Multistakeholder platforms for natural resource governance: lessons from eight landscape-level cases

. Multistakeholder platforms (MSPs) are the subject of increasing attention and investment in the domain of collaborative natural resource governance, yet evidence-based guidance is slim on policy and investment priorities to leverage the MSP approach. We provide a comparative analysis of eight landscape-level MSPs spanning seven countries (Peru, Brazil, India, Tanzania, Ethiopia, and a cross-border case from Kenya and Somalia), representing a diversity of resource systems covering forests, rangelands, and multiuse agricultural landscapes. Applying an adapted social-ecological systems framework, our synthesis identifies the influence of these MSPs on patterns of stakeholder interaction and draws implications for the design and organization of MSPs that are both appropriate and effective. From the cases, we distill lessons addressing: (1) how to design an MSP in relation to the governance context, including the fit between institutional and ecological dimensions of the system and with attention to cross-scale linkages; (2) how to implement inclusive processes that address power inequities, including through capacity building and procedural rules; and (3) how to support adaptive learning to expand the MSP’s influence over time, including monitoring outcomes, adapting the scope of stakeholder engagement, and investing in MSP durability.


INTRODUCTION
Multistakeholder platforms (MSPs) are the subject of increasing attention and investment in the domain of natural resource governance. The 2030 Agenda for Sustainable Development, for example, promotes multiactor partnerships under SDG 17 (United Nations 2015). The Voluntary Guidelines on the Responsible Governance of Tenure of Land, Fisheries and Forests (Food and Agriculture Organization 2012) give national MSPs an implementation and monitoring role. Many regional policy frameworks give similar emphasis; for example, the Framework and Guidelines on Land Policy in Africa (African Union 2010) "promotes the need for a shared vision among all stakeholders of a comprehensive and coordinated land policy as a major factor in national development." Responding in part to these global and regional agendas, MSPs are increasingly being incorporated into subnational resource governance by both governments and civil society (Franco and Monsalve Suárez 2017, Boyd et al. 2018, Ros-Tonen et al. 2018).
However, evidence-based guidance on policy and investment priorities to leverage the MSP approach is slim (Bodin 2017, Sarmiento Barletti et al. 2020b. We aim to shed light on the potential and challenges of MSPs to support inclusive natural resource governance while contributing to restoration of the commons. We view MSPs as sustained, intentionally created, long-term spaces to promote dialogue, deliberation, and collaborative action among social groups and organizations ("stakeholders") who stand to be meaningfully affected, either positively or negatively, by decisions of public importance within a defined domain. Our overarching hypothesis is that MSPs, appropriately and effectively designed and implemented in the frame of adaptive learning, have the potential to contribute to transformative change in landscape governance and management. Our focus of analysis concerns both the appropriateness of the MSP in its context (good design or "fit") and the factors contributing to effectiveness (good implementation). What makes an MSP "appropriate" and "effective" depends strongly on the ways the MSP and organizers account for and address these power dynamics.
In our analysis, we focus on landscape-level MSPs that bridge civil society, government, and private sector actors and aim to improve landscape governance. Specifically, the cases are aimed at strengthening the governance of renewable natural resources for resilient rural livelihoods and enhanced food security and for ecosystem services provision for society as a whole. We provide a comparative analysis of eight landscape-level MSPs spanning seven countries (Peru, Brazil, India, Tanzania, Ethiopia, and a cross-border case from Kenya and Somalia) and covering forests, rangelands, and multiuse agricultural landscapes. We apply an adapted institutional analysis and development framework that was developed with the expectation that it could be applied to assess diverse MSPs to substantiate and validate lessons for practice. We provide a first empirical application of the framework, demonstrating its practical value as an analysis tool and an aid to critical practice.
The framework allows us to explore questions of equity, power, and stakeholder dynamics in a comparative analysis of the cases, focusing on designing for context, designing and implementing processes for inclusion, and using adaptive learning to support improvements in dialogue processes and ultimately in longer term outcomes. From the cases, we distill lessons addressing: (1) how Fig. 1. Dynamics of multistakeholder dialogue and adaptive learning in natural resource governance. Lowercase letters denote the focus of analysis addressing particular linkages. Adapted from Di Gregorio et al. (2008) and Ostrom (2005).
to design an MSP in relation to the governance context, including the fit between institutional and ecological dimensions of the system, and with attention to cross-scale linkages; (2) how to design and implement inclusive processes that address power inequities, including through capacity building and procedural rules; and (3) how to support adaptive learning to expand the MSP's influence over time, including monitoring outcomes, adapting the scope of stakeholder engagement, and investing in MSP durability.

COMPARATIVE FRAMEWORK AND RESEARCH METHODS
Governance concerns the institutional framework, both formal and informal, in which power is exercised over matters of public importance. Natural resource governance, therefore, focuses on the exercise of power as it relates to the control, use, and management of natural resources (Larson and Soto 2008). Governance fundamentally shapes the relations among actors in social-ecological systems, which means that it can be a powerful enabler of, or obstacle to, system transformation.
Recent research on natural resource governance (Andrachuk and Armitage 2015) has looked at transformation both as process and outcome. Systems transformation typically entails processes of disruption and confrontation as well as collaboration and cocreation between actors (Dentoni et al. 2017). Amid this diversity of change processes, MSPs are particularly suited to facilitate collaborative action and co-creation, complementing other strategies (e.g., social movements, protest politics, advocacy campaigns) that explicitly aim to disrupt and confront established power relationships, however inequitable. Co-creation increases shared ownership and acceptance of solutions identified together, which makes it more likely that subsequent behavior changes (Schut et al. 2013). Bringing different actors together in structured dialogue also reveals power asymmetry and sources of vulnerability to participants (Pelling 2011, O'Brien 2012; that awareness can yield more inclusive and accountable natural resource governance. These effects create an environment in which transformative system change becomes more likely, in ways that favor both equity and sustainability (Frantzeskaki et al. 2012).
Our focus is on the role and potential of MSPs to contribute to transformative change as a deliberate, inclusive, and accountable response to deficiencies in natural resource governance, positioning actors and institutions as key forces both responding to the governance context and capable of instigating change within it (Giddens 1984, Olsson et al. 2008). This is not to imply that MSPs necessarily yield such results (Warner 2007, Sartas 2018. Rather, it sharpens the focus on comparative analysis to understand the ways in which MSPs in practice affect stakeholder relationships and behaviors. We present a comparative case study application of a conceptual framework developed for analyzing MSPs that address natural resource governance (Fig. 1). The framework takes a relational systems perspective (Lerner and Schmid 2013), asking that we consider not only the characteristics of the MSP but also the context in which it operates. In this respect, the landscape framing (i.e., a geographically defined social-ecological system) is a critical feature of that context and distinguishes MSPs focused on landscape governance from those that take a national or issuebased focus. It also brings into sharp relief the practical consequences of diverse resource uses and user groups operating https://www.ecologyandsociety.org/vol27/iss2/art2/ in close proximity (Kusters et al. 2020). Drawing from the institutional analysis and development framework (Ostrom 2005), the context includes the attributes of the resource system and history of resource use, characteristics of resource users, and current governance arrangements, which influence the scope for MSP design and functioning (Fig. 1, arrow a). We refer to designing an MSP with attention to these contextual factors as "designing for context".
Within an MSP, seen as a purposefully designed action arena, the patterns of stakeholder interaction, and thus the outcomes, are influenced by actors' characteristics, the action resources available to each actor, and the formal and informal rules at play (Di Gregorio et al. 2008). In assessing the role of the MSP in shaping patterns of stakeholder interaction (Fig. 1, arrow b), we give particular attention to inclusivity in representation and equity in decision-making to address power inequities, which have been demonstrated to influence stakeholder commitment and institutional durability (Faysse 2006, Brouwer et al. 2013). This aspect we term "designing for inclusion".
We recognize MSPs as dynamic by nature, incorporating feedback loops (McGinnis and Ostrom 2014), with outcomes of stakeholder interaction within an MSP influencing future stakeholder interactions. The way that competing interests, rights, and power constellations of diverse actors are managed is of special importance at this outcome level. Strengthening the adaptive learning capacity of stakeholders is critical for creating lasting impact. The learning process creates trust, helps to find balanced solutions, and empowers actors to manage future change (Hage et al. 2010, Reed et al. 2010, Scholz and Steiner 2015. These aspects concern the potential for adaptive learning to improve multistakeholder dialogue in future interactions (Fig.  1, arrow c).
Though it is not our focus in this comparative analysis, the framework also encourages consideration of how, over time, MSP processes can contribute to: outcomes that shift resource status and trends; attributes of actors such as livelihood assets, wealth, and vulnerability; and governance arrangements that are enduring (Bodin 2017). This process occurs as a feedback loop ( Fig. 1, arrow d). Assessing such factors typically requires a longer time frame than most of the case studies included here, as well as robust baseline data. Nevertheless, where there are preliminary indications of such dynamics of systems change according to independent monitoring or observations of the actors involved, we do note them.
While the framework visualizes an abstract set of relationships among context, action arena, and outcomes, in practice, these elements are characterized by real tensions and power dynamics (Bodin 2017, Denney et al. 2018. Power distribution and politicoeconomic interests play important roles in the design of institutions and the rules that emerge to manage stakeholder interaction, a factor that the original institutional analysis and development framework does not adequately consider (Clement 2010). Rather than viewing MSPs as technical, managerial, neutral, or apolitical spaces, therefore, we argue that the potential of such arenas can only be understood and realized through better understanding of such MSPs as socio-political constructs (Lefebvre 1991) that evolve in relation to their context (Warner and Verhallen 2007). Thus, we seek to understand the spatial context in which power operates (Brenner et al. 2003) and how MSPs are (re)shaped by existing power structures and relationships. Viewing the action arena as an arena of power struggle, we look at an MSP as a potential means to promote deliberative decision-making and as an organizational tool to open and create political spaces fostering inclusive institutional innovation (Cleaver and Whaley 2018).
The study methods entailed purposive case study selection (Patton 1990) and qualitative, theoretically informed comparative case analysis (Mangen 1999). Each of the eight cases represents a focus of analysis or engagement within the flagship program on governance of natural resources within the broader CGIAR Research Program on Policies, Institutions and Markets (CGIAR Research Program on Policies, Institutions, and Markets 2021). In contrast to a comparative review based on prior published literature, our analysis benefits from the in-depth engagement of coauthors and their research teams in the various cases. This engagement provides a window into often unpublished practitioner experiences facilitating MSPs at the landscape scale. This involvement offers the potential for more nuanced insight into multiyear stakeholder engagement processes and identification of emergent lessons that can contribute to strengthening those and other efforts.
Before undertaking the individual case studies, the research team developed a joint comparative framework (as summarized above) drawing upon a review of the literature, and a template with a guiding set of questions to structure each case study (see Appendix 1). Individual case study researchers developed the case study write-ups, drawing upon existing research engagement, supplemented by additional interviews to capture insight on emerging outcomes and lessons. Sources of evidence thus include documentation from available reports in each case, workshop notes, focus-group discussions, and key-informant interviews. To validate outcomes, including observations of challenges and success, researchers aimed to triangulate the perspectives of multiple actors engaged in these processes whenever feasible. Where judgments of different actors within a particular case differed in assessing the MSP's value or outcomes, interviews probed the reason for the divergence as a route to drawing further lessons, and these differences are noted. In the analysis that follows, our aim is to highlight findings from each case that are most salient in relation to the key dimensions of the comparative framework. In particular, we focus on the first three elements of the framework: designing for context, designing for inclusion, and adaptive learning.

