REDD + policy networks : exploring actors and power structures in an emerging policy domain

Policy making is often neither rational nor solution-oriented, but driven by negotiations of interests of multiple actors that increasingly tend to take place in policy networks. Such policy networks integrate societal actors beyond the state, which all aim, to different degrees, at influencing ongoing policy processes and outcomes. Reducing Emissions from Deforestation and Forest Degradation (REDD+) can be considered as such an emerging policy domain, in which actors cooperate and conflict in network structures, build coalitions and try to control information and finance flows relevant for REDD+ decision making. This special feature is the result of an extensive comparative research effort to investigate national level REDD+ policy processes and emerging policy networks. This unique collection of seven country cases and a comparative study provides evidence on how power, coalitions, and different interactions among actors in policy networks enable the transformational change required for an effective, efficient, and equitable national REDD+ design. However, as we will see in most of the cases, where the dominant coalitions fail to tackle the drivers of deforestation and forest degradation, they also hinder such major policy reforms required for REDD+. The aim of this editorial serves four purposes: first, we provide an argument about “why” policy network analysis is highly relevant to the study of REDD+ policy processes; second, we explain “how” policy network analysis is used in this special feature to investigate policy processes in this domain; and third, we explore the “so what?” or how a policy network lens helps us understand the political opportunities and challenges for REDD+. Finally, we provide an outlook for the relevance and future research design of policy network analysis when applied to REDD+ and to policy network structures more broadly.


INTRODUCTION
Policy processes are inherently linked to power struggles, where policy making is driven by political disputes and interest competition rather than being a rational, solution-oriented process (Mayntz 1993(Mayntz , 2001)).This becomes particularly visible in complex, multiactor, multisector, and multilevel policy processes, typical of environmental policy arenas.Here, negotiation of interests often takes place in emerging policy networks that integrate societal actors beyond the state, each aiming at influencing ongoing policy processes and outcomes.
In recent years, the analysis of policy networks, by which we mean the patterns of interactions and resource interdependencies between policy actors (Smith 1997), gained increased attention by scholars of environmental governance (Bulkeley 2000, Weible and Sabatier 2005, Sandström and Carlsson 2008, Fawcett and Daugbjerg 2012, Gale 2013).In policy networks, power, understood as the ability to influence policy outcomes (Arts 2003), is exercised, in large part, through different forms of interactions among a variety of policy actors (Daugbjerg 1998, Bulkeley 2000, Weible and Sabatier 2005, McClurg and Lazer 2014).Although today there is an increased understanding of the role of interactions, e.g., the exchange of information or financial resources, and policy network approaches are becoming more common, few policy network analyses have employed a comparative design to investigate policy processes (Knoke et al. 1996, Kriesi et al. 2006, Broadbent 2010, Broadbent and Vaughter 2014), in particular in developing countries.This special feature is the result of an extensive comparative research effort to investigate national level policy processes and emerging policy networks around the forest-related climate change mitigation mechanism, known as Reducing Emissions from Deforestation and Forest Degradation (REDD+) in seven countries.This collection of seven country cases and a comparative study provides evidence about how power, coalitions, and different interactions among policy actors shape national REDD+ policies.Despite the variety of network related issues addressed in this special feature, all papers presented use a shared analytical lens and build on a common research methods design, contributing to fill the research gap for more systematic policy network research (Knoke et al. 1996, Kriesi et al. 2006, Ingold and Fischer 2014).Such an approach provides increased explanatory power of the individual case studies and allows for comparative analysis of policy networks.
The aim of this editorial serves four purposes: first, we provide an argument for "why" policy network analysis is highly relevant to the study of REDD+ policy processes; second, we explain "how" policy network analysis is used in this special feature to investigate policy processes in this domain; and third, we explore the "so what?," i.e., how a policy network lens helps us to understand the political opportunities and challenges associated with developing an effective, efficient, and equitable national REDD+ policy design (Angelsen andWetz-Kanounnikoff 2008, Angelsen 2010).Finally, in an outlook section we draw some implications on the main findings from the papers of this special feature for the relevance and future research design of policy network analysis.http://www.ecologyandsociety.org/vol19/iss4/art29/THE "WHY" OF POLICY NETWORK ANALYSIS ON REDD+ REDD+ became part of the international climate change negotiation framework at the 13th Conference of the Parties (COP 13) of the United Nations Framework Convention on Climate Change (UNFCCC) in Bali in 2007.Since then a number of tropical forest countries have been engaged in the design and implementation of a national REDD+ architecture as part of what was then considered a cost-efficient response to the problem of climate change (Stern 2007).In order to achieve effective emission reductions, REDD+ will require transformational change, e.g., incentives for major policy reforms in various sectors of the economy that drive deforestation and forest degradation (Kanninen et al. 2007, Angelsen andMcNeill 2012).A variety of policy actors, including government agencies, domestic and international NGOs and civil society organizations, "green" business, research and intergovernmental organizations engage in realizing such changes in domestic policy arenas.At the same time, some interest groups are trying to maintain existing structures that fuel deforestation and forest degradation by seeking political support and building coalitions to realize their economic interests (Ross 2001, Nepstad et al. 2013).
National REDD+ policy progress has so far been slower than expected and uncertainty on global and national policy directions, conflicting interests, real and perceived tensions between REDD+ and economic development objectives, and considerations about distribution of costs and benefits have been identified as some of the major challenges (Angelsen and McNeill 2012, Luttrell et al. 2013, Korhonen-Kurki et al. 2014).In the absence of obvious short-term win-win outcomes for all, at the national level such tensions emerge especially during policy negotiations about the details of the national REDD+ strategies (Brockhaus andAngelsen 2012, Brockhaus andDi Gregorio 2014).To understand what hinders or enables the transformational change required for realizing effective, efficient, and equitable REDD+ strategies it is important that we understand the processes that underlie such challenges.
Increasingly, research has adopted the lens of policy networks to investigate complex policy domains linked to social-ecological systems and environmental policy problems such as climate change and forest governance (Bulkeley 2000, Gallemore andMunroe 2013, Gale andCadman 2014).Some of the reasons for these developments are explored below, with specific reference to the REDD+ policy domain.A policy network approach can help to investigate some key determinants of policy making in this domain.Among other issues, it provides a conceptual lens to: 1. Investigate in detail features of policy processes in multiactor policy domains 2. Explore how resource exchange, and pooling of resources, is used as a basis for political negotiation

