interests in the design and implementation of social forestry mechanisms. As their seemingly disparate and heterogeneous objectives collide, disperse and realign, we assess how the deficiencies in social forestry are problematized at different points in time, how the solutions proposed for their improvement shift or hold constant, and how the processes of territorialization can have direct equity implications. Our specific research questions are thus twofold: • How do the key actors in the SF assemblage impose, negotiate, contest and/or align their interests in design of SF policies and territorialization processes at particular times? • What are the changing notions of equity related to changing discourses around SF, resultant SF solutions (policies, mechanisms and territorialisation processes)?

Data and Methods
Data for this paper was collected under a project on social forestry and climate change 3 . As part of the project's activities between 2012 and 2019, we've carried out reviews of national policies and published and grey literature related to social forestry, conducted interviews with policy and civil society actors across different levels of governance, and carried out field research in Vietnam, Laos, Indonesia and Malaysia Sabah 4 . From this work, a series of articles related to diverse local forest governance systems and on the intersections of social, economic and political dynamics within forest-agriculture frontier systems in Southeast Asia have been published (see for example Bong et al. 2019, Cole et al. 2019, 2017, Maharani et al. 2019, Moeliono et al. 2017). In addition, the authors have engaged in policy discussions related to social forestry at the national level with the state-led Vietnam Community Forestry Working Group, networks such as Indonesia's Working Group on Social Forestry and the Sabah Social Forestry working group (SASOF), and participation in regional policy events such as the annual ASEAN Working Group on Social Forestry and ASEAN Senior Officials in Forestry meetings from 2012 to 2019.
For this paper, we leveraged on findings from reviews of national policies, interviews conducted with policy and civil society organizations, and our observations in national and regional policy events. We also carried out new analyses of strategic plans and policy materials published by key actors within the social forestry policy arena, and ASEAN policy documents (see Table 1). The key actors were identified through their regular and active engagement in ASEAN and national social forestry events. Within the data corpus, we have identified the actors' interests and discourses through a close reading of supported policies and vision statements and objectives in strategic plans; their rhetoric in how the problem of forests and people are framed in the documents' background and contextual sections; and how this is linked to the proposed actions or solutions for improvement (see Table 2).
We describe the three case studies in the following section, highlighting in particular the problem and solutions identification, actors, interests and agenda setting within the social forestry assemblages in Indonesia, Vietnam and Sabah.

Case studies of social forestry schemes
In all three case studies, social forestry is being implemented within boundaries of conservation areas, forest reserves, village forests, fallows and agroforest systems, areas around tree plantations and timber concessions, which suggests different materialities, politics and exclusions.
Social forestry took hold in Indonesia in the early 1990s in line with the global ideals of community based management being promoted through projects funded by international organizations such as the Ford Foundation (Seymour and Rutherford, 1993 Indonesia did not initially consider social forestry a policy for decentralization of rights but rather, as a "development" program for villages located in forested areas where forests had been used or managed in some form or other (Moeliono et al. 2017). SF was formally included in the new Forestry Law of 1999 and in subsequent policies and laws. Indonesia's conceptualization of social forestry today is legalized in five distinct schemes or permits, each with its specific management regime, permitting requirements, eligible applicants and extent of rights allowable in specific forest zones (Erbaugh 2019, Moeliono et al. 2017. Vietnam moved towards social forestry in the 1990s as forest exploitation led to a rapid decline of forest resources. The implementation of a logging ban to address this problem resulted in a financial crises for most state forest enterprises and a collapse of the centralized forest management. The 1993 Land Law and trade liberalization under the Doi Moi socio-economic reforms promoted 'forest socialization', a Vietnamese term that emphasizes the contributions of societal stakeholders to achieve good forest governance (Dang, Turnhout and Arts 2012). The 2004 Law on Forest Protection and Development legalized the allocation of natural forests to community management and encouraged local people to increase their revenues through community forestry (Pham et al. 2012) despite restrictions on use. The Forestry Law 2017 further acknowledged religious and customary forms of forestry but the Land Law 2013 continues to provide only limited rights to local communities Sabah has carried out intensive, industrial-scale logging since the 1950s, reaching its peak in the early 1980s. During this period, almost 90% of the state's budget was derived from timber revenue. Social forestry was first introduced by the Sabah Forestry Department (SFD) in the early 1980 ′ s to address issues of "illegal" settlement and swidden cultivation in forest reserves (Toh and Grace, 2005) despite the fact that communities would have lived in the area before their designation as forest reserves and had customary rights or claims to the land (Lunkapis 2013). SF projects were initially funded through Community Forestry Cess Fund, established in 1989, through fees collected from timber companies based on the volume of timber exported or processed (Toh and Grace, 2005) and have become a consideration for sustainable forest management licence agreements (SFD Annual Report 2016).