Case overview
Each case study has its own context, characteristics, challenges, and corresponding purpose for the MSP (Table 1; additional case details in Appendix 2). We hereafter refer to the cases by site name. All cases concern an ongoing, multiyear MSP initiative that functions in convening dialogue and has a role in resource management. In three cases (Madre de Dios, Acre, and Pará), the MSP is a formal mandate of government through official decree or law, reflecting policy toward decentralization or greater involvement of historically marginalized stakeholders in resource governance, in which government plays the role of convener. In other cases (Gujarat, Odisha, Chemba, and Oromia), the MSP is a result of a nongovernmental organization (NGO) initiative, gaining legitimacy through engagement and commitment of government bodies. One case (Tana-Kipini) concerns a binational initiative supported by an external funding agency, which entails both local and national actors under a trans-boundary framework.
None of the case study MSPs have official decision-making authority, though the proximity to official decision-making varies considerably. At one end of the spectrum is Chemba: Because the MSP has no formal legal status, it has no power to enforce decisions, so it relies on engagement with and influence on local authorities, building social norms to support compliance with agreed action plans and links to civil society policy networks at the national level. The Gujarat and Odisha MSPs play a coordination and advisory role to raise priorities in local planning for issues that are high on the communities' agenda. Although the outcomes are nonbinding for local government, through continuous engagement, the MSP can often influence the larger development agenda, have recommendations adopted in official block-level plans, and launch efforts that require collective action among local communities.
The three government-mandated MSPs have a more direct advisory function. In Acre, MSP products, including territorial zoning plans, need further approval from other official decisionmaking bodies (three state councils and the state legislative assembly). MSP organizers and most participants identify the zoning plan as a policy-guiding tool to recommend and orient subsequent investment projects and actions (Gonzales Tovar et al. 2021a,b). In Pará, the MSP results typically take the form of recommendations, but these have significant weight, mainly because of the presence and participation of the Public Ministry in the proceedings (Londres et al. 2021). In Madre de Dios, the MSP convenes dialogue sessions to reach agreements and raise awareness among stakeholders of the deforestation and degradation challenges faced by the Communal Reserve, a type of protected area (Palacios Llaque and Sarmiento . Agreements may be implemented as recommendations toward Building positive relations among stakeholder groups and giving increased visibility and value to indigenous peoples, traditional populations, and small-scale farmers Forum was activated to agree on a plan for ecological and economic zoning that would address historical conflicts over land and resources and create positive engagement through "forest citizenship" Forum validated land-use mapping, database, and land-use policy recommendations, which were implemented by the state government; landscape improvements were reported following restoration action Pará, Brazil Structured and improved dialogue among stakeholders, but with grassroots communities and organizations frequently excluded Forum mediated negotiations between the private sector and public prosecutor and between municipalities and the state government Collective action by political and economic elites to reduce deforestation, but results challenges by environmental and grassroots organizations, citing tenure insecurity Odisha, India Local federations took ownership of convening and leading the multistakeholder platform early in the process; state then took up convening role, expanding the model to other blocks and districts Rules on forest fire management agreed; block-level rules formed to govern open grazing; mining companies are not actively engaged and usually not present Improvements in landscape restoration reported as a result of collective rules adopted to control open grazing, reduce biomass loss, and restore vegetation on communal lands

Gujarat, India
Multistakeholder platform enabled community peer-to-peer learning, community capacity and ownership, and increased trust between communities and government actors Local participants cited increased harmony within villages and confidence in collective solutions to local resource management problems Improved agriculture and planting techniques (crop diversification, spacing, mulching, drip irrigation) reported on common lands, contributing to increased farm productivity Chemba, Tanzania Improved confidence and trust among local communities, local government, and other actors, including nongovernmental organizations playing role as facilitator or intermediary Resolved conflict over extent of protected area; greater commitment from local government to intervene in land-based conflicts and issues; fewer conflicts between land users Greater incentives established for community investments in improving land productivity and for respecting boundaries, rules, and regulations of protected area Oromia, Ethiopia Improved representation of women and private sector agriculture and forestry investors; improved coordination between different nongovernmental organization projects in the ecoregion Newly established cooperatives promoted livelihood diversification, aiding rule enforcement and reducing conflicts; tensions over power imbalances remain an obstacle Collaborative enforcement efforts among state and local actors credited with helping to reduce deforestation; increased use of alternative energy sources such as fuel-efficient stoves Tana-Kipini, Kenya and Somalia Peaceful dialogues initiated across borders; community management and development plans agreed, and community monitoring system created Following history of violent civil conflict, the dialogue process helped communities move beyond immediate resource disputes, enabling greater cross-border trade and development Progress in collective agroforestry, rainwater harvesting, farmland and protected areas rehabilitation through natural regeneration; reduced incidence of human-wildlife conflict reserve management and may subsequently become legally binding agreements between all stakeholders participating in the management committee or may inform national policy on the topic (e.g., the intercultural co-governance of the reserve). As part of its mandatory tasks, the management committee also approves the master plan for the reserve.

Summary outcomes
A comparative analysis of the eight cases identified three common areas of influence on patterns of stakeholder interaction (in the action arena, following the categories of the comparative framework): inclusion, collaboration, and trust; conflict management; and collective action for conservation. These areas are considered proximate changes for which causal connections to the MSP process are readily established (Table 2; additional details in Appendix 3).
Inclusion, collaboration and trust: All of the MSP cases show at least some improvements in communication, collaboration, and trust between different stakeholders. In many cases, this improvement involved strengthened links between different community groups, between local communities and government, and across different sectors. In many of the same cases, however, various forms and degrees of exclusion remained because of existing power asymmetries and conflicting livelihood priorities.
Conflict management: Although some forums did not include conflict management as an explicit aim, they all reported measures that served to reduce at least some aspects of conflict. In several cases, conflict resolution was cited as an important outcome. Various rules and mechanisms that reduce tensions were defined through dialogue among stakeholders, ranging from zoning rules to forest fire management and cross-border trade. However, the success is mitigated by the failure to include stakeholders involved in some important conflicts, including small-scale resource users such as informal miners (Madre de Dios) or larger commercial actors such as mining companies (Odisha).
Collective action for conservation: Participants in all cases noted increases in collective action to address conservation priorities, often connected to targeted livelihood improvements, though not always involving all essential stakeholders. There were reported Ecology and Society 27(2): 2 https://www.ecologyandsociety.org/vol27/iss2/art2/ improvements in landscape restoration and related incentives for ongoing action to maintain and protect ecosystem functions in six of the cases (Pará, Acre, Oromia, Gujarat, Odisha, and Tana-Kipini). Although much of this influence was positive in terms of resilience goals, significant shortcomings were also identified.