3.
Understand how policy actors exercise power and influence through interactions in policy processes

Analyze the form and role of policy coalitions in influencing policy outcomes
Multiactor policy domains A policy network approach provides a way to integrate and investigate in detail policy actor systems that are broader than just formal institutional structures of the state (Kenis and Raab 2008).It is therefore a concept that operates at the meso-level [1] , focusing on interactions among organizations (Rhodes 1997, Evans 2001).For example, realizing emission reduction through REDD+ requires policy action and reforms in a large number of sectors including forestry, agriculture, and infrastructure development, which affect different business interests, smallholders, local communities, and government plans (Di Gregorio et al. 2012).It also requires planning, participation, and compliance by actors operating at different jurisdictional levels (Pacheco et al. 2011, Korhonen-Kurki et al. 2012).With the expansion of the role of business, civil society, and international actors in national policy making, diverse forms of consultation have become institutionalized and policy decisions are no longer the exclusive domain of state actors.This is, for example, reflected in the decision making around the operationalization of REDD+ safeguards (Roe et al. 2013, Arhin 2014, UNFCCC 2014).

Resource exchange for political negotiation in policy networks
In these emerging policy domains, interactions of different actors and interests are used to mobilize and pool resources and support, in order to shape policy outcomes and realize their interests (Kenis and Schneider 1991).Material and nonmaterial interdependencies reveal how organizations bargain with one another to affect policy outcomes (Laumann and Knoke 1987).In fact, social action can be understood as a negotiating process in which actors interact, exchange resources, and use them to realize their interests (Hanf andScharf 1978, Coleman 1990).These resource exchanges are used to consolidate existing policy positions and maintain established power structures as well as facilitate policy changes (Friedland and Alford 1991, Börzel 1997, Marsh and Smith 2000).Financial resources for REDD+ come predominantly from international sources (multilateral and bilateral) and are channelled through national government, NGOs, and civil society sectors to a variety of actors at different levels.Other resources are provided in-kind, such as scientific knowledge and expertise.
Given the high level of technical knowledge required to set up the institutional structures, e.g., forest inventories and monitoring systems, to support REDD+ implementation, governments rely in part on nonstate actors including international and national research and civil society organizations to mobilize such expertise.