Case study 1: Social forestry and agrarian reform in Indonesia
The presidency of Joko Widodo, which began in 2014, introduced   Table 2 Key actors and their interests, rhetoric of the problem at hand and proposed actions, based on review of social forestry related policy documents. a

Core interests Problematization Proposed actions and solutions
Regional actors: ASEAN "A competitive, inclusive, resilient and sustainable Food, Agriculture and Forestry (FAF) sector integrated with the global economy, based on a single market and production basse contributing to food and nutrition security and prosperity in the ASEAN Community." [1, p.6] Priority for Forestry sector is sustainable forest managment [1, p. 14] with emphases on market mechanisms (certification, value chains) and leveraging on REDD+; PES and climate finance.
Sustainable forest management as a triple win for production of forest goods and services, biodiversity conservation and their optimal use, and climate change mitigation/ adaptation [2] "… in particular, deforestation, largely driven by increasing demand for agricultural land and aggravated by illegal logging. These developments now threaten ASEAN's enviable current position as a region with high levels of food security and a net exporter of food, industrial crops and forestry products. "…to help shape and influence policies that promote more efficient and cost-effective NTFP production. It will tap networking options for communities and for their products and enterprises. Biodiversity-friendly and ethical green products will be promoted for the niche markets and tourism … will tap more strategic market opportunities, but at the same time advocate for ethical domestic, regional and international markets for NTFPs and creative.  It's mission is to motivate the development of CF in this perspective through information dissemination, concept development, capacity building and policy formulation. [8,9] Community forestry is one solution to the problems of tenure and structural poverty to the destruction of forest resources.
Social Forestry has yet to address the fundamental problems faced by the customary and local communities. The perpective that must be changed into a new 'Community Forestry' perspective followed by decentralization and evolution, access to forest resources for customary and local communities, adoption of the forest resource and forest ecosystem management systems. [8] FKKM strategy is to extend community forestry and shared learning process; support conflict mediation; develop the means for information exchange and prmotion of community forestry; and facilitate policy reforms to meet the principals of community forestry [9] Working Group on Social Forestry (WG-SF) The WG-SF is a multi-stakeholder platform, formalized by the 2016 government decree on Social Forestry Acceleration Working Groups [10].
As part of the official SF strategy, every province need to establish a Social Forestry Acceleration Working Group (POKJA PPS) consisting of government, NGOs, academia, and interested parties [10].
The WG-SF supports the government's narrative as it is tasked to raise awareness, and support local governments and local communities in the verification and formal processes required to obtain SF permits.
Mobilizes POKJA PPS at the provincial level, with responsibility to the provincial governors [10].
Enables a transparent verification and monitoring process.
Country actors: Sabah Sabah Forestry Department (SFD) Mission of SFD is to "effectively and efficiently plan and implement the management of the State's forest resources in accordance with the principles of sustainable forest management (SFM)" [12, p.6] Commitment to ensure at least 50% of land is "designated and protected for sustainable forest use and tree cover for environmental protection, biodiversity conservation and socio-economic well-being." "The Forest Policy strengthens the commitment and direction in managing designated forest areas and tree cover through sustainable forest management. This policy takes into account environmental, social and economic sustainability, through good forest governance and best management practices, to ensure forestry remains an integral and competitive land use in Sabah. G.Y. Wong et al. programs to respond to the idea of developing the nation from its margins. These included agrarian reform programs, of which SF is a major component. This reinvention of SF promoted the idea that by providing access to land, business opportunities and vocational training, poor rural farmers and forest communities will be able to escape poverty (MOEF 2018 This newest iteration of the SF policy is distinct. In the 1970s and 1980s, SF was formulated as a 'development' program for communities living in or around forest plantation concessions to resolve land conflicts and social discontent to further the interests of state and private corporations. This changed in the 1990s and 2000s with a series of ministerial regulations and national forest laws that laid out formal SF schemes for the devolution of rights and access aiming to empower communities to manage forests (MOEF 2018, Moeliono et al. 2017 SF is implemented according to the five formal schemes as per the Ministerial Decree P.83. Three schemes are based on management permits issued to village enterprises (Village Forests), communities/groups (Community Forest) and individual or cooperative plantations (Plantation Forest), one scheme regulates co-management of forests in protected areas (Partnership Forest), and only one recognizes full ownership of forests by customary or 'adat' communities (Adat Forest). However, the process to obtain the certificate of adat ownership involves the laborious process of gaining official recognition as 'adat' community by the district government and a territorialization process that can exclude certain groups and their claims (Myers et al. 2017). Further, adat forests comes with the full responsibility to protect the forest but with little government support to do so (Fisher et al. 2018).