EMERGING LESSONS
Recognizing this diversity of outcomes, each of the case experiences was examined individually and in comparative perspective to draw implications for the design and organization of MSPs that are both appropriate and effective. In relating the lessons from the cases to the conceptual framework, we examine the first three causal relationships of the framework in turn: designing in relation to the broader governance context (Fig. 1, arrow a), designing (and implementing) processes for inclusion ( Fig. 1, arrow b), and using adaptive learning to strengthen multistakeholder dialogue (Fig. 1, arrow c).

Designing a multistakeholder platform in relation to the governance context
In reviewing the way in which context affects the action arena of the MSP in the case studies, or how the MSP organizers addressed context, we identified three lessons addressing: the geographic scope of the MSP, linking across scales, and filling governance gaps.

Geographic scope of the multistakeholder platform may be defined either by ecosystem or administrative boundaries
Multiple authors have argued that the concept of a landscape defined by biophysical features is problematic for governance and management (e.g., Görg 2007, McCall 2016, or that "naturalizing" the basin or watershed scale may disempower certain actors (Warner et al. 2014). Some of the MSP cases were formed around landscapes boundaries, whereas others coincided with administrative or jurisdictional boundaries. The case results suggest that operating at a landscape scale is not a problem, as long as the relevant government authorities are involved. The jurisdictional scale is more likely to guarantee state involvement and a connection to relevant processes, but landscape approaches do not preclude similar connections.
In Acre, the Ecological-Economic Zoning process is, by definition, a jurisdiction-wide effort; it is a legal requirement of Brazil's states. The Green Municipalities Program in Pará also operated at the jurisdictional level as a state-led process to bring the region's 144 municipalities into compliance with Brazilian law on deforestation. The program was scaled across the state after the municipality of Paragominas implemented a successful MSP process that resulted in its removal from the "blacklist" of highdeforesting municipalities that was put together as part of Brazil's Action Plan for the Prevention and Control of Deforestation in the Legal Amazon (Londres et al. 2021). In this case, the upscaling from municipal jurisdictions to a coordinated state-level approach facilitated the response to central government policies.
Other MSPs were initiated to address a particular resource challenge rather than a jurisdictional mandate. Such was the case in Chemba, where the MSP aimed to provide a platform for resolving conflicts that were increasing in number and degree of violence. The district administration provides the forum for addressing these disputes, though it is influenced by decisions in neighboring districts and by national policy and legislation. The MSP cut across divisions between government sectors and different actors in land governance. As one village leader said, it "is a bridge that connects villagers at the grassroots level with high decision makers. It helps the village authority to understand their responsibilities on land issues better." The Oromia MSP aims to conserve biodiversity and ecosystem functions and services in the Bale ecoregion and, as such, is not based within a single jurisdictional boundary (Yami et al. 2021). Similarly, in Madre de Dios, a multistakeholder management committee was set up to support co-management of the Amarakaeri Communal Reserve and to bring a wider set of stakeholders into dialogue with the reserve's co-management partners. Following legal regulations for Communal Reserves, comanagement is shared by the Peruvian natural protected areas authority (Servicio Nacional de Áreas Naturales Protegidas por el Estado, or SERNANP) and ECA-Amarakaeri, an indigenous organization representing the ten indigenous communities inhabiting collectively titled territories within the Reserve's buffer zone.
Linking to other platforms and multiple scales presents opportunities to maximize impact Effective MSPs sometimes generate substantial momentum that builds across the larger landscape, contributing to cross-scale linkages that are pivotal in promoting transformational change (Robinson et al. 2017). MSPs at the same scale promote mutual learning across similar groups, whereas multiscalar approaches allow targeted planning and differential impacts based on the potential of each different arena. Scaling up is seen as a way to promote greater policy impact. Three of the cases demonstrate this kind of momentum.
In Gujarat and Odisha, MSPs were initially very local in scale and gradually developed links to higher levels of governance. Initiated by the NGO Foundation for Ecological Security (FES), the Odisha MSP encompasses all the villages within a hill range, bringing together different village federations. Similarly, the Gujarat MSP brings together village federations from three watersheds and a hill range. Both MSPs evolved over time to address the higher scale of planning at the block (subdistrict) level, bringing federations together with various government agencies, NGOs, and the private sector. Working at the block level encourages mutual learning and knowledge exchange between different stakeholders, especially different village communities, on best agricultural practices and resource governance. Bringing federations together has also revived intercommunity cooperation in resource governance through development of intervillage conservation action plans; this scale is local enough to allow for regular interaction and large enough to achieve conservation outcomes for the landscape.
The Chemba MSP was conceived explicitly as a pilot effort to validate the potential for dialogue to resolve resource conflicts in Tanzania, and an MSP in nearby Iringa district was established with similar objectives. Building on the success of these two pilot projects, MSPs are being established in additional districts. The NGO Tanzanian Natural Resource Forum (TNRF) provides a link between these MSPs, for example, by facilitating exchange visits among the districts to share experiences and good practices. It also fosters links to higher level processes and forums such as the National Engagement Strategy coordinated by the Tanzanian https://www.ecologyandsociety.org/vol27/iss2/art2/ Land Alliance and supported by the International Land Coalition. The aim is to influence national policy and legislation to address, for example, conflicting regulations that contribute to the root causes of local problems.
In the Tana-Kipini case, the MSP invested heavily in capacity building and training on environmental management and biodiversity conservation, which enabled a nested approach to coalition-building. Village-level partnerships linking diverse stakeholders were aggregated at community, cluster, county, and regional levels, working from the bottom up to build coalitions to then work on broader landscape issues within the MSP. This process was complemented by efforts to establish links between agencies at county and national levels within Kenya and Somalia. The progress has been especially notable in view of the history of conflict in the border region in recent decades.

Multistakeholder platforms can fill important governance gaps
MSPs are often intended to solve coordination problems in a variety of ways. Most commonly, this process involves actors in a landscape, including different government offices, who have no other effective space for coordination to address land use and related resource governance challenges (Ros-Tonen et al. 2018). The Oromia MSP succeeded in addressing the lack of integration among different sectors operating in the ecoregion, with an emphasis on getting different NGO projects to coordinate efforts. Interviewees also cited previous top-down approaches used in land-use planning, and the "isolated" nature of interventions in the past, as key constraints to sustainable land use in the ecoregion. By providing a platform for stakeholders to negotiate their interests and priorities, the initiative succeeded in reducing conflicts. The same is true for the Tana-Kipini landscape at the international level. The MSP, supported by the European Commission, served the function of an interim governance mechanism in the absence of a formal governance system in the cross-border area.
In the Chemba case, addressing the governance gap entailed the creation of important new governance bodies. Compared to other countries in sub-Saharan Africa, Tanzania has reasonably robust land policy and legislation based on the principles of: recognizing existing rights, including customary rights; decentralization of land administration and dispute settlement to the local level; and facilitating registration of land and titling of land rights to enhance tenure security and promote land markets (Luhula 2017). However, implementation is beset with challenges. District authorities maintain a dominant role over village authorities and their decisions, frustrating decentralized decision-making. There is also generally poor understanding of power and the role that the village council could and should be playing under the formal policy framework. Indeed, many villages have not yet established the village governance structures (including land governance structures) required by law. In 2017, for example, only 30% of villages in Chemba district had village executive officers, and few had land committees. MSP support for coordination among responsible land governance and administration institutions in the district, combined with the establishment of village land councils, have helped to realize the intent of decentralization in land and natural resource administration.
In the Madre de Dios case, the gap addressed concerns about the past exclusion of indigenous peoples from protected area management arrangements. The case is connected both to Peru's policy transition toward more equitably managed protected areas in indigenous territories and to the wider political demands for a greater say by indigenous Amazonian organizations over the management of protected areas, which also helped to advance Peru's Law of Prior Consultation, passed in 2011. The introduction of the co-management model, which includes indigenous organizations and communities in the management of Communal Reserves, was not initially considered as part of the law that introduced those reserves. It was tested in Madre de Dios, following demands by local indigenous peoples, and was later introduced more widely in the Peruvian Amazon.