Exerting power and influence through interactions
To understand progress in national REDD+ policy making, such as why, for example, substantial policy reforms in the sectors driving deforestation are slow to be realized, we can start by investigating the structure of power that underlies national REDD+ policy networks.Policy networks "reflect past power distributions and conflicts," yet, at the same time they "shape present political outcomes" (Marsh and Smith 2000:6).Policy networks, therefore, reveal the institutionalization of power relations of actors that are part of the network and the constraints formed by the broader political context.By analyzing the form that policy networks take, and the relations between different actors and their positions in these policy networks, we can gain a deeper understanding of policy actors' roles and power in a particular policy domain and can infer possible effects on policy http://www.ecologyandsociety.org/vol19/iss4/art29/outcomes (Knoke 1990, Marsh and Rhodes 1992, Knoke et al. 1996, Broadbent and Vaughter 2014).
For example, the papers in this special feature: identify whether REDD+ policy networks are dominated by a few very influential actors, as opposed to representing an inclusive and diffuse distribution of power; identify those actors that facilitate, control, or restrict information flows within and among coalitions; investigate who these central actors are and whether they support business-as-usual or policy reforms; these are some of the ways in which the policy network lens can gain deeper understanding of policy processes.Marsh and Rhodes (1992:267-268 al. 2011, Di Gregorio et al. 2012, Indrarto et al. 2012).

Policy coalitions for change: ideas and interactions
Policy network approaches can identify policy coalitions, and how these influence policy change (Matti and Sandström 2011).In this sense, networks are not just the reflection of institutional structures and are not static (Marsh and Smith 2000).For example, policy actors recruit like-minded allies to help form powerful coalitions, they argue with one another to influence the ideas of other actors, and bargain and negotiate to break up networks and underlying power structures (Marsh and Smith 2000, Mische and Pattison 2000, Weible and Sabatier 2005, Ingram et al. 2014).Policy change is most often driven by coalitions of actors that share similar beliefs, values, and policy objectives (Österblom and Bodin 2012).To translate their ideas into public policy, coalitions have to coordinate and interact (Sabatier and Jenkins-Smith 1993).Combining the analysis of relations between actors with information about their ideas and beliefs provides a useful way to identify political coalitions and assess the power and dominance of these coalitions, and their position in policy networks (Matti andSandström 2011, Leifeld andHaunss 2012).Where powerful coalitions are advocating for policy reforms, such reforms are more likely to occur (Fischer 2013).Within the REDD+ policy domain this would translate, for example, into more consistent and effective progress in the development of REDD+ strategies and changes in policy directions from business-as-usual to tackling the major drivers of deforestation and forest degradation (Brockhaus and Angelsen 2012, Brockhaus et al. 2013, Babon et al. 2014).
In summary, the policy network lens can be useful to characterize single organizations that are part of the policy domain, while at the same time providing a complete overview of network structures.The combination of detail and breadth allows us to understand meso-level policy processes, from which one can develop targeted policy recommendations for effective policy changes for REDD+ that take into account the structural constraints and opportunities, as well as the potential of policy actors to facilitate policy change.However, it is important to note that not all of the actors forming the policy network are elected nor necessarily adhering to any institutional requirements, and so the issues of transparency and legitimacy become evident.

THE "HOW" OF POLICY NETWORK ANALYSIS IN REDD+
The policy research component of the Global Comparative Study on REDD+ led by the Center for International Forestry Research, uses policy network analysis as one of its central theoretical and methodological approaches (Brockhaus and Di Gregorio 2012).
By analyzing policy networks to investigate actors relevant in REDD+ policy processes, their relations, and the structural conditions in the specific REDD+ policy arenas, we aim to understand the political structures that support or hamper transformational change and REDD+ policy progress.We explore how resistance can be overcome to build a carbon effective, cost efficient, and equitable REDD+ regime.
This special feature presents some of the results obtained from data collected over one thousand interview hours, focusing on the national level policy network analyses for seven countries (Brazil, Cameroon, Indonesia, Nepal, Papua New Guinea, Tanzania, and Vietnam), one study at the subnational level in Indonesia, and a comparative study.Although each study in this special feature followed the same research design, each one focuses on distinct questions around national REDD+ policy processes and uses different techniques and social network measures.Countries were purposefully selected based on their relevance for and early engagement with REDD+, the presence of pilot projects, and the presence of established partnerships to conduct the research.At the time of data collection (2010)(2011)(2012), all countries included were in the phase of institutional and pilot project set-up, i.e., REDD+ Readiness Phase, or moving into the REDD+ implementation of policies and measures (Meridian Institute 2009).