Invisibility of local forest governance practices. Although official SF policies responsibilize local communities to manage forests, participation at the community level is often weak or limited to village elites (de Royer et al. 2018). Further, the SF territory comes with state-imposed restrictions on sites, use and access (De Royer et al. 2018, Erbaugh 2019, thus making impossible for diverse local governance practices and informal rights long practiced in many forested sites to be recognized by, or drawn into, formal social forestry structures (Bong et al. 2019). This incompatibility has created conflicts and contestations over use and access (Cummins and Yamaji, 2019). Despite the restrictions, local communities are still inclined to obtain SF permits for achieve some security of access to the forest land and resources they are dependent on (Myers et al. 2017).
Bureaucracies, politics and interests. To obtain a SF permit, communities have to develop and submit a technical proposal documenting required criteria such as proof of a community institution or cooperative to manage the forest, the boundaries of the area to be managed, and an approved management plan (Moeliono et al. 2017). This territorialization process excludes local groups who may not fit, or choose not to participate, in these "neat" community institutions or cooperatives, or those who may demand more from or contest the specific discourses of forest stewardship within the formal SF policy (Myers et al. 2017, Sahide et al. 2020). There is also discontent amongst local (district) agencies who used to play substantive roles in forest governance and now excluded under the new Law 14 on Regional Government (Fisher et al. 2018).
In addition to SF, Indonesia's push for agrarian reform is spearheaded by the policy on Identification of Land for Agrarian Reform policy (Identifikasi Tanah Obyek Reforma Agraria, TORA). TORA involves the certification of 4.5 mil ha of land informally managed by individual farmers, the re-distribution of 4.1 mil ha of state forest lands, and the provision of use rights for 0.4 mil ha of idle or abandoned lands (Resosudarmo et al. 2019, KSP, 2016. The latter two categories are forest areas that are earmarked for crop plantations, unproductive forests slated for conversion and potential areas for paddy rice farming (KSP, 2016). This has implications for potential disputes between the MOEF and the Ministry of Agrarian and Spatial Planning (Sahide and Giessen, 2015). SF permits are being awarded or planned for 12.7 mil ha of the State Forest Area while these same areas are also being 'released' as private property to communities and individuals for farming. While contestations over TORA has not provided much space for the involvement of civil society organizations (CSOs), the SF policy process on the other hand has embraced their participation.
A loose multi-stakeholder platform, established since the 1990s of CSOs, academics, individuals and businesses actively engaged in all aspects of social forestry, was institutionalized as the formal 'SF Acceleration Working Groups' (POKJA PPS) through a decree by the Director General of Social Forestry (P.14/PSKL/SET/PSL.0/11/2016) to support fast-track implementation of SF schemes on the ground. This act effectively subdued the platform's critical dissent of the policy by rendering its members as technical experts and agents to implement the SF policies. The POKJA PPS engages in the entire territorialization process of SF at the local and provincial levels, including facilitating permit applications, supporting the administrative and biophysical verification and monitoring processes, and reports to the provincial governors (MOEF 2018, Resosudarmo et al. 2019) again bypassing local government agencies. In interviews, members of the multi-stakeholder platform suggest that priority is to first secure SF permits, and then work on other equity issues later. In addition, SF Community Business Groups were formally established to assist in the development of social forestry enterprises (MOEF 2018), which are considered as a successful outcome of a SF scheme. The Indonesian assemblage with its diverse actors, interests and high levels of interconnectedness and contestations suggest that different possibilities could be negotiated at specific moments and places.

Case Study 2: A State-driven Social Forestry in Vietnam
The Vietnamese Constitution defines the State as manager of all forests and land,. Insights from interviews with government actors between 2016 and 2018 suggest that social forestry in Vietnam can be understood in two different ways: social forestry (lâm nghiệp xã hội, xã hội hóa ngành lâm nghiệp) and community forestry (quản lý rừng cộng đồng or lâm nghiệp cộng đồng). The term "community forestry" implies attention to communities as the focus of the policies while "social forestry" focuses on mobilizing finance from private sectors and social actor groups. Findings from interviews conducted with central and provincial government actors as well as from reviews of policy documents suggest that the central government emphasizes "social forestry" and forest policies prioritise collaborative models with local communities and households through an economic and production lens. Moreover local governments tend to apply a "community forestry" approach in their forest protection and development plans emphasizing support to and empowerment of farmers, households, and communities to manage forests. The distinction between the two terms suggest differentiated priorities for local actors in the management of the state's forest estate.