Implementing inclusive processes and addressing power inequities
Lessons in this section address the ways in which the MSP is structured to enable inclusive dialogue and yield more equitable patterns of stakeholder interaction, with the aim of promoting outcomes that favor social-ecological resilience (Fig. 1, arrow b).

Recognizing barriers to inclusion is key for effective knowledgesharing and dialogue
Criticism of the participatory paradigm that preceded the current interest in MSPs notes that some of the main power imbalances challenging MSP-like processes are at the level of technical knowledge held by different participants and in the ability of more powerful participants to decide what kind of knowledge is more important than others (Cooke andKothari 2001, Edmunds andWollenberg 2001). Across the cases, differences in knowledge as an action resource were identified as a critical source of power imbalances.
In the Acre case, research revealed clear efforts to mitigate the effects of differences in access to technical knowledge. The state environment agency (Secretaria de Estado de Meio Ambiente, or SEMA) and the private sector (who hired experts to give them technical support in the process) were identified as the actors with the highest technical knowledge and, thus, with the highest ability to influence the zoning commission. Their situation was in stark contrast to representatives from indigenous organizations, traditional populations, and smallholder farmers. Nevertheless, participants valued the way that SEMA guided discussions and acted as a mediator, making the technical knowledge available in a more neutral way and ensuring that traditional knowledge was also valued in the commission. Organizers made efforts to present technical information using nontechnical language to facilitate the participation of indigenous and rural worker representatives as well as to explain technical concepts clearly. Additionally, smaller and more frequent meetings were organized by thematic groups and chambers, enabling more focused discussions and negotiations. These meetings included spaces with more homogeneous participants (actors working on similar issues or sectors), where it was easier to solve conflicts and reach agreements.
The Gujarat case illustrates how investing in locally led systems for data collection and analysis can provide a basis for more democratic decision-making. Training community resource persons to collect and interpret data enabled them to tap knowledge as an action resource. Village federations were concerned by the provision of government services such as healthcare, schooling, and irrigation infrastructure. Official government data did not seem accurate from the communities' https://www.ecologyandsociety.org/vol27/iss2/art2/ point of view. Youth from various villages were appointed as resource persons to collect local data using simple tablet-based applications, which enabled real-time data aggregation and analysis. This process revealed, for example, locales where more teachers were needed or where remote villages were disadvantaged and lacked the necessary services such as healthcare centers. Drawing on these community-generated data, the facilitating NGO then helped in data visualization and presentation, increasing the credibility and impact of the information. The supported local communities were able to gather data to back up their claims regarding priority challenges, including lack of infrastructure and access to services. These data boosted the communities' credibility and helped build support from other stakeholders, prompting government to recognize and respond to inequalities between villages as part of block planning.
Several cases also show an important link between women's empowerment and capacity building in MSP processes. In Oromia, there has been special emphasis on including women because of their dependence on forests for firewood as an energy and income-generating resource. Indigenous women, in particular, are seldom represented at the MSP, which is a reflection of the unequal access to governance spaces for women in indigenous communities and women's disparate opportunities to build capacities for effective participation (e.g., schooling and public leadership roles). Similarly, in Gujarat, MSP participants related inadequate levels of women's participation to gender norms, women's time constraints due to domestic work, and distance to travel to meeting sites. It is noteworthy, however, that members of women's self-help groups have higher rates of participation and express feeling empowered to speak out at these events. This situation points to the importance of capacitybuilding and collective action efforts outside of the MSP that may build a foundation for effective participation.

Procedural rules and facilitation strategies can help mediate power relations
Several of the cases illustrate explicit mechanisms within the MSP to ensure attention to the priorities of groups that are frequently disadvantaged in other forums. In Acre, for example, during the second phase of the zoning commission, organizers established a parallel "ethno-zoning process" in response to a demand from indigenous peoples for respect for their right to selfdetermination. The process was run for indigenous peoples by indigenous representatives as a space to discuss their own priorities with regard to the key issues being addressed by the MSP. Indigenous participants emphasized the importance of the MSP as a purposeful space for them to hold discussions with the government, address differences in technical knowledge, and ensure that their voices were heard and that the zoning decisions benefited indigenous peoples. Even with these measures and the decentralized workshops held in municipalities, other indigenous representatives considered the representation of indigenous peoples incomplete.
Other rules aim to preclude hijacking of the process by the most publicly vocal actors. In the case of Gujarat, an explicit rule bans political agendas and rhetoric from the MSP in an area with polarized party politics. To reduce power imbalances between different stakeholders, including different castes and tribes, women and men, and government officials and village members, there are no chairs at meetings so that everyone is seated on the floor, and participants note that this measure is quite effective in reducing power asymmetries.
Finally, there are rules aimed at ensuring balanced engagement and participation. To address power imbalances in the Pará MSP, the organizers acted as mediators, ensuring that all participants had the right to speak and to vote. Every meeting's agenda was discussed beforehand, and a quorum rule was adopted requiring that attendance in meetings include 50% civil society and 30% government representation. Nevertheless, important actors, including indigenous peoples and local communities, were left out of the MSP and the programs it supported. The Green Municipalities Program initially failed to differentiate its approach to the different municipalities, but adaptations were made over time, including preparatory workshops, for example, to understand locally relevant drivers of deforestation and build local capacity to articulate and address these distinct challenges. Rules were also adopted to enable difficult issues to be deliberated in smaller working groups that would provide recommendations during the MSP's plenary. This process enabled more intensive deliberation and consideration of a broader range of perspectives to feed back into the process.
Several cases also highlight the importance of effective facilitation to create conditions for inclusive dialogue, with attention to the formal and informal relations between participants, including cultural and gendered power dynamics. In the Chemba MSP, organizers purposefully invited the District Commissioner to chair the MSP sessions, and government and community leaders decided on who should participate from their constituencies. There was also a separate role for a facilitator to focus on opening spaces for all actors to contribute and to influence decisions. Preparatory meetings were key for certain groups to share knowledge on the issues to be discussed and to strategize joint courses of action. In addition, targeted training (for example, on women's land rights) and support for community-based land monitors also helped to build deeper participation.
Frank recognition of conflicts among participants, and structured dialogue to address them, can also help to establish the value of the MSP. In Madre de Dios, for example, roundtables were organized around divisive issues such as the construction of a road across the reserve's buffer zone. Participants appreciated the MSP's success in addressing divergent points of view and eventually reaching agreements.

Roles of government actors must be balanced to lend legitimacy and enable follow-up action without controlling the process
We considered government engagement to be key to MSP legitimacy in all of the cases. Such engagement includes providing a connection to broader and parallel plans and processes, bringing attention to the MSP's concerns, and providing support for or enforcement of agreements reached in negotiations. Whereas most MSP organizers clearly want government engagement, they do not want government control. Also, even if the government brings legitimacy to the MSP, the process may not be equitable. A variety of cases illustrate these tensions.
The driving force behind the Chemba MSP was TNRF, a domestic NGO respected for its knowledge of land issues and considered a relatively neutral actor in the highly politicized land arena in https://www.ecologyandsociety.org/vol27/iss2/art2/ Tanzania. However, TNRF recognized that for the MSP to be both effective and sustainable, it was vital that the government play a central role to assist in enforcing decisions. Because the MSP was not a legal entity, it had no power to enforce compliance with the agreed action plan. Recognizing that the District Commissioner was the key official with power to enforce decisions made in the MSP, organizers invited the commissioner to chair the MSP. Having assumed this responsibility, the commissioner was more likely to attend the meetings, thus providing an opportunity for participants to provide full briefs on the different issues in a balanced way and secure the commissioner's buy-in and commitment to take forward and, where appropriate, enforce agreed action points.
Similarly, in both cases in India, NGO leadership helped bring government to the table. In Odisha, after the MSP was jointly organized by the village federations and FES for several years, the government became interested in adopting the annual event, taking on the role of convener. FES and the federations remained as key participants, assisting with planning and follow-up. Importantly, the government has initiated similar MSPs at the block level in other blocks of Odisha state, as well as at the higher district level. Participants viewed this development positively because it increased legitimacy and recognition from the state for the MSP, its members, and the co-management process, especially given that the village federations still have ownership of the process and are able to influence the MSP agenda. Nevertheless, organizers were concerned that government agencies might not refrain from controlling the process when convening MSPs, which poses risks for power imbalances in the future.
The two cases in Brazil represent MSPs that were, by contrast, initiated by government. The role of the government in the Acre MSP was central to its legitimacy because the state government needs to approve the zoning plan. Government commitment to both the inclusive, deliberative process and its goals were key factors leading to the vast majority of participants reporting that the MSP was both equitable and effective. This situation is based on the political history of the state: the Workers' Party won on a political platform supporting social justice and consolidated a socio-environmental alliance with NGOs and the Catholic Church. The MSP, although organized following Brazilian law, followed a process purposefully designed to promote sustainability, forest conservation, and the participation and empowerment of historically underrepresented groups. Acre's zoning commission was implemented within a national and regional political environment that enabled the participation of civil society and multisector collaboration while supporting sustainable development in territorial planning.
Similarly, the Pará MSP was organized and led by the state government as a response to a national government program. This leadership and fit to national policy gave the effort substantial legitimacy among those involved. The program was effective in bringing together different sectors of society that do not normally dialogue, negotiate, and generate joint solutions, such as large producers' organizations, the Public Ministry, and large NGOs. In fact, as one of the largest environmental platforms in Brazil, the program placed Pará in the international spotlight and is commonly perceived as having sophisticated and integrated mechanisms to combat deforestation. Nevertheless, smaller NGOs, subsistence farmers, indigenous peoples, and black communities were largely excluded, and interviews identified profound historical animosity between those who participated and those who were excluded.