Boundary setting and policy events
The first step in the policy network analysis was to identify the policy actors that are part of the national REDD+ policy domains.An initial identification of domain actors was based on two earlier research outputs, i.e., a country profile and the analyses of media articles on REDD+, and country researchers' knowledge (Brockhaus and Di Gregorio 2012, Brockhaus et al. 2012, Di Gregorio et al. 2013).The complete list of actors was then validated by a panel or five to six experts in each country.To set the boundaries, only "core actors" were asked for interviews.We identified core actors using the criteria of relevance as: an organization that defines itself, and that is perceived by others, as a part of the national REDD+ policy domain.The criteria of relevance suggests that other members in the policy domain need to take into account a specific actor when they make decisions, even if the first is not directly involved in decision making.
For each country, a full roster of all relevant organizations (i.e., actors) was drawn up, and these organizations were then contacted to take part in a survey.We recorded respondents' perceptions of the influence of listed policy actors, a measure http://www.ecologyandsociety.org/vol19/iss4/art29/called "perceived influence" or "reputation power" (Knoke et al. 1996, Kriesi andJegen 2001) and of five types of interactions: exchanges of financial (and in-kind material) resources, general and scientific information, collaborative ties, and conflictual interactions.
Research teams also identified up to five key national REDD+ policy events.We defined a policy event as "a critical, temporally located decision point in a collective decision-making sequence that must occur in order for a policy option to be finally selected" (Laumann and Knoke 1987:251).This includes both policy proposals being discussed and actual major policy decisions taken in the relative country.The list of policy events was used in the survey to ask whether policy actors participated in these events, in which role, and the extent to which they were satisfied with the outcomes (Fillieule and Jiménez 2003).
Some limitations of the methods need to be highlighted.First, the policy network analysis conducted investigates network structures at one point in time, the time of the survey.Policy processes are, however, dynamic processes.Although policy networks analysis suggests that network structures tend to be stable over time, this has also been an area of contention (Marsh and Smith 2000).In particular, "new" policy domains, such as REDD+, are likely to be more dynamic than long established policy domains.We need therefore to interpret the results of the policy analysis within the temporal context of the survey.Second, the response rate of the survey varied from 56% to 100%, which means that it was difficult in most countries to obtain information for all policy actors identified.Such missing information can impact social network measures, in particular when very relevant actors did not participate in the survey (Burt 1987, Wasserman and Faust 1994, Borgatti and Molina 2003).These limitations are recognized and to reduce the bias of missing responses the papers include the observed incoming ties of nonrespondents when measuring reputational power (Costenbader and Valente 2003).
In the comparative studies we assess the extent of missing data in all the case studies (see Appendix in Brockhaus and Di Gregorio 2014).

Survey tools and data collection
Each country team administered an organizational survey and, where possible, a semistructured interview.Both the survey and interview were undertaken with a senior representative of each organization that was involved in or had good knowledge of national REDD+ policy processes.

Organizational survey: stances, networks, and protest events
The survey elicited responses in a fixed choice format, including questions about: (1) the position of organizations on key REDD+ issues; (2) identification of the actors that they considered particularly influential in the REDD+ policy domain; (3) the five different types of interactions with other actors in the policy domain, i.e., exchange of information, providing and receiving funds and in-kind resources, scientific information, collaboration, and disagreement; (4) participation in five previously selected REDD+ policy events and protest events; and (5) characteristics and resources of organizations (see Appendix 1).
To elicit information on the position of organizations on key REDD+ issues, position statements (or stances) were formulated to facilitate a response either in agreement or disagreement (non-neutral statements).Position statements covered key issues debated in international and national REDD+ circles, REDD+ specific elements such as sustainable forest management and enhancement of carbon stocks, the delivery of REDD+ cobenefits, e.g., poverty reduction and biodiversity, governance issues, prominent REDD+ policy challenges, and the role of science in policy, for a total of 35 stances.

Semistructured interviews
The semistructured interviews were based on a guide of openended questions, which encouraged respondents to talk in some depth about four main topics.They were asked to discuss: (1) the interest of the organization in the REDD+ policy domain; (2) their perception on what the main policy challenges and opportunities are in REDD+ policy making, with particular attention given to governance aspects; (3) the dynamics and effectiveness of the consultation processes linked to the development on the national REDD+ strategy; and ( 4) their assessment of national REDD+ policy processes (see Appendix 2).The interviews were recorded and transcribed for analysis.Analysis of transcripts was done with open coding techniques using the qualitative analysis computer software package Nvivo (Mayring 2004, QSR 2012).