The forestry sector, as with other sectors, is managed by the State to align with the political interest of the Communist party, whose views towards social forestry are not static. Following independence from the French in 1945, forests were exclusively managed by the State through its state forest enterprises (Directive No. 15 CT/CTCW 1961). The early 1970s saw emergence of a political shift towards collaborative forest management (Forest Protection Act 1972), which led to the coercion of local people to support the state as temporary laborers for state forest enterprises often with no direct benefits in return (Pham et al. 2012). Decision No. 184-HĐBT 1982and Forestry Development Strategy 1986 recognized the role of non-state actors (cooperatives, communities and individual households) to manage forests but without accompanying rights. In the late 1980s, transformation of the country's economy from state central planning to market orientation, Doi Moi, triggered liberalization in the agriculture sector and large-scale changes in household ownership of products and land, but this was not translated to the forest sector with "ownership" still limited to use rights. The processes of territorialization with selective inclusions and exclusions is to facilitate economic production. Significantly, the open door policy of Doi Moi enabled foreign organizations to introduce new narratives and demands for devolution and participation (Dang, Turnhout and Arts 2012).
Increasing openness but not for all. Early manifestations of social forestry started in the 1990s (Forest Protection and Development Law 1991 and Land Law 1993) emphasized local communities' duties rather than the benefits they could gain from managing forests (Clement and Amezaga 2008). Many central and provincial government interviewees claimed that "In the past, no one wants to protect and manage forests but central and provincial government assign forests to them as political tasks. Local people saw this as a political mission rather than for their own benefits". In forest areas belonging to state forest enterprises, local people were contracted to work on forest protection or planting (Decree 01/1995/ ND-CP, Decision 661/1998/QD-TTg), with a share of benefits depending on time and labor invested (Ngo and Mahdi 2017). Community forestry was officially recognized in the Forest Protection and Development Law 2004 where "ownership" rights of non-state actors were specified, and the first state-led community forestry pilots were implemented in 17,000 ha across 64 villages in 2007 (RECOFTC 2014). These pilots drew on lessons from the legacy of past forestry policies and projects and the contestations over changing rights and exclusions (Sikor and Nguyen 2011). The government considered community forestry to be a solution to the conflicts between the state and local people in forest and land governance, between conservation and development practices, and an opportunity to address high rate of deforestation and degradation (Decree No. 02/1994/ND-CP, Decree No. 01/1995. The expansion of collaborative forestry approaches in Vietnam was in large part shaped by the policy narratives of international actors, throughthe influence of bilateral aid programs on social forestry funded by Switzerland (1992Switzerland ( -2002 and the Trust Fund for Forest funded by Finland, Netherlands, Sweden andSwitzerland (2004-2015). In particular, MARD interviewees (2018) attributed the government's shift towards social forestry evident in the Forest Protection and Development Law 2004 to the Swiss-funded program on social forestry. Declining state budget allocations to the forest sector pushed the government to consider alternative financing from non-state actors (Pham et al. 2008, Pham et al. 2013, leading to the development and implementation of the national Payment for Forest Environmental Services policy (PFES Decision 380 and Decree 99). International donor-funded programs on PES (Germany and the US) and REDD+ programs (Norway) set the tone on the need to engage meaningfully with communities, indigenous people and private sector. However, State dominance in these processes leaves limited political space for non-state actors, e.g., NGOs and CSOs, in Vietnam to exert an influence on the final policy outputs (Pham et al.2014).
The gap between 'community forestry' and 'social forestry'. The Forestry Law 2017 (enacted on 1/1/2019) gave formal recognition of religious and customary rights over forests for the first time in Vietnam's legal framework. While policies have increasingly emphasized the rights and important roles of communities, individual households and private sector as forest owners and managers, this shift remains rooted in the government's interest to ensure sustainable revenues for the forestry sectorthus reflecting the disparity between the Vietnamese meanings of 'social forestry' and 'community forestry'. For example, social forestry as defined in the PFES policy is a process where all actor groups contribute their financial resources to protect forests and reduce the burden on the state budget (Pham et a. 2013). The Vietnam Forestry Development Strategy (2006-2020) emphasizes an economic and entrepreneurial approach in all forestry actions with local communities and households to advance state development objectives. However, forest "ownership" still only refers to limited forest use rights (except in plantation forests where households have tenure) and little autonomy in decision-making related to their forests (Dang, Turnhout and Arts 2012). An international donor agreed (2019): "Yes, you might see community forestry and social forestry everywhere in policy documents and in the speeches of policy makers, the fact that the state and its state forest enterprises, protected areas and national park and state management boards own and manage more than 85% of forests, mostly the best quality forest in Vietnam while communities and households receive the most degraded forests pose significant challenges for non-state actors to engage in forestry protection and development". By the end of 2018, less than 2% of total forest had been officially allocated to communities (Vietforest, 2018). While there are claims and forms of resistance at different levels, the participation of local people and CSOs in formal SF planning and implementation processes remains limited, despite the participatory rhetoric in policies (from interviews with CSOs in 2019) thus highlighting the state's dominance in the Vietnam assemblage, which effectively dictates territorialization processes to maximize its benefits from Vietnam's forests.