Supporting adaptive learning
Lessons in this section focus on the learning feedback, i.e., the ways in which structured reflection on the MSP outcomes can contribute to improved management of the MSP and its longerterm durability and impact (Fig. 1, arrow c).

Monitoring both processes and results of multistakeholder platform implementation can improve effectiveness
Monitoring results of MSP implementation as part of adaptive learning processes is important for MSP development and strengthening (Kusters et al. 2018). Many MSPs have widereaching goals to influence governance beyond the MSP. Taking a step back to review the big picture and wider context, including any changes in the political and social landscape, stakeholders, and their interests, provides an opportunity to realign the MSP and strengthen its impact pathways. Giving time for reflection and ensuring that there is flexibility in the structure, design, ambitions, and goals of an MSP are vital for achieving governance transformation.
Among the cases we compared, explicit monitoring and evaluation of outcomes and formal, structured reflection on lessons were inconsistent and often missing altogether. However, there was broad evidence of learning and adaptation across the cases, signaling that informal learning processes were engaged. Most case study MSPs relied on informal processes of selfreflection (and occasionally protest and hot debate) as the main way of monitoring the quality of the process. An exception was the MSP in Madre de Dios, which must undergo an annual monitoring exercise carried out by a SERNANP employee. Even this measure, however, seems to be secondary to the main learning process. Some participants noted that the official tool monitors issues that they did not find important and produced results that they were unable to use to support their annual planning processes. This gap took MSP participants to collaborate with Center for International Forestry Research researchers in developing a participatory monitoring tool to support reflexive learning about their MSP's process, priorities, and progress, which was implemented for the first time in 2020 (Sarmiento Barletti et al. 2020a).
In both cases in India, village federations typically discussed the MSP's outcomes to gather lessons and improve the MSP for the next year. As part of systematizing its support to these MSPs and similar block-level platforms in other states, FES is refining indicators to track different levels of achievement, encourage more regular reflection, and share experiences on obstacles and strategies across the different MSPs. The aim is to strengthen formal monitoring and evaluation to aid in comparison across multiple sites while retaining the open reflection and participatory planning that are strengths of the current practice.
Participants in a range of other MSPs noted the need for more structured processes for monitoring and evaluation when they were lacking. In Chemba, although TNRF had no structured plan for monitoring the results of the MSP, the cycles and reporting requirements of projects financing the MSP and the cycle of https://www.ecologyandsociety.org/vol27/iss2/art2/ seeking new funding highlight the need for monitoring (and reporting) of impacts and reflection on these together with how the MSP could be improved. In this case, the local MSP being nested within a wider network provides opportunities for annual reporting and contributing insights into planning of the wider network, such as the national engagement strategy on land governance in Tanzania.

Adapting the multistakeholder platform's scope and stakeholders engaged can help it evolve toward transformational change
With or without structured, regular, and deliberate monitoring and reflection, many of the MSPs have adapted the scope of the MSPs and the stakeholders engaged in them. For example, the MSP in Gujarat was initially conceived more modestly as a platform for information exchange and awareness building on issues of ecology and livelihoods, but it evolved to target blocklevel development planning and lasting improvements in local democratic governance more broadly. This transformation required incorporating additional issues such as health and education that were high priorities among local actors. It also meant that the practices of participatory governance became more institutionalized.
In the case of Chemba, membership has been quite fluid depending on the issues being discussed or the agenda for the year. At the start, meetings tended to focus on a particular case or conflict, so the participants would be those directly involved. Over time, however, new stakeholders were included, and more regular membership was established, with representation of key stakeholders across the district and communities. This change provided a distinctive opportunity for village representatives to meet and engage with district officials and directly address their problems in a swift manner. This process not only saved community members time and resources (avoiding numerous trips to district offices), but also gave a higher profile to the problems, encouraging the district or other officials to find solutions more quickly.
In the Tana-Kipini case, we considered that expanding membership in response to emerging priorities is an important aspect of adaptive learning. Initially, only agencies working with community members were included in local and regional cluster coalitions. Later, however, realizing the importance of integrating peace and security issues in the region, organizers decided to invite the internal security agency into the MSP at the subnational level in both countries. The MSP agenda continues to evolve in response to issues at hand. Currently, a key area of focus is the development of protocols and agreements on the management of biodiversity hotspots. By working jointly on conservation, livelihoods, and environmental security, the MSP is expanding the range of agencies that take part and increasing its influence.
The limits of membership in the MSP can become the focus of considerable debate, especially when it concerns powerful private sector actors. In Madre de Dios, while the MSP focused on comanagement of the Amarakaeri Communal Reserve, both government and the indigenous communities eventually agreed on the importance of engaging representatives of the informal mining sector, which represents a key threat to the implementation of management plans. Notably, however, some nonparticipants felt the MSP was overlapping with other management bodies and said that they chose not to join for that reason. While expanding the scope of an MSP may enable greater influence, sometimes the priority lies in negotiating and clarifying the relationship to other forums for resource governance.

Strengthening the multistakeholder platform durability requires attention to factors that sustain trust and commitment
Time is required to realize meaningful change, particularly in highly politicized environments. However, the cases reveal that neither a mandate from government, the initial enthusiasm of participants, nor the support of a capable external organization ensures the institutional durability required over time to ultimately identify and seize upon opportunities for transformation. Also important are continual investment in the MSP and in the action resources of its participants, including mutual trust and relationships, to sustain commitment through times of difficulty.
Consistent funding was frequently cited as critical to durability and long-term impact. In Oromia and Chemba, uncertainty in future funding and dependence on external donors were noted as key risks. To date, the MSP in Chemba has been funded entirely by its supporting NGO, TNRF, and the district has not provided any funds. Because funding is scarce, it now convenes only once per year, undermining its ability to respond to issues as they emerge. Project funding is likely to conclude within a year, and efforts are underway to build up the capacity of the district officials to coordinate the MSP in the future. A similar situation exists in Oromia, where government relies on donor project funding channeled through NGOs to support agreed actions.
Where government has committed substantial time and resources, it offers a chance for institutionalizing new practices, but exclusive reliance on government financing can also raise concerns about the character of the MSP. In Acre, the state government's resources for the zoning commission have supported an inclusive forum, as well as important parallel processes of engagement. Noting that indigenous peoples and smallholder farmers were unable to cover the costs of travel to MSP meetings, decentralized sessions were held in all of the state's municipalities. In Pará, however, NGO interviewees noted that only those municipalities with funds to invest in effective technical teams and capacity building were able to benefit from the program. Also, funding constraints have meant that not all participants who wanted to engage could be accommodated, particularly to the detriment of nonindigenous actors who cannot cover their own costs, including travel to the state capital.
The India cases illustrate the value of an organization that provides long-term support to local MSPs and simultaneously invests in building the relationships to influence state and national policies. Apart from the particular substantive issues of debate, much of the work centers on building mutual trust. In the case of Gujarat, as the MSP's negotiation and influencing capacity increased, village federations were able to bring relevant government officials to the table. The benefits of this connectivity were confirmed from the other side of the table as well: one government participant described an important value of the MSP as making "communication with villages in border areas and interior regions easy." Lastly, members' commitment toward and ownership of the MSP were frequently cited as critical factors in determining durability Ecology and Society 27(2): 2 https://www.ecologyandsociety.org/vol27/iss2/art2/ and impact, and these qualities are often seen as consequences of trust, relationships, and achievement. In Oromia, for example, participants described seeing the influence on natural resource management decision-making as motivation to stay engaged. In the Tana-Kipini case, long-term commitment of MSP participants was cited as a key enabler of impact. Organizers need to pay attention, however, to the depth of that commitment beyond direct participants in the MSP. In the case of Madre de Dios, indigenous organizations are mandated by law to participate in co-management of the Communal Reserve. However, focus groups in two communities in the reserve's buffer zone revealed that inhabitants knew very little about the MSP, its functioning, or how they were represented in it. This situation underlines the importance of understanding the value of an MSP, the challenges it faces, and, in turn, its potential for impact, from the multiple perspectives of actors at different levels.