Network analysis
Policy network analysis builds on quantitative social network analysis techniques, a method used to investigate social structures (Börzel 1997, Wasserman andFaust 1994).There are three main levels of analysis: (1) network measures investigate overall structural features of a network, (2) group level measures investigate features relating to the composition and characteristics of subnetworks formed by groups of actors, and (3) actor level measures assess the positions and roles of specific actors within a network (Scott 2000, Prell 2012).
Network level measures such as cohesion, density, or centralization characterize the network as a whole.They can indicate a number of different tendencies in networks.For example, highly centralized policy networks, such as one centered on a specific government agency, may facilitate coordination, but are also likely to be highly exclusive, marginalizing dissenting forces (Diani and McAdam 2003, Sandström and Carlsson 2008, Bodin and Crona 2009).Networks with low cohesion characterize networks where subgroups are disconnected.They indicate fragmentation, obstacles to collective action, and likely low levels of consensus (Granovetter 1973, Borgatti and Foster 2003, Bodin and Crona 2009).Network level measures are particularly useful to compare across the different network relations within a policy domain, e.g., REDD+.For example comparing between information and collaboration networks may reveal particular obstacles that hamper another network relation (Heimeriks et al. 2003, Weible and Sabatier 2005, Saunders 2007, Bushley 2014, Gebara et al. 2014).
Actor level measures, e.g., betweeness and in-degree centrality, etc., "zoom in" and investigate positions and roles of specific policy actors in a network.As an example, high levels of centrality of an actor, i.e., in-degree, which indicates the total number of ties of an actor, or betweeness, which measures how often an actor appears on shortest paths between other actors in the network, or high brokerage scores, referring to actors that act as mediators between two other actors in a network, characterize actors that http://www.ecologyandsociety.org/vol19/iss4/art29/are able to exert influence over others in the network, control resource flows between other actors, and might be better situated to, e.g., access information or other resources (Gould and Fernandez 1989, Degenne and Forsé 1999, Scott 2000, Burt 2004).These measures were applied in most of the papers in this special feature (Babon et al. 2014, Bushley 2014, Dkamela et al. 2014, Gebara et al. 2014, Moeliono et al. 2014, Rantala and Di Gregorio 2014).Actor level measures are also used to investigate associations and causal relationships between relational and other variables.Causality can be investigated using approaches such as quadratic assignment procedure and Exponential Random Graph Models, as applied in two of the papers in this special feature, to model organizational collaboration (Lusher et al. 2012, Gallemore et al. 2014, Moeliono et al. 2014).
There are many different ways to analyze subgroup structures in networks.Blockmodeling techniques rely on structural equivalence, grouping actors that occupy similar structural positions in a network, meaning that they are connected in the same ways to related actors (Wasserman and Faust 1994, Doreian et al. 2004).Combined with measures of influence of actors, blockmodeling can help to determine actor groups that dominate and those that are more peripheral (followers or marginalized) in a policy domain (Kriesi and Jegen 2001).This technique was applied in the global comparative paper in this special feature to compare networking patterns across countries and investigate how different political structures impact REDD+ policy development in the seven countries (Brockhaus and Di Gregorio 2014).
In addition, combining network measures with other attributes, nonrelational information about actors, and more detailed qualitative information from interviews provides a way to explore networking in more depth, including, for example, to investigate the meaning of relations, e.g., based on shared ideas or beliefs (White et al. 2007, Crossley 2010).Such analysis underpins the papers in this special feature that investigate homophily in policy coalitions (Moeliono et al. 2014, Rantala andDi Gregorio 2014).
Finally, social network analysis can be used to explore affiliation networks, which investigate the relations between two different sets of categories, e.g., actors and policy events (Borgatti andEverett 1997, Carrington et al. 2005).This technique is used to identify levels of inclusion in policy events in the Vietnam case study (Pham et al. 2014).

THE "SO WHAT?" OF POLICY NETWORK ANALYSIS ON REDD+: EVIDENCE FROM SEVEN COUNTRIES
The case studies in this special feature reflect a range of policy research questions that can be explored through a policy network lens.Key issues that were analyzed center around information flows, policy coalitions and discourses, and power in national REDD+ policy domains.We draw on these studies to illustrate how using a policy network analysis as a point of departure helps us to understand the political opportunities and challenges for transformational change in REDD+.

Information flows and information holes
Information network measures can generate insights into the extent to which organizations are working together in the REDD+ policy arena.Gallemore et al. (2014) demonstrate this by examining the extent of collaboration and information sharing between organizations in Central Kalimantan, Indonesia.They identify "discursive divides" that constrain the cross-scale collaboration necessary for effective REDD+ policy development.Moeliono et al. (2014) utilize information exchange measures to identify potential barriers to transformational change.They find that influential government agencies are fairly isolated from other actors and do not seem to be interested in seeking outside sources of information.The absence of brokers between state and nonstate actors leads to a fragmented REDD+ policy arena characterized by top-down decisions, which is unlikely to leverage the adaptive management required for REDD+.

