International forest policy by international and transnational organizations

By the end of the 1980s, Argentina was in the middle of a severe economic crisis. In 1991, the Deregulation Decree, which steered the political economy toward a new neoliberal policy, dismantled the Argentine National Forestry Institute (IFONA), an autonomous bureaucracy responsible for forests. The aim of this study is to analyze the influence the World Bank exerted on domestic forest policymaking and bureaucratic reform in Argentina. We selected the interventions of the World Bank in the Argentinian forest and agricultural policy that started in the early 1990s and still continues today. We use a qualitative case-study design building on content analysis of policy documents. The World Bank interventions through funding figured prominently, in the form of new forest laws that, as a whole, benefited plantation forests and regulated soy production expansion. This policy was found to be supported by a coalition of the World Bank, agricultural interests, and private landowners. ARTICLE HISTORY Received 11 April 2014 Accepted 26 July 2015

v ACKNOWLEDGEMENTS I wish to thank the following people who over the last three years and a half offered me their help and support in different ways.
In the first place I would like to thank Prof. Krott for giving me the opportunity to join his group. His critical comments throughout my PhD and vast knowledge in forest policy allowed me to improve my work as well as giving me the opportunity to learn a lot. All our many discussions helped me understand a lot about bureaucratic politics, actors´ interests and power. Although we did not always agree, he always gave me the opportunity of discussing our ideas and by this he helped me understand different theories of forest policy analysis. I would also like to express my gratitude to Prof. Daniela Kleinschmit for reviewing this thesis. During my PhD I had the chance to meet with Prof. Kleinschmit in different conferences and in every occasion her critical questions in my presentations helped me to improve my work considerably. By her always positive attitude as well as her excellent scientific career she became a role model for me and I am very proud of having the chance to have her as an examiner of my dissertation.
As well I would like to thank Prof. Bürger-Arndt for agreeing to review this thesis. Her background in nature conservation and experience in the field gives an added value to this dissertation. I would specially like to thank Dr. Lukas Giessen, my daily supervisor, for his constant support and guidance, encouraging and helping me to grow as a researcher and scientist. In our daily interactions he has been always available to discuss ideas, manuscripts and field work assistance. By working with him I was able to develop many different skills in international relations, forest policy and governance as well as public policy, encouraging me by valuing my work and introducing me to a new research discipline to me, which I learned to enjoy a lot. Thank you for three wonderful years of constant learning.
I also want to thank all my colleagues in the International Forest Policy Working Group: Jacqui, Alif, Saifur and Agung and for a short but great time Carmen. We grew together professionally in the last three years by sharing many discussions about our progress and supporting each other in the process of writing our thesis. As well I want to thank all the other PhD students in the Chair of Forest and Nature Conservation Policy: Janina, Kenji, Huong, Budi, Carsten, Manjola and all other PhD students, some who have already finished for sharing with me daily work activities.
A special "Gracias!" goes to all my colleagues and friends from the Laboratorio de Sistemas Ecológicos y Ambientales (LISEA) from the National University of La Plata in Argentina, for making me feel that I never left and giving me their usual support and friendship over skype meetings, emails and chats. To Carolina Perez, Marcelo Arturi and Marcelo Barrera for their support from the beginning, not only during these past years but ever since I first started working in the LISEA allowing me to be part of this extraordinary working group. I especially want to thank Juan Goya and Pablo Yapura, vi who not only replaced me in my teaching activities in Argentina, but also encouraged me and helped me to grow ever since I was a forestry student giving me later the chance to work with them, and by this join this incredible group of colleagues and friends.
I will be eternally grateful to the National University of La Plata (Universidad Nacional de La Plata) for having trained and educated me gratuitously and for granting me a leave of absence that allowed me to come to Germany to do my PhD studies. I am extremely proud of belonging to the UNLP as a student, graduate and teacher.
I also want to thank my friends and family (de sangre y de ketchup) back in Argentina. For the past three years you supported me over the phone in so many occasions, always being with me in the good and in the bad times. One thing I learned over these past years is that not even an ocean can bring us apart. Gracias por estar siempre! I would like to thank Kerstin Wiegand, Katrin Meyer and all the Ecosystem Modelling group for including me in so many activities and sharing so many fun and delicious lunches in the Nordmensa together making me feel as part of the group.
Last but not least, I want to thank my friends in Germany for all the nice times spent together. Without you I would not have been able to enjoy these past three years the way I did, hanging out in our living room in Schröders watching football and drinking beer. Thanks to you I feel at home in Göttingen. A special thank goes to Bettina for being the most amazing friend to me since I first arrived, introducing me to basically everyone I know in Göttingen, teaching me to color eggs for Easter and as Daisy said, showing me that black and yellow are the best colors ever! To Daisy, for sharing so many moments and talks about our experiences and feelings, even though you like black and I like pink and for teaching me a completely new Spanish. Gracias Daisyta por estar siempre! To Yasmin, my official German translator, for all the emotional and practical help throughout these years, without them I would have been lost. Sonderbar and I are very grateful for having you in our lives! To Natalia, for all the long and crazy talks that make it impossible to be bored or sad. Nubecito, te quiero mucho! I don´t want to forget anyone, Niki, Kaufi, Torben, Paul, Meik, Marc, Matteo, Federico (the disgusting rat), Iris, Urs, Maria (Rose), Nele (our sweet Nonna), Mike (Irish rat), Martyna, Per, Anna, Thomas, all of you are the reason life in Göttingen has been fun and warm.
For all of you….