DISCUSSION
The modified institutional analysis and development framework employed in this analysis enabled a structured comparison of cases, distinguishing findings and lessons that relate to each element of the framework. These elements are, namely, the relation between the design of MSPs, conceived as structured action arenas, and the broader governance context; how processes for inclusion are designed and implemented; and the degree to which adaptive learning is fostered to improve the MSP's performance. Here, we draw out further cross-cutting insights from this analysis, illustrating how the conceptual categories of the framework and our emphasis on the dynamics of power within (and beyond) these MSPs can be applied to support reflective practice.
Our analysis shows the importance of assessing the context, including attributes of the resources, actors, and governance arrangements, in the design of MSPs. These attributes should influence decisions on the appropriate role for the MSP as a purposefully designed action arena, and its potential objectives in establishing new patterns of stakeholder interaction and resultant decisions regarding resource use and allocation. This insight accords with the conclusions of a systematic review by Sarmiento Barletti et al. (2020b), who argue that trying to address a resource-related challenge by implementing a new way of doing things without taking into account existing ones (both formal and informal) may hinder an MSP's ability to reach its goals. In a study of land-use planning in Laos, Suhardiman et al. (2019) similarly found that when local land-use planning processes apply a strict approach to land categorization without taking into account existing knowledge of overlapping boundaries between forest and agricultural land, there is a high probability that the defined land-use plan will not be implemented. The cases in our analysis include MSPs that were established by government mandate as well as those that grew from NGO initiatives, some that were entirely financed by state budgets and others that primarily relied on external funding. With appropriate attention to the particularities of the context, each of these arrangements can deliver positive outcomes.
The cases also illustrate how building action resources of knowledge and social capital may lead to successful outcomes. Social network analysis in multistakeholder governance systems has demonstrated the importance of social capital as an action resource that contributes to influence (Schiffer et al. 2010). The two cases in India illustrate the gradual, bottom-up approach to weaving relationships of trust within a context of strong social hierarchies and gaps in effective local governance. Village federations first established by FES with a narrow, local focus provided a foundation for later engagement with multiple stakeholders in MSPs enabling knowledge exchange, negotiation, and landscape governance. Long-term and repeated engagement between FES, local federations, and various government agencies both before and after the initiation of the MSP allowed enough time for building trust and capacity of different stakeholders. In Chemba, a number of projects undertaken in the district, including training on land issues, best practice exchange visits, and help with forming village land councils were able to build the capacity of village governance institutions to take up responsibilities in the MSP and influence its direction.
Whether conceived with such explicit aims or not, MSPs have the potential to challenge or reinforce existing power relationships. This aspect concerns both the immediate dynamics within the scope of the MSP's activities and possibly in interactions that go beyond individual MSPs, within the broader realm of the systems they seek to address. Thus, the work of designing and strengthening MSPs cannot be divorced from the values and objectives of those who facilitate and support these processes (Dewulf et al. 2019). In some instances, this aspect can lead to strikingly different assessments of failure and success. It helps to make these criteria explicit. The Pará case was a top-down organizing process that addressed the contextual factors that its organizers thought were important to the MSP goals, which were in turn based on a national program to prevent deforestation. However, in doing so, they missed other key factors surrounding conflicts over land in Pará, including the land tenure priorities of subsistence farmers.
The cases further illustrate how shifts in the overall governance context can markedly influence the effectiveness of an MSP and the equity of its outcomes. The most striking case is Acre, where more equitable zoning was favored by political conditions in Brazil over the past two decades with the growth of participation mechanisms and indigenous rights and environmental movements, including a strong multiactor environmental alliance within Acre. The initiative now faces a much more difficult scenario with the Workers' Party election loss, both in Acre and at the national level, to the far-right party of Jair Bolsonaro.
Similarly, rules and characteristics of the MSP influence the patterns of interaction and hence the chances in inclusive dialogue and equitable decision-making. This effect is illustrated, for example, by efforts to address imbalances in access and deployment of technical knowledge (Acre, Gujarat) and to build the capacity of typically marginalized groups to participate effectively in dialogue activities, including preliminary efforts to address barriers to women's participation (Oromia, Chemba). In some instances, the aim of addressing historical inequities has been built into the formal mandate and procedural rules of the MSP, as seen in the "ethno-zoning" process in Acre, the requirements for strong civil society representation in Pará, and the roundtables focused on conflict management in Madre de Dios. https://www.ecologyandsociety.org/vol27/iss2/art2/ Thus, differences in outcomes cannot be assigned simply to either the context or the characteristics of the MSP alone; it is a complex interplay. In recognizing this dynamic, MSP organizers must address multiple tensions, working simultaneously to "fit within" the prevailing governance context to ensure relevance and influence on key decision-making processes and also to "purposefully disrupt" existing patterns of stakeholder interaction in support of more equitable and sustainable forms of resource use and benefit sharing. Legitimacy of the process may be judged on very different grounds from the perspectives of indigenous resource users, private sector associations, or government bodies at state or local levels. Adaptive learning processes should aim to integrate these multiple perspectives, addressing both the institutional design questions (the "fit" of the MSP in relation to its context) and the dialogue process questions (the way the MSP's rules and principles are implemented in practice).