Policy beliefs, discourses, and coalitions
In national REDD+ arenas, as in other policy domains, it is primarily policy coalitions, as opposed to individual organizations, that drive resistance to, as well as the direction of, policy change.In particular, combining evidence from interactions among actors with their ideas and beliefs can help reveal why some policy domains resist change and others facilitate policy reforms.
Drawing on the advocacy coalition framework, Babon et al. (2014) examine potential pathways to transformational change in Papua New Guinea.They argue that members of different coalitions form "coalitions of convenience."Organizations may defect from one coalition to another, bringing their influence and resources with them.Their findings suggest that, despite the early presence of high-profile REDD+ policy entrepreneurs and the government's pro-REDD+ position at the international level, the most powerful coalition of actors in the domestic policy sphere continue to defend existing institutional structures that support business as usual, often under the guise of agricultural development.Complementing social network with discourse analysis, Rantala and Di Gregorio (2014) show how in Tanzania there are two main opposing policy coalitions that support distinct directions for REDD+.They analyze the discourses of these coalitions and show how the content of National REDD+ Strategy reflects the bargaining power of these two opposing coalitions.

Power and influence: driving change?
The investigation of how influence is exerted through interactions lies at the center of the policy network concept.In this special feature "power" is conceptualized as the ability of actors to influence forestry and land-use decisions such that the outcomes of these decision processes serve their interests (Arts 2003, Arts and Van Tatenhove 2004, Biermann 2010, Krott et al. 2013).It is operationalized mainly as the way in which actors use reputational power and their position in material and immaterial resources exchange networks to exert influence (Marsh and Rhodes 1992, Fischer et al. 2007, Heaney 2014).Power over the REDD+ policy process has many different expressions.As an example, the degree of concentration of power in particular actors can either reflect the autonomy of the nation state from the interests driving deforestation and forest degradation, the extent of national versus international ownership over reform processes, or the level of inclusiveness of policy processes.The case studies from Cameroon (Dkamela et al. 2014), Vietnam (Pham et al. 2014), Nepal (Bushley 2014), Brazil (Gebara et al. 2014), and the comparative paper (Brockhaus and Di Gregorio 2014) all investigate issues related to distribution of power.http://www.ecologyandsociety.org/vol19/iss4/art29/In Cameroon, the international actors are central in controlling and facilitating information flows across organisations, whereas state actors are less involved and domestic civil society organizations are peripheral.The limited presence of domestic actors among the leading organizations indicates that national ownership of the REDD+ process is low (Dkamela et al. 2014).
The lack of domestic leadership in REDD+, low levels of inclusiveness, and the absence of a substantive national REDD+ coalition mirror similar weaknesses experienced in earlier forestry law reform processes, and reveal path-dependent power structures.
The analysis of participation in decision making in key policy events reveals a very different reality in Vietnam (Pham et al. 2014).Government agencies dominate REDD+ policy making, leaving limited political space for nonstate actors such as NGOs and civil society organizations, yet, even within these circumstances nonstate actors have been able to present alternative policy options.This suggests that the REDD+ consultation processes have had some effect, yet there is the need for more inclusive and accountable policy processes.
Measures of reputational power, information exchanges, and collaborative ties between REDD+ actors are used to assess whether REDD+ is threatening to recentralize forest management in Nepal (Bushley 2014).Here, REDD+ policy is largely shaped by the interactions among a triad of government agencies, peak civil society organizations, and international actors, while the nonforest related government agencies, the private sector, and community organizations represent weaker sections of society remain in the periphery of the REDD+ policy network.Such features indicate that REDD+ might be reversing some of the gains of decades of decentralized community forestry in Nepal.In the Brazil case study, Gebara et al. (2014) combine evidence from social network and qualitative analysis to examine the polarization of views on key REDD+ issues and the extent of coordination among government agencies, NGOs, and the private sector.They use a variety of centrality and brokerage measures from different network relations to identify who is shaping the policy design of the national strategy.They find that while government agencies and national NGOs are the most central actors in the REDD+ arena, coordination roles are limited to a few international donors and domestic NGOs.Polarization and lack of coordination across sectors and with the private sector are identified as the main constraints to transformational change.
Power structures in a country affect its progress in national REDD+ policy making, yet the broader political context and the stage of REDD+ policy development shape these power structures in turn.The comparative paper identifies the different power structures in the REDD+ domain in the seven countries based on two dimensions: the level of distribution of power and the dominant type of interaction, i.e., either cooperation or conflict.This analytical lens is used to map how different power structures impact the progress of REDD+ decision-making processes (Brockhaus and Di Gregorio 2014).