Case study 3: Finding space for Social Forestry within economic development and tenure conflicts in Sabah
The Sabah Forestry Department (SFD), with support from the German Agency for Technical Cooperation, developed and implemented a pilot case for sustainable forest management (SFM) in Deramakot Forest Reserve in 1989 as part of a Forest Stewardship Council (FSC) certification process to address community development issues (UNDP 2008, italics by authors). This pilot approach was considered as a model for SFM, and the acquired FSC certification in 1997 provided impetus for SFD to apply similar 'community forestry' approaches in other FSC forest concessions. The year also marked a significant shift of SFD's core policy approach from industrial logging towards adoption of a SFM strategy and social forestry was institutionalized within this. Under the SFM License Agreement (SFMLA), if there are communities living in or around the forest concession boundaries, the license holders are required to verify community benefits, provide dispute resolution systems and set aside land for community forestry projects within the forest concessions (Toh and Grace, 2005). While acknowledged as an improvement to previous conflicts between industrial logging and local communities, this requirement effectively renders social forestry as a legal and technical obligation for the success of the SFMLA concession. Toh and Grace (2005) argues that it can also be a mechanism for controlling community land use within the forest reservesa territorialization process where forest managers assign strict boundaries for agriculture plots where only approved tree and cash crops can be cultivated.
Persistence of colonial legislation for economic growth. Conflicts around forest access and rights stem from two key colonial laws: the Sabah Land Ordinance (SLO) of 1930, which recognizes Native Customary Land Rights and includes provisions for the granting of Native Titles for indigenous individuals and households and Communal Titles for shared access, and the Ladang Act of 1913 which was meant to control shifting cultivation. Both pieces of legislation signaled the North Borneo Chartered Company's (NBCC) intention to settle Native land rights and identify "idle" lands suitable for plantation agriculture (Majid Cooke 2012, Bernard andBissonnette 2011, Doolittle 2007). The issues of Native rights proved much more difficult for the NBCC however, and this tangled legacy of unresolved rights has led to anxieties, contestation and multiple overlapping claims to forested lands today. Applications for individual Native Titles often take many years to be resolved while logging and plantation licenses are processed quickly 6 . During the peak of palm oil plantation expansion in 2009, there was a reported backlog of 285,000 Native Title applications (Majid Cooke and Toh 2012). The complexities passed on from colonial rule have paved the way for territorialization processes of zoning and the demarcation of land and forests for the long-term state objectives of land commodification and centralized control of natural resources (Lunkapis 2013, Majid Cooke 2013, Doolittle 2007. In 2010, the Sabah government stopped issuing Native Titles to prioritize communal titles (SUHAKAM 2013). In total, 96 communal titles totaling 61,6201 ha were issued to 13,789 recipients from 271 villages in 15 districts in Sabah (Borneo Post, 2018). The shift to communal titling was built on familiar discourses of low productivity, idle land and poor decisions of natives (Majid Cooke 2012) to push forward the State's strategy for rural development and poverty alleviation (SUHAKAM 2013). An amendment of the SLO in 2009 to include the phrase 'any state land planned by the government for the natives of Sabah' made it "possible for the creation of communal titles in cases where the government plans to develop the land, albeit for the benefit of natives" (Lunkapis 2013, p. 202). The communal title, issued through the Fast Track Land Alienation Program, requires the participants to sign an agreement to transfer their rights to state agents for creation of joint ventures for oil palm plantations on the Native lands (Lunkapis 2013). The communal title came with 'special conditions', including expectations that the land has to be cultivated with commercial plants or trees approved by the Director of Agriculture for the duration of tenure (SUHAKAM 2013, Majid Cooke andToh 2012). Large tracts of forest land degazetted for Native communal titles were effectively converted to oil palm plantation and the process was deemed to have "violated the rights of natives" (SUHAKAM 2013, p. 93) 7 . The rapid territorialization and de-territorilization of native communal titles legitimized by policy were intended to benefit the state (and convergence of state leaders' and private sector actors' interests) at the expense of the indigenous population.