CONCLUSION
Much of the enthusiasm fueling the promotion of MSPs in land and natural resource governance stems from their perceived value in reaching agreement on sustainable resource use and enabling new forms of collaboration to support effective policy implementation amid many competing goals (Ros-Tonen et al. 2018). Whether or not conveners ascribe an explicitly political agenda to these efforts, they function in political environments. This milieu presents challenges to evidence-based analysis of MSPs because the context and actors' motivations are rarely well documented and vary widely. Most of the documentation comprises project or "grey" literature, often with a promotional bent, rather than independent analysis. Even less common is comparative evidence with reference to a common framework.
An advantage of our comparative analysis is that researchers have drawn upon their detailed familiarity with the cases and the actors involved to tease out divergent perspectives on the history, challenges, and achievements of each MSP. We drew additional lessons from discussion among the authors familiar with each case study and iterative reference to the framework. Our approach complements the findings of a systematic review of the published literature (Sarmiento Barletti et al. 2020b). An important contribution is the ability to trace the evolution of these efforts and characterize differential outcomes based on actors' own perspectives, particularly for a domain of social action in which the contextual complexities make it exceedingly difficult to define robust natural experiments (e.g., different approaches in 'comparable' states). Our study also complements a multicase comparison of organizers' own perspectives on landscape MSPs in the forestry domain, which concluded that the transformation potential of such MSPs depends on explicit strategies to identify and address power differentials, ensure meaningful participation of under-represented groups, and plan for longer term engagement (Sarmiento . We have sought to draw lessons from this comparative analysis that can help to guide future efforts to design MSPs at landscape scale and to aid organizers of existing MSPs to reflect, improve, and adapt their efforts for increased success. We structured these lessons to reflect this practical intent: addressing MSP design in relation to the governance context ("designing for context"), designing inclusive processes to address power inequities ("designing for inclusion"), and supporting adaptive learning to improve and expand influence over time. The structure also mirrors the approach of Kusters et al. (2018), whose framework for monitoring and evaluating MSPs is structured to enable "looking ahead" (to define appropriate goals), "looking inward" (to assess equity and effectiveness of the process), and "looking back" (to draw insights on successes, failures, and requirements for future adaptation).
Our study also shows that the modified institutional analysis and development framework we applied can flexibly capture a wide range of contextual factors, MSP attributes, and outcomes as a basis for comparative analysis. The cases clearly illustrate how MSPs at landscape level are able to influence the patterns of stakeholder interaction, as measured by shifts in collaboration and trust, conflict management, and collective action for conservation. The limitations of a retrospective analysis, as well as the limited duration of a number of the cases, means that we did not attempt to assess MSP contributions toward longer term shifts in resource status, attributes of resource users such as poverty and assets, or enduring characteristics of the governance context. Evaluating these causal connections, illustrated by the longer term feedback loop in our analytical framework, would require a longitudinal research design with baseline measures and detailed process tracing (Befani and Mayne 2014), likely at multiple intervals.
Future research should examine such longer term causal relationships among the quality of MSP design, implementation, and outcomes related to transformational change. It should explore, in additional cases, the interplay between context and MSP characteristics in influencing variable outcomes, including through mixed quantitative and qualitative assessment tools. It should also seek to probe the conditions that allow for MSPs to achieve impact beyond the scope of the directly engaged actors, either through horizontal replication of governance innovations (as is beginning to happen in the case of block-level planning in Odisha) or cross-scale linkages (as is emerging between the Chemba case and national policy processes in Tanzania, as well as local and cross-border collaboration in the case of Tana-Kipini). Analyzing such scaling processes will help develop current understanding of the potential ways in which MSPs at landscape level may catalyze enduring governance transformation to support social-ecological resilience, and the ways in which they may fail. Building such a robust body of research is critical to continue to validate and refine evidence-based lessons for practice.
Responses to this article can be read online at: https://www.ecologyandsociety.org/issues/responses. php/13168 In Madre de Dios, Peru, the Amarakaeri Communal Reserve (ACR) is an official natural protected area (NPA) part of Peru's National System of Natural Protected Areas (SINANPE). It is decreed to protect the watershed of the Madre de Dios and Colorado rivers in the southeastern Peruvian Amazon and ensure the stability of lands and forests to maintain water quality and quantity, and an adequate environment for the development of the indigenous Harakmbut communities that inhabit this area. This landscape is situated within Manu province in the Madre de Dios region. ACR is co-managed by the state's National Service of Natural Protected Areas (SERNANP) and ECA-Amarakaeri (an indigenous organization of ten native communities located in the ACR's buffer zone). The multi-stakeholder Management Committee of the ACR was set up in 2014 to provide support to the co-management of the reserve, between the ECA-Amarakaeri and SERNANP, and to approve the Master Plan for the ACR's management.
In Acre, Brazil, economic development policies in the 1960s and 1970s aided construction of highways towards Amazon regions and granted credit incentives for cattle ranching, logging, mining, and settlements. This led to deforestation and displacement of indigenous and local populations in Acre, while large-scale farmers acquired land held by rural communities through both legal and illegal means. This resulted in conflicts in the late 1970s, and groups of indigenous and local peoples whose livelihoods depended on these forests organized themselves to resist deforestation and expansion into their territories. The MSP thus emerged from a violent history of social relations in the state and eventually became a vehicle to transform a history of land tenure insecurity and social-environmental conflicts by building positive relations between different actors. Under the leadership of the Workers' Party for 20 years, it aimed at empowering under-represented groups. Acre's Ecological-Economic Zoning commission's second phase (2003)(2004)(2005)(2006)(2007) set the goal of recognizing and protecting indigenous and local peoples' land rights by addressing deforestation drivers and including and empowering historically marginalized and underrepresented groups in the zoning process. The commission was charged with designing the map for territorial management within the state.
Similarly, in Pará, Brazil, historically, government-driven expansion of the economic frontier into the Amazon led to high deforestation rates, displacement of indigenous and local peoples, and social conflicts related to land rights. In 2011, the state government of Pará launched the Green Municipalities Program, an MSP that aims to reduce deforestation, increase the areas registered under the state's Rural Environmental Registry, and improve local governments' capacity to combat deforestation. It's goals include creating a municipal working group to combat illicit deforestation, conducting field inspections of illegal deforestation, maintaining the annual rate of deforestation below 40 km², and registering 80% of the state's area in the Rural Environmental Registry. It also aims to address some of the overlapping land tenure issues.
In Odisha, India, the Foundation for Ecological Security (FES), a domestic nongovernmental organization focused on local democratization and landscape restoration, has engaged with local communities to form village institutions and federations (informal collectives of village institutions formed by representative members from each village) at the "block" level to help communities claim their rights to forests and collectively manage commons. (The block is an administrative division for rural development below the sub-district, which typically includes around 100-150 villages.) Federations in the case study area represent all villages within the Charmalik hill range, densely forested and situated on the buffer zone of a wildlife sanctuary. The area is also rich in minerals and faces increasing pressures and threats to its ecological health from coal, aluminum and steel mining industries. The MSP was initially co-initiated in 2005 by FES and federations to reinforce and revive traditional farmer seed exchange networks. It evolved to serve as an exchange platform for good agricultural practices, a forum to address shared threats such as forest fires, and a channel for dialogue with government and other stakeholders for block-level landscape resource and development planning.
Similarly in Gujarat, India, FES has been engaging with communities since 2013 to build village-level and block-level institutions. The village federations represented in this MSP are centred around three watersheds and a hill range. It is a relatively homogenous population of small-scale farmers of the same tribe, who also depend on the forest for subsistence. Communities' level of trust in government was low, mainly because of major displacement due to the building of the Kadana dam in the late 90s. The MSP was initiated by FES, initially to bring together different communities within the watershed to collectively discuss resourcerelated issues. Overtime, it transitioned to involve government actors and other stakeholders. It serves as a communications platform to bridge the gap between government and local village institutions and for block-level planning of resource management, while maintaining its original role as a knowledge-exchange platform where different community members share and learn from each other's experiences.
In Chemba, Tanzania, (often violent) conflicts over land use between pastoralists, farmers, settlers and land-based investors increased in intensity over the last decade. Key drivers included lack of tenure security, inadequate land use planning, poor leadership by local authorities, and a lack of space to bring stakeholders together to dialogue and resolve conflicts. The majority of conflicts were over land use, including between conservation and pastoralismfarming-settlement; between pastoralism or farming and settlement; between farming and pastoralism; and between all of the above and land-based investors. Spearheaded by the nongovernmental Tanzanian Natural Resource Forum (TNRF), the MSP established in 2015 brings together representatives from around fourteen villages in the district-typically including the village chairperson of the village council, the village executive officer and a land rights monitoras well as officials from the District Commissioner's office, the District Executive Director, officials from the land, legal, migration and livestock offices, police officers, officials from the Swagaswaga Game Reserve, representatives of pastoralists and farmers from Kwamtoro division, representatives from UMAKWA (a local CBO based in Kwamtoro division), and the media.
In Oromia, Ethiopia, the top-down approaches used in land use planning and the isolated nature of past interventions aggravated deforestation and forest degradation. Agricultural expansion, migration and illegal settlements were some of the key drivers of deforestation. MSP participants noted that conflicts over land use were increasing in the ecoregion due to the growing population, which they understood as leading to land and resource scarcity. The SHARE-BER multi-stakeholder platform was established in 2014 as a component of the SHARE-BER project, a three-year project funded by the European Union. The project aimed to conserve biodiversity, ecosystem functions and services in the Bale ecoregion, and increase the livelihoods of local communities. It also sought to explore the benefits of multi-sectoral and integrated approaches to enable sustainable land use in the ecoregion. The MSP approach was based on the assumption that multi-sector approaches would address the multifaceted problems that families and communities in the ecoregion face and lead to more sustainable land use and reduce deforestation.
The Tana-Kipini Laga Badana Bush Land and Seascape (TKLBBS) covers the coastal border region of Tana-Kipini in Kenya and Laga Badana in Somalia. Established in 2016, the TKLBBS MSP aims to promote cross-border cooperation in natural resource management and to facilitate the establishment of a transboundary protected area, in a zone of high biodiversity under facing significant threats of degradation. The border is economically underdeveloped with high incidence of poverty, consequently imposing pressures on land and ecosystem services. It also has a history of conflict, increasing the need for (and the challenge of) bringing different civil society stakeholders and government departments into dialogue. The MSP includes participation by community interest groups clustered at landscape level, plus NGO technical partners and government agencies at county and national levels, such as Kenya Wildlife Services, Kenya Forestry Department, Somalia Ministry of Environment, and Ministries of Foreign Affairs of Kenya and Somalia. The persistent threat of terrorism in the region has hindered the management and full utilization of forest ecosystems. The area is still heavily militarized, and this makes most livelihood activities difficult. This situation also poses a challenge to biodiversity conservation work, including field research and engagement with communities.