OUTLOOK
We draw on the papers herein to provide an outlook for the relevance and future design of policy network analysis when applied to REDD+, but also when applied to policy networks more broadly.This special feature provides a unique collection of individual case studies of national REDD+ policy networks that use a common methodology and are complemented by a comparative analysis.The reader can gain insights of countries' individual and shared opportunities and obstacles for promoting a national REDD+ agenda that is carbon-effective, cost-efficient, and equitable, while delivering cobenefits.All case studies show to higher or lesser degrees how national REDD+ policy arenas are still dominated by powerful business-as-usual interests, reflected in some countries' reluctance to undertake the larger policy reforms that would enable the required change and effectiveness in tackling the often underlying causes of deforestation and forest degradation.The individual papers in this special feature also provide insights and possible ways forward to overcome these challenges.In doing so, the case studies use a range of analytical angles and perspectives that policy network analysis allows for, in a context of highly diverse political regimes and governance structures.
The studies provide evidence of how power structures that are embedded in interactions among a variety of different policy actors with varying interests and ideas contribute to REDD+ progress.The power of agency remains somewhat underexplored in the policy network literature.However, its application to the investigation of policy coalitions provides some insights into agency and drivers of policy change, and powerful forces that hinder such change, as shown in most of the REDD+ policy domains investigated in this special feature.The position of powerful actors in policy interactions, the distribution of power, and the features of the dominant policy coalitions in national REDD+ contribute to explain REDD+ outcomes, or, as most case studies indicate, the lack of achieving these.However, these aspects shed less light into how changes in power structures occur and why and how REDD+ policy networks change over time.
Policy network analysis has been criticized for focusing primarily on structural and stable features of policy domains (Börzel 1997).
To help understand how past processes have impacted current policy networks structures and are likely to impact future developments, the case studies in this special feature use a combination of networking data and qualitative data on policy processes.However, the use and advances in longitudinal analysis of policy networks, where network data are observed at different points in time (Snijders et al. 2007, Ingold andFischer 2014) provides a way to further investigate policy network change over time.
Finally, the comparative analysis of policy networks in different countries presents a number of methodological challenges, because social network analysis is best suited to compare different relations among the same actors as opposed to networks formed by different actors altogether (Knoke et al. 1996).To advance the value of policy network analysis for understanding policy processes and outcomes, and to strengthen its explanatory power as theory and method of investigation of environmental governance, longitudinal analysis, a stronger recognition of the role of agency in policy networks, and additional efforts for comparative analyses are required. [1]Meso-level policy analysis focuses on "how policies come to be made, who puts them on the policy agenda, and the structure of institutional arrangement in which policy is defined and eventually implemented" (

Organizational Efforts on REDD Introduction
This research project studies the positions and activities of organizations concerning issues related to Reduced Emissions from Deforestation and Degradation (REDD).There are many different opinions about this subject.We would like to talk with you today about the positions and activities of (ORG NAME) concerning these issues.Because interviews are being carried out with many organizations, we will follow an interview guide to make sure we cover all the questions.The name of the respondent will be kept confidential.
Question 1: Please indicate the amount of effort your organization typically devotes to each theme related to land use and REDD, using the six response categories to the right.

Organizational Stance on International and National REDD Issues
Question 4 Next we want to talk about the stance or position of (ORG NAME) on important international and national issues regarding REDD.In Table Q4, these issues are stated in a one-sided manner to allow agreement or disagreement.
Please circle the number that best reflects the level of agreement/disagreement of (ORG NAME)'s typical policy stance with the issue as stated.Please note that not all organizations are engaged with the vast variety of issues related to REDD.If you feel there is no specific or implied position of your organizations on one of the issues you can tick 'neither agree nor disagree'.If the topic is unknown to you and your organization please tick 'not known/no response' NOTE: These response categories differ from the previous ones.{the choice 'not known/no response' should also be ticked in cases when the respondent does not know or does not respond}  Semi-structured interviews of the core organization/actors respondents will complement the Social Organization survey, they will be administered to a high ranking representative of the organization who is politically-knowledgable about REDD policy debates.
Interview might be conducts will all core organization that are part of the policy domain, or a subset of peak organizations with a balanced representation among government, business, civil society and international actors.These interviews will provide more depth and details with regard to main the stances, and preferred policy options of the organization.The semi-structured interviews guide is organized in 4 sections.
In the first set of questions, the respondent will be asked to describe the main beliefs, interests and activities of his/her organization in relation to the REDD debates.
In the second section, the respondent will be asked to comment about the main challenges posed by a REDD policy strategy and about the existing opportunities provided by the introduction of REDD policies.The third section asks to assess governance aspects of the consultation process leading to national REDD strategies.Finally the respondent will provide and assessment of existing REDD related policies and of likely policy directions in terms of effectiveness, cost-efficiency, equity and other co-benefits.These interviews can be analysed with a full-fledged "grounded theory" approach using open coding technique.The teams will record and have these interviews transcribed into text form, and will then code them using NVivo (or another computer-assisted qualitative data analysis package) for the categories of thought and value judgment that the actors use, as well as the specific substantive information they mention.