Summary
In recent decades, globalization and internationalization led to an increase in the number of international policies and regimes as well as the creation of a new form of governance based on voluntary agreements, attempting to influence national behaviour over many different issues. The complexity and fragmentation in global governance arrangements matches the complexity at the domestic level with whom they interact. Findings from public policy analysis suggest that especially the interaction between international regimes and domestic, mostly national bureaucracies is crucial being analyzed since it is in this interaction where the politics of competing domestic bureaucracies have a strong bearing on the influence international actors and institutions will be able to develop.
By using the case of the international forest regime complex this thesis seeks to evaluate how an international organization such as the World Bank and private institutions of forest certification influence domestic forest policy. The cases of Argentina and Armenia were selected as examples of developing countries open to international influences with weak forest sectors that went through a recent administrative restructuring. In so doing this dissertation seeks to answer how do international and transnational organizations influence domestic forest policies?
In order to address this question, four articles addressing specific questions in selected study cases were developed. Article 2. Burns, S. L. and L. . Dismantling Comprehensive Forest Bureaucracies: Direct Access, the World Bank, Agricultural Interests, and Neoliberal Administrative Reform of Forest Policy in Argentina. Society and Natural Resource 29(4): 493-508. This article analyzes the influence the World Bank exerted on domestic forest policy-making and bureaucratic reform in Argentina resulting in new forest laws that, as a whole, benefited plantation forests and regulated soy production expansion. This policy was found to be supported by a coalition of the World Bank, agricultural interests and private landowners.
Article 3. Burns, S. L., Yapura, P.F. and L. . State actors and international forest certification policy: Coalitions behind FSC and PEFC in federal Argentina. Land Use Policy 52: 23-29. This publication analyzes the role state agencies played in the forest certification standard development processes in Argentina, including the related coalition building strategies.
Article 4. Burns, S. L., Krott, M., Sayadyan, H. and L. Giessen. The World Bank´s power on environmental and natural resources policies in Armenia: neoliberal changes and post-soviet heritage. Under review in World Development. This article analyzes the World Bank´s power resources and its resulting influence on environmental and natural resources policies under the umbrella of poverty alleviation projects in a post-socialist fragile state using Armenia as a case.

viii
In order to answer the main research question these four articles addressed more detailed questions: Who are the dominant actors in forest policy? II.
What domestic allies, state and non-state, do international and transnational organizations have? III.
Which coalitions made by international and transnational organizations are more frequent? With state bureaucracies or with non-state allies? IV.
How stable are these coalitions? Are they permanent or dynamic? V.
What are the factors to break up coalitions?
Two cases with strong political dynamics in the forest sector and strong intervention by the World Bank and with additionally good field access, Argentina and Armenia, were selected with the aim to obtain empirical rich evidence for testing the hypothesis needed. The thesis focused on two main sources of data: written documents and semi-structured interviews with experts. In total approximately 60 interviews were carried out, 50 in Argentina and 10 in Armenia between October 2012 and September 2014.
In addition around 150 documents from both public and expert sources were analyzed in the same time period. Process tracing was used to understand how international and transnational organizations influence policy change in the domestic level through direct access and market mechanisms. In all cases data was analyzed by means of qualitative content analysis.
The results show that, first, which issues become relevant in different countries depends on the domestic actors and their interests leading to differences in the relevant issues in each country, as observed from the issues detected in Argentina which were not the same as the relevant subjects found in similar studies conducted in Indonesia and Bangladesh although some similarities were observed. Second, by means of funding and capacity building, the World Bank, as an international organization, assisted the coalition of agricultural and profit-oriented bureaucracies and landowner associations (of forestry and agriculture) that emerged after the economic crisis in the country at the beginning of the 1990s. The intervention of the World Bank in the reform of forest policy in Argentina figured prominently in the form of new forest laws that, as a whole, benefited plantation forests and opened land for agricultural production. Third, the influence of the World Bank in Argentina and Armenia through direct access by funding projects aiming at poverty alleviation and sustainable natural resources management pushed the forest sector towards deregulation. Following a framework of neoliberal conservation governance, promoting market-based mechanisms that aim to motivate local users to address environmental concerns by means of incentives, the World Bank promoted the intensification of natural resources production both in the case of Argentina and Armenia. Although seen as a fundamental step towards sustainability, these neoliberal conservation policies have been criticized for informally promoting decentralization, privatization and commodification that allow corporations and international actors to increase their influence over local natural resources. Fourth, state bureaucracies play an important role in the implementation of transnational regimes at the national level. In forest certification processes international actors seek coalitions with different state bureaucracies. However, in situations where new promising venues for political influence open, private actors prefer ad hoc coalitions with weaker bureaucracies that would not be permanent and with less technical knowledge. Fifth, the political system of federal countries like Argentina provides multiple institutional access points for policy change that international ix and transnational regimes try to use in order to influence the domestic level, consequently changing the power balance of the domestic networks. Domestic actors then can use these regimes to shift the power balance in their favor. This was observed in Argentina where national bureaucracies benefited by gaining influence through the implementation of the International Forest Regime Complex.
x xi