Potential for a more equitable form of social forestry and native land claims? Within this dynamic context of Native land as the frontier for development objectives, SF in Sabah occupies a fraught space. The active engagement of civil society organizations and regional ASEAN Working Group on Social Forestry (AWG SF) over the past decade has led to the establishment of SASOF with the SFD in 2016. Serving as a multi-stakeholder platform to discuss and align SF activities and to bring new perspectives and ideas to definitions and understandings of SF, SASOF was instrumental in the development of Sabah and National Social Forestry Roadmap (SFD Annual Report 2016). The Roadmap is undergoing consultation and revisions, and in the latest version, SF is defined as "the involvement of indigenous peoples and local communities on their own initiatives or in collaboration with relevant stakeholders in forest management from the aspects of social, economic, culture and environment for sustainable livelihoods" (MNRE, unpublished draft). At the same time however, social forestry and issues of communal tenure or rights remain conspicuously absent from the Sabah Forest Policy 2018 (SFD 2018) and policy actions only emphasize increased consultations and dialogue with local and Native peoples in forest management, preferential contracting and employment in forest industries, and development of native entrepreneurship (Ibid, p.31). With funding from UNDP, the proposals of CSOs, also SASOF members, to transfer global principles of ICCA (territories and areas conserved by indigenous peoples and local communities) to Sabah provide fresh impetus to integrate Native knowledge and governance practices within SF policies, and is considered by local communities as another mechanism to advance their claims for Native rights (Majid Cooke and Vaz, 2011).

Discussion
From our interrogation of the three case studies across Southeast Asia, the framing of problems and proposed solutions in social forestry closely align to what Li (2007) identified as assemblage practices. In particular, we find elements of: (1) forging alignments between state, international donors, CSOs and communities on social forestry as a mechanism to govern and control forests, to enable participation and improve local capacity, and as a promise to legal enabling rights. This alignment, though fragile and dynamic, has upheld the social forestry endeavour in its various incarnations for the last four decades; (2) the rendering technical of diverse rights and forest governance practices into formal, standardized and enforceable SF programs through bureaucratic and territorialization processes of permitting, licensing and the redrawing of boundaries that potentially excludes more than they include; (3) an anti-politics approach to contain the assemblage by integrating non-state actors (and their criticisms and policy concerns) into agents for piloting, implementing and actively supporting formalized social forestry ideals and programs; (4) the reassembling of social forestry to promote entrepreneurship and the grafting of win-win-win narratives using market ideals of land reforms for growth (TORA), climate mitigation and ecosystem service payments (REDD+/PES), and value chains (SFM).
These practices are shaped and shaping the interplay of the different public and private actors, their interests and their interactions of contestations and consent in the case study regionsand translate into diverse notions of (in)equity and specific pathways to hold state and private sector accountable to the promises made in the name of social forestry, as we discuss in the following sections.

Actors, interests and the many translations of the notion of equity
In our review of state policies and the stated visions and missions of civil society organizations in relation to social forestry (see Table 2), it is clear that states explicitly maintain authority over forests in the design of formal social forestry mechanisms and its territorialization processes. The three governments in our case studies have had varying roles and prominence in the assemblages over time, with other actors' interests emerging to have influence. Vietnamese state actors evolved from complete authoritarian interests to allow (limited) recognition of traditional and customary rights; Sabah state actors were complicit to external capital interests that benefit the powerful but a short-lived reformist government has state actors partially aligning interests with CSO actors; while the Indonesia government appears fairly consistent in its problematizations and territorialization processes despite various reforms and seemingly high levels of engagement with external actors. However, a not unexpected picture of consensus emerges in the states' problematizations of deficiencies and promises for improvement, highlighting their hard work and graft in forging and maintaining the many elements and social-material connections of the assemblages (Li, 2007). Social forestry is held up as a promise for improving the welfare of local communities and to counter contemporary societal problems such as land scarcity, rural poverty and under-development, but also to advance the States' various interests and claim over forests for economic production and growth. While equity is explicit in policies such as the 'Equitable Economy (Ekonomi Pemerataan)' in Indonesia, social forestry and TORA programs under this umbrella are de facto mechanisms to formalize forest access and land ownership, or as Scott (2009) argues, to project state power and bring into control "ungoverned regions". In addition, by providing divergent sets of rights and access, the two programs are in actuality likely to increase the potential for conflict: local communities in SF programs have realized that gaining access to state forests also means acknowledging the state's legitimate authority over forests (Fisher et al. 2018, Myers et al. 2017. Local communities in our research sites have expressed preference for TORA titles which promises access to business opportunities and financing, instead of the management burdens that come with social forestry schemes (field interviews, Kapuas Hulu, 2019 and 2020). Territorialization processes associated with formal bureaucracies of SF mechanisms have reduced the initial notion of equity and social-environmental justice related to the devolution of rights and empowerment, to simplistic notions of fair access to markets with restrictive rights and access, and emphasis on economic entrepreneurship and production.