INCLUSION, COLLABORATION AND TRUST
Active involvement of local governments in the Oromia MSP has strengthened inter-sectoral coordination and rule enforcement in the ecoregion. Before establishment of the MSP, local administrative bodies and agencies responsible for health or irrigation development had no direct working relationship, such as planning and implementation of interventions. FARM Africa and PCI among other NGOs have worked alongside the government on such issues, focusing on conservation and economic development. There is also an agricultural cluster of experts that include the irrigation, water and energy, land administration, environmental protection offices, and the cooperative promotion agency. The result is a clear commitment from government to multi-sector collaboration on land use and land-use change issues.
In Chemba, the MSP also cut across divisions between government sectors, and provided a space for actors with different responsibilities in land governance to come together and engage. Land councils established in 45 villages, creating space for otherwise marginalized groups to contribute to decision making on land use and management. The MSP built confidence between local communities and local authorities, reducing mistrust. As one village leader said:

"The [Chemba] MSP is a bridge that connects villagers at the grassroots level with high decision makers. It helps the village authority to understand their responsibilities on land issues better."
Over time, the government took more responsibility in convening and leading the MSP, helping to institutionalize it as a forum for dialogue and action. The MSP has subsequently been replicated in other districts.
In Pará, the MSP enhanced multi-stakeholder dialogue between the private sector, government institutions, NGOs and the Public Ministry Service-groups that rarely coordinated previously. This produced a reorganized rural cadaster (environmental register) better addressing municipal contexts and priorities. But grassroots organizations and communities report being excluded from the process, resulting in much more critical assessments of progress from the perspectives of these groups.
In Gujarat, continuous participation in the MSP improved cooperation between different actors, particularly between various government agencies and village federations, especially as the MSP gained recognition and secured participation of higher-level officials such as state ministers. This also enabled local communities to raise livelihood issues related to the environmental commons on issues such as water allocation and access. Because FES, the supporting NGO, has been active within Gujarat for years, their previous commitment played a role in fostering trust among the community members. This also enabled FES to convene different actors and reduce community skepticism towards government officials and vice versa.
Conversely in Odisha, after years of participating in the MSP led jointly by the communities and their NGO supporter, the district administration and agriculture department adopted the responsibility of organizing and convening it and replicating the model in other districts. Previously, local federations had displayed agency in organizing the MSP events, including preparing the agenda to highlight priorities concerning resource trends, livelihoods and community rights, which enabled significant peer-to peer learning among communities. The new arrangement gave official recognition to the platform to set the official block development agenda. Yet, communities' roles were reduced from organizers to invitees, which poses a potential risk to local voice and agency.
Roundtables were organized around specific issues in Madre de Dios, such as the construction of a road across the communal reserve. Participants appreciated the MSP's success addressing actors' divergent points of view, enabling them to discuss and reach agreements. However, the lack of interest and participation from different sectors in the subnational and national governments hindered the effectiveness of the MSP. By contrast, the Tana-Kipini MSP has provided an important link among transboundary communities, enabling diverse groupings in terms of tribe, clans, interests and areas and nationalities to sustain a dialogue on environmental sustainability and livelihoods. The MSP is recognized for helping to build institutional capacity through its members representing various stakeholder groups and sectors at different levels, fostering engagement in decision making around transboundary issues.

CONFLICT MANAGEMENT
In the case of Tana-Kipini, in a region with a history of violent civil conflict, the dialogue process presented opportunities for communities to move beyond disputes over resource use and management, moving towards a more collaborative vision involving cross-border trade and development. The MSP has been active for only a few years but has initiated peaceful dialogues joining Somali and Kenyan counterparts. At the onset of the project, mistrust was such that it was not possible to link the different partner ministries in joint management of the project. However, once the multi-stakeholder formation process was implemented at cluster level, different sector partners were able to relate better through specific activities jointly undertaken. The platform has now developed an action plan to ensure continuity of planned activities. Improved working relationships among MSP members has accelerated project implementation, and generated ripple effects beyond the initial biodiversity agenda. Some of the spinoffs were increased internal and cross-regional trade, and the revival of cottage industries such as honey production and horticulture.
In Madre de Dios, the establishment of the platform was itself a significant achievement, after 12 years of convening attempts that were hindered by conflicts. Interviews with MSP participants revealed that a positive outcome has been the MSP's function in resolving conflicts in the communal reserve. The platform established dialogue roundtables for conflictive situations, such as the construction of a road across part of the reserve's buffer zone. This goes beyond what is mandated by law, but most participants agree that the resolution of conflicts is a vital role for the MSP. Interviewed participants appreciated the MSP's success in addressing divergent points of view that actors may have had, who were able to meet, discuss and reach agreements. Nevertheless, some respondents noted that the actions of the MSP and the comanagement of the communal reserve in general benefitted some communities more than others, and that principally its approach failed to address long-standing inequities or positively impact their livelihoods.
The Chemba MSP was instrumental in resolving more local conflicts between land users. For example, a conflict existed between Handa village and Swagaswaga Game reserve was addressed by amending the Game reserve boundaries, returning part of the terrain to the village. Not only did this resolve the conflict but it gave the actors greater insight and understanding of their (at the time) opposing parties, helping to build greater trust and solidarity between them. In addition, most of the villages established village land councils for resolving local land conflicts, following the commitments made during MSP meetings. As one participant said: "To me, the MSP has been an eye opener. In the beginning I did not know how to deal with land use conflicts, how to address the issues, whom to talk to, but after the MSP I now know whom to talk to and where to go. As for my organization, the MSP has helped us to be recognized in the district where, otherwise, we would not be known." However, the platform failed to resolve the conflict between pastoralists and farmers in some areas involved in the joint village land use planning, such as the Lahaki villages. Participants cite district leaders' lack of clear vision, understanding, neutrality, and willingness to resolve the issue.
The Chemba MSP also took up the complex issue of women's land ownership and access rights, promoting more equitable distribution and access to resources and helping to avert potential localized conflict. Village communities' understanding of gender issues has improved, including that gender equality is a responsibility of both men and women. As a result, gender issues are now included in local development plans. The MSP has assisted women to retain family assets after death of their husband, or in times of separation or divorce. Land inheritance laws have also been discussed; as there is no single law on inheritance in Tanzania, the MSP provided advice on relevant legislation.
In the two Brazilian cases, conflict management efforts were more limited. The Acre forum was activated specifically to agree a plan for ecological and economic zoning, seen as essential because of historical conflicts over land and natural resources. While efforts were made to cultivate the notion of florestania or 'forest citizenship', less attention was paid to monitoring compliance with the plan or resolving subsequent disputes. In the case of Pará, the focus was on negotiations between the private sector and various government bodies at state and municipal levels. Grassroots communities and organizations were excluded from participating, however, and found challenges to access the land registry supported by the forum, which increased conflicts and uncertainty over access to land, cited as an ongoing driver of deforestation.

COLLECTIVE ACTION FOR CONSERVATION
The Pará program is described by its participants as having sophisticated and integrated mechanisms to combat deforestation, which led to five out of 17 municipalities being removed from the national blacklist of most deforesting municipalities. MSP participants and nonparticipants noted that at least part of the reduction in deforestation rates can be attributed to the MSP, and in particular collective action by political and economic elites (but excluding indigenous and local communities). Another portion is attributed to the state's command and control measures, including increased enforcement activity. Participants noted that this was because activities were tailored to each municipality's context. However, a fundamental flaw is that it failed to address the regularization of collective territories for indigenous peoples. NGOs and local non-participants noted that the land issue and the need for a fair and organized agrarian reform in the Amazon remains critical to address the root causes of deforestation.
In Gujarat and Odisha, gains in agricultural productivity and income stemmed in part from earlier efforts to establish village institutions and local federations, but participants claim these benefits have increased through influence of the broader MSP. Long-term and repeated engagement between the NGO, local federations and various government agencies gave way to parallel outcomes below the block level, eventually contributing to larger landscape-level goals.
In Odisha, a traditional farmer seed exchange network has been revived, and financing was secured from the district administration for enhanced irrigation. Moreover, communities reached an agreement on new rules to regulate open grazing. As a result, communities were able to cultivate a second annual crop, improving farmer incomes. In Gujarat, the MSP has also influenced the direction of public fund investments towards issues collectively voiced by communities. Community rights to forests and non-timber forest product (NTFTP) extraction have also been recognized, bolstering access to key livelihood resources.
In Oromia, MSP participants noted that major outcomes include decreased deforestation and degradation of forests using the multi-sector taskforce for rule enforcement established through the MSP, as well as increased livelihood diversification. Interviews with MSP participants and non-participants also noted that the MSP increased awareness among local communities on sustainable land use and helped increased the use of alternative energy sources such as energy-saving stoves.
On the Kenyan side of the Tana-Kipini landscape, agroforestry and rehabilitation of protected areas and farmlands have been initiated. Following identification and promotion of local actions through dialogue, farmers have established woodlots and planted trees around their farms, and protected areas were enriched through natural regeneration and replanting of degraded areas. Training to build awareness of the 2013 Wildlife Act resulted in reduced cases of human-wildlife conflict, and a new community monitoring system was introduced. On the Somalian side of the landscape, rainwater harvesting techniques have been piloted, honey value chains revived. Biodiversity status has also been assessed, resulting in a vision and roadmap for the establishment of a conservation area in Somalia. Landscape conservation committees have used the dialogue processes to engage various stakeholders in developing rules and ensuring higher participation in safeguarding natural resources. Participants cite the MSP's strong point as enabling the emergence of leadership in various conservation and development sectors that are deemed crucial for livelihood improvement.