GUIDING QUESTIONS:
Section 1: Organizational Interest First, I would like to ask you about your organization's main interest and activities

Question 1
What is the main mandate or mission of your organization?

Question 2
When did your organization first get involved in the REDD debate and why?Section 3: Consultation Process With respect to consultation processes undertaken so far for the development of the national REDD strategy….

Question 9
Can you please comment of the role, nature and effectiveness of these consultation processes?
(Probe on access to information, inclusiveness of stakeholder participation, quality of participation)

Question 10
Do you feel the position and claims of your organization were seriously considered in the decisionmaking process?
Section 4: Policy assessment and 3 Es

Question 11
Looking at existing REDD related policies and likely directions of policy design and implementation.Do you think these policies will effectively address the drivers of deforestation and degradation?Will they be effective in reducing national level carbon emission from REDD-Plus related sectors?

Question 12
So far, do you think proposed policies effectively address issue of cost-efficiency in the design and implementation of REDD related?Can you give examples on why this is the case?

Question 13
Can you comment on the extent to which equity (fairness) issues are or will be taken into account in design and implementation?
(Probe on the extent to which co-benefits -as poverty alleviation, protection of biodiversity and other environmental services, and governance improvementswill/should be included as objectives) Hudson and Lowe 2009:11, emphasis in the original).http://www.ecologyandsociety.org/vol19/iss4/art29/

TABLE Q1 .
ORGANIZATIONAL EFFORT -LARGE THEMES

TABLE Q4 .
ORGANIZATIONAL STANCE ON NATIONAL REDD ISSUES

of the main challenges for an effective REDD national strategy is …
Below, are a few further stances with refer to very technical aspects of REDD.Not all organizations involved in the REDD policy domain will be familiar or engaged with these issues, which are usually only relevant for more scientifically specialized organizations.If your organization is not familiar with these issues please tick '0' 31.REDD schemes are also likely to help countries to

cope or adapt to the impacts of climate change
Please look at Table5below.This is a list of organizational activities.Please indicate how much effort your organization devotes to each type of activity related to REDD, using the six response categories to the right.

TABLE Q5 .
ORGANIZATIONAL ACTIVITIES ON REDD section investigates REDD policy networking activities.It identifies and traces different types of relations between organizations.Questions refer to the influence of other organizations, communication/exchange of information, sources of scientific information, organizations with opposing positions, and organizations with whom there is collaboration} Network questions introduction:Now we want to ask about (ORG NAME)'s relations with other organizations and actors.Please answer each question by putting a tick by the relevant organizations from the following list.There is also space at the bottom for you to add important organizations/actors if they are not on the list.
{This Question N 2:{Regular COMMUNICATION and INFORMATION EXCHANGE } Please indicate those organizations with which (ORGNAME) regularly or routinely discusses and exchanges information about national REDD policy matters?Question N 3: {RECEIVE expert scientific information about} Sometimes organizations need expert scientific information about REDD that can only be provided by other organizations or particular scientists.On which organizations does (ORGNAME) rely on to obtain reliable scientific information on REDD? (if information is received from independent consultants not affiliated with any specific organizations, please add the name of the consultant at the bottom of the list and tick it) Question N 4: {GIVE FUNDS } To which organizations on List 1 has (ORGNAME)

contributed substantial funds or other in-kind resources as
payments and cost-sharing for services or goods, or for voluntary contributions including co-sponsoring of activities of common interest.

with other interests (human rights, gender etc,, farmers' orgs, other grass-root orgs , indigenous groups, other CBOs) International Environmental NGOs and Networks International NGOs (non-environmental) International Government Organizations (e.g. associations of local governments)
Question P 2: Now I would like to ask you to fill in the protest answer sheet.It lists relevant protests that have been held on issues that might affect REDD policies, and it should be filled in for any other recent major protest

events related to REDD in which (ORGNAME) participated.
Please read through the list, check off the protest actions in which (ORG NAME) participated, and for those rows fill in the required information.
# Link: if the protest event is linked to one of the above policy events, enter the event code number.

Section 2: Policy Challenges and Opportunities Question 5
What is your organization's main position on REDD and what are your main objectives in terms of policy design and implementation?What are the main activities in the REDD arena that your organization is engaged in?What are the main challenges for REDD in [COUNTRYNAME]?In particular, are there, and if so which are the main governance challenges that need to be overcome to design and implement effective REDD policies in [COUNTRYNAME]?(Probe on aspects related to coordination, transparency and accountability of policy actors, institutional and civil society capacity) Do you believe there are important opportunities that REDD can provide to[COUNTRYNAME].If so, what are these opportunities?What do you think will be the consequences of the introduction of a national REDD policy strategy for [COUNTRYNAME]?