The evolution of social forestry in Vietnam closely mirrors the country's own transformation from a closed socialist state to an open socialist market economy, from a form of socialist forestry where local people provided labor to state forest enterprises to a market-oriented social forestry that channels development aid and financing from PES and REDD+ schemes to local people to manage and rehabilitate state forests. The Government of Vietnam takes pride in its progressive policies, for example being the first country in Southeast Asia to set up a national PES policy, and setting a new precedent official inclusion of ethnic minorities in forest land 'ownership', with the caveats of the limitations in what the term ownership implies. Social forestry continues to be largely framed within principles of economic empowerment as a solution to what are identified as the problems of underdevelopment and rural poverty. However political pressures to increase forestry productivity, mobilize domestic and international investments, and collaborate with enterprises in production to meet centrally planned development targets are the more important State objectives. Equity, thus, is related to the enabling of economic success, limited to 'success' in the current globaland now national -economic framework.
Sabah is still finding its way with social forestry after decades of intensive logging and state-sanctioned 'land grabbing' for oil palm plantation expansion. SFM and the focus on forest certification appears to provide a middle ground to achieving the multiple economic, social and biodiversity goals stated in the latest Sabah Forest Policy 2018. A new state government of the opposition coalition elected in 2018 promised fresh reforms and a progressive outlook, but political machinations in 2020 have led to a return of the former political power and with it, anxiety of potential re-territorialization processes for continued forest exploitation. While economic and biodiversity conservation goals are emphasized through distinct and elaborated actions and plans in the Policy, there is not a single mention of social or community forestry and rights in the policy document. Hope remains in the Sabah Social Forestry Roadmap that has been undergoing deliberations since 2017, but with little expectation that there will be a reverse in the processes of territorialization that has only benefited the powerful.
In our review of civil society organizations (CSOs) strategy documents, the key civil society actors working in the social forestry space tend to come from the grounded perspectives of environmental justice, forest rights and equity (see Table 2). All the CSOs articulated the detrimental impacts of specific forms of forest conversion and economic development on local peopletheir exclusion, displacement, increased vulnerability and conflictsand the environment, and posit the potential role of community or social forestry to address these issues. Notions of empowerment, justice, governance and maintenance of culture and traditional knowledge are evoked, alongside interests and actions to developing forest product enterprises, creating market linkages and enabling private sector partnerships. These latter interests converge with State objectives of social forestry for economic development. CSO actors who have long engaged with policymaking at different government levels, also aimed to promote transformational market reforms that ensure safeguards and secure rights and benefits for local and indigenous communities throughout Southeast Asia (NTFP-EP, RECOFTC). Is this convergence a strategy which CSO actors choose in order to advance their call for reforms or are market-based solutions the only pathway for empowerment in social forestry? Is the notion of equity merely relegated to fair and equal access to markets?
The relationship between the State and CSOs is more dynamic. In our case studies, policymakers in both Indonesia and Sabah have found ways to work with CSOs to advance their social forestry objectives while Vietnam has largely chosen not to engage with non-State actors, except for select donor institutions contributing finance to the forest sector. CSOs in Vietnam are also more constrained in their activities, often scrutinized by the government. The Indonesia Working Group on Social Forestry started as an independent group of practitioners, researchers and activists with diverse interests in varying issues related to social forestry such as community development, indigenous rights and forest conservation, and in its institutionalization as POKJA PPS, its members are now deployed as experts to facilitate the establishment of formal social forestry schemes to help meet the Government target. SASOF on the other hand is a collaboration between the state (SFD), civil society organizations and academia, first initiated after an ASEAN social forestry network conference held in Sabah in 2014, and formalized in 2016. Their agreed mandate was to develop and promote social/ community forestry; customary and local stewardship, governance and management of forests and other natural resources; and sustainable community enterprises and livelihoods. The CSO members of SASOF are advocating for ICCA principles of recognition of native customary knowledge and rights, while conscious of the limits within Sabah's policies. In both case studies, the CSOs are embedded within social forestry policy processes, but it is not yet clear whether their voiced concerns of social justice are fully recognized and distinct within the discourses of policy making or whether they are co-opted into a political process to enhance State legitimacy by enabling/facilitating the implementation of pre-determined social forestry models that have generally not benefitted local forest communities in an equitable manner (Bong et al. 2019, Moeliono et al. 2017, Sahide et al. 2020. State actors engaged in these networks have interests of varying convergences with CSO and international actors while maintaining formal allegiance to state goals, and these interactions are perhaps the glue to holding together the assemblages. Whether the belief that equity can be found within the formal structures of Indonesia's SF schemes or the call for recognition equity is heard in Sabah, CSOs remain the only actors likely to advocate for such claims.

Public interests, private invisibilities and accountability
In our examination of the social forestry assemblage in the case studies, one group is conspicuously silent: market linkages, tourism, NTFP production and community-based enterprises are mentioned in all institutional and policy objectives, and the private sector actor is invoked everywhere, yet there is little visible engagement of the private sector actors in social forestry policy processes. CSOs provide some definition of their private sector of interest: NTFP-EP identifies "ethical domestic, regional and international markets for NTFPs and creative industries" as a target for the community produced NTFP products (NTFP-EP 2016, p. 17) and RECOFTC identifies "both large enterprises and SMEs [small medium enterprises] and responsible investors" as business and investment partners for the community based enterprises (RECOFTC, 2018, p. 32). However, the private sector actor or market linkage is somewhat hidden in policies that focus only on creating the 'enabling' conditions: permits for the Indonesian formal scheme of Village Forests (Hak Pengelolaan Hutan Desa) can only be requested by a village cooperative or enterprise (Fisher et al. 2018, Moeliono et al. 2015, signaling that village forests are to be managed to generate revenues, while Vietnamese policies emphasize increasing productivity, creating flexible models of collaboration with enterprises and promoting specialized forest production zones. The belief that markets and entrepreneurship is a solution to the problematizations of conflict, poverty and increasing vulnerability of local forest communities and forest degradation is a prevalent discourseincluding throughout ASEAN level policy agreementsand this possibly assigns power to the distal private sector actors while responsibilities of this actor group remain undefined in social forestry policy processes. Our case study of Sabah highlights how the fast-track process of communal titles were designed to make land available to private or public-private development interests, and the return of old political power may yet re-enable these convergences. Other studies have shown how states do clearly represent private sector interests in climate and forest policy decisions , Ross 2001, and this is a critical area for further study, and theoretical and methodological advances. The evidence on outcomes from increasing market access and entrepreneurship is still sparse and Governments often do not have systematic monitoring in place for assessing multi-dimensional outcomes of SF despite their rhetoric. This raises interesting questions on who would be held accountable if the impacts are less than desirable or even negative: the local communities for lacking necessary skills to properly engage with markets, the State and their 'enabling' policies, the CSOs that promote the partnerships, or the private sector actors who are absent and hence do not enable scrutiny for society in social forestry schemes to hold them accountable.

Conclusion
The endeavor of social forestry has forged alignment across state, civil society and international donor actors, and with local communities. Social forestry was initially framed around objectives of equity, empowerment and devolution of rights to local communities to continue their stewardship and governance practices over the forests, while providing a basis for their livelihoods and development. Social forestry was later re-assembled and entangled with the initiatives of agrarian reform in Indonesia, SFM in Sabah and PES and REDD+ narratives in Vietnam, with the underlying narrative that entrepreneurship and fair access to markets is the equitable solution towards economic empowerment and prosperity. Meanwhile, the private sector as a desired new actor in the reassembling of SF is absent from the public discourses but may exercise most power in profiting from economic exploitation of forests and land. While there are voices calling for social justice within the assemblage, these tend to be muted in the anti-politics approach of being assimilated into the practice or implementation of policy as in Indonesia and Sabah, or in their exclusion from the policy process altogether as in Vietnam.
Our work contributes to the as yet limited scholarship deploying assemblage approaches in forestry. The case study findings suggest that a multi-dimensional notion of equity should involve recognition of the diverse governance practices and institutions of local forest communities, instead of the imposition of formal and structured social forestry models from the top-down and discourses inherited from colonial times. Recognition of this diversity would mean that social forestry models has to move current territorialization processes of exclusion to consider property arrangements that may involve different constellations and notions of property and rights and accommodate meanings of forest and landuse beyond economic production values (Bong et al. 2019, Sikor andMüller, 2009). While the principles of ICCA being advocated in Sabah could be a start, social forestry in Southeast Asia is far away from meeting this notion of recognition equity. Most SF schemes, rendered technical through formal bureaucracy and territorialization processes that lead to exclusions and stricter rulesand enforcementon forest use, may in fact reinforce inequities as the poorest are often most affected in these forms of decentralization (Lund et al. 2018).
We suggest that future research focus on impact assessments of social forestry in Southeast Asia, beyond simplistic metrics of hectares and incomes but by asking (and measuring) who benefits from social forestry shemes' implementation. Crucial will be an understanding of how SF policies and practices of territorializations have affected inclusions and exclusions over rights, participation and access over time, have influenced notions of local equity, changed or reinforced contestations and differentiation in forest access and use; and of considerations of power relations in governance practices and market linkages in dynamic forest landscapes. Application of the assemblage approach has allowed for a more nuanced understanding of the heterogenous, contesting and intersecting interests within the social forestry, and this is particularly critical in Southeast Asian rural forest community-state contexts where pervasive State power, do not always allow for conflicting and diversity of views.