Introduction and research questions
In recent decades, globalization and internationalization led to an increase in the number of international policies and regimes attempting to influence national behaviour over many different issues (Lindstad and Solberg 2010). This rise in international regimes dealing with issues as different as development cooperation, health, trade, agriculture, environment and forests among others brought along a burst of research on regime compliance and effectiveness both in the field of international relations as well as in environmental policy (Hisschemöller and Gupta 1999, Sprinz and Helm 1999, Helm and Sprinz 2000, Young 2001, Hovi et al. 2003, Sprinz 2005).
An international regime is defined as a "set of implicit or explicit principles, norms, rules and decisionmaking procedures around which actors´ expectations converge in a given area of international relations" (Krasner 1982, 186). Although originally these regimes aimed at constraining or modifying the external behaviour of states, a new generation of international regimes which boomed mainly after the Rio Summit in 1992, attempt to influence domestic practices, policies and policy-making processes .
At the same time international regimes boomed, and with the increased participation of non-state actors, a new form of governance based on voluntary agreements was created. This new form of private governance has been referred to as transnational regimes. The main difference between international and transnational regimes is that in the latter, non-state actors generate the set of norms and rules instead of states (Pattberg 2012). As a result, current governing arrangements are no longer represented by single-issue regimes with states as only actors, but rather by a mixture of mechanisms including legal, non-legal, governmental and non-governmental arrangements .
The new complex global governance arrangements led Cashore (2010, 2012) to propose a change in focus, moving away from regime compliance and effectiveness into influence. According to the authors the means by which international regimes may exert their influence at the domestic level can be categorized into "four pathways of influence, each with its own causal logic: international rules (the traditional focus of regime effectiveness literature); international norms and discourse; creation of or interventions in markets; and direct access to domestic policy processes" (Bernstein and Cashore 2012, 587). It is "the interaction of mechanisms and processes, sometimes along multiple pathways, that create collective influence" (Bernstein and Cashore 2012, 603).
The complexity and fragmentation in global governance arrangements matches the complexity at the domestic level with whom they interact. Compliance and effectiveness approaches have so far considered the interaction in only one direction, that is, from the international level to the domestic level. However, influence in the opposite direction also occurs when national states exert influence over the international processes blurring the boundaries between global and domestic politics (Lindstad and Solberg 2010, . Findings from public policy analysis suggest that especially the interaction between international regimes and domestic, mostly national bureaucracies is crucial being analyzed (Hofmann 2002, Krott 2005, McDermott et al. 2010. It is in this interaction where the politics of competing domestic bureaucracies have a strong bearing on the influence international actors and institutions will be able to develop . The International Forest Regime is a good example of a complex and fragmented regime without a single legally binding agreement, but rather a set of different instruments dealing directly or indirectly with different aspects of the world´s forests (Humphreys 1999, Arts and Babili 2012. The complexity of this regime led scholars to introduce the term International Forest Regime Complex (Glück et al. 2010, Howlett and Rayner 2010, Rayner et al. 2010. A regime complex can be defined as a "set of specialized regimes and other governance arrangements that are more or less loosely linked together, sometimes reinforcing each other but at other times overlapping and conflicting" (Giessen 2013, 62). So far the international forest regime complex has been described as fragmented (Humphreys 1999, Glück et al. 2010, Howlett et al. 2010, Rayner et al. 2010, hollow (Dimitrov 2005, Dimitrov et al. 2007, ineffective (Humphreys1999, 2006, Dimitrov 2005) and a failure (Dimitrov 2005, Smouts 2008. However,  claims that the main characteristic of the international forest regime complex is its fragmentation, which also explains the other three attributes. Fragmentation can be defined as "a patchwork of international institutions that are different in their character (organizations, regimes, and implicit norms), their constituencies (public and private), their spatial scope (from bilateral to global), and their subject matter (from specific policy fields to universal concerns)" (Biermann et al. 2009, 16). According to Biermann et al. (2009) all global governance arrangements show some degree of fragmentation. These characteristics of the international forest regime complex might be the result of a global tacit agreement on having a weak and hollow forest regime and not just a lack of consensus between actors leading to what has been defined as a non-regime (Dimitrov 2005, Dimitrov et al. 2007). As pointed out by many authors, these fragmented regime complex is the result of the different interests of the actors involved. While it benefits both Northern and Southern states  it also benefits transnational non-state actors and institutions (Arts andBuizer 2009, Pattberg 2007). The benefits of the fragmentation are not only observed at the international level but also at the domestic level, where different bureaucracies competing among each other over a certain issue would benefit differently from the regime .
For a fragmented regime complex, as the international forest regime complex, to influence domestic policies at national levels, it depends on international and especially domestic coalition partners (Humphreys 2006, Bernstein and. These can include international as well as domestic state and non-state actors. Special attention, however, must be paid to domestic bureaucracies, as they have to invite international actors due to sovereignty principles and in the domestic policymaking process they are instrumental in developing policies including also regulatory instruments. A research program on international forest policy that examines the pathways and politics through which international institutions influence such domestic policy change is currently being developed (Arts and Babili 2012;.
By using the case of the international forest regime complex this thesis seeks to evaluate how an international organization as the World Bank and private institutions of forest certification influence domestic forest policy. The cases of Argentina and Armenia were selected as examples of developing countries open to international influences with weak forest sectors that went through a recent administrative restructuring.
In so doing this thesis poses the following research questions:

How do international and transnational organizations influence domestic forest policy?
In order to answer the main research question more detailed questions are proposed: x Who are the dominant actors in forest policy?
x What domestic allies, state and non-state, do international and transnational organizations have? x Which coalitions made by international and transnational organizations are more frequent? With state bureaucracies or with non-state allies? x How stable are these coalitions? Are they permanent or dynamic?
x What are the factors to break up coalitions?
These questions will be addressed in a cumulative PhD dissertation consisting of 6 peer-reviewed publications. A brief description of the four core publications and the contributions made by the author of this dissertation (marked in bold) as well as by the other authors is detailed below.
Article 1. Burns, S. L. and L. . Identifying the main actors and their positions on international forest policy issues in Argentina. Bosque 32(2): 163-173. This article identifies the most relevant forest issue elements in Argentina as well as the main actors involved and their positions towards them as a result of their interests. Burns, as the first and main author of this publication applied theory, hypothesis and methodology to the case as well as findings regarding the implementation of the main forest issue elements from the international forest regime complex in Argentina. Giessen developed the general theory, hypothesis and methodology.
Article 2. Burns, S. L. and L. . Dismantling Comprehensive Forest Bureaucracies: Direct Access, the World Bank, Agricultural Interests, and Neoliberal Administrative Reform of Forest Policy in Argentina. Society and Natural Resource 29(4): 493-508. This article analyzes the influence the World Bank exerted on domestic forest policy-making and bureaucratic reform in Argentina resulting in new forest laws that, as a whole, benefited plantation forests and regulated soy production expansion. This policy was found to be supported by a coalition of the World Bank, agricultural interests and private landowners. Burns, as the first and main author of this publication applied theory, hypothesis and methodology to the case as well as findings regarding the influence of international organizations on domestic forest policy change. Giessen developed the general theory, hypothesis and methodology.
Article 3. Burns, S. L., Yapura, P. F. and L. . State actors and international forest certification policy: Coalitions behind FSC and PEFC in federal Argentina. Land Use Policy 52: 23-29. This publication analyzes the role state agencies played in the forest certification standard development processes in Argentina, including the related coalition building strategies. Burns, as the first and main author of this publication applied theory, hypothesis and methodology to the case as well as findings regarding the influence of transnational regimes and the role of state bureaucracies in an initiative that was so far claimed as non-state private governance. Yapura applied methodology to the case and produced case findings regarding the development of FSC in Argentina. Giessen developed the general theory, hypothesis and methodology.

4
Article 4. Burns, S. L., Krott, M., Sayadyan, H. and L. Giessen. The World Bank´s power on environmental and natural resources policies in Armenia: neoliberal changes and post-soviet heritage. Under review in World Development. This article analyzes the World Bank´s power resources and its resulting influence on environmental and natural resources policies under the umbrella of poverty alleviation projects in a post-socialist fragile state using Armenia as a case. Burns, as the first and main author of this publication developed and applied theory, hypothesis and methodology to the case as well as findings regarding the influence of international organizations in fragile states through environmental policy change. Krott and Giessen developed the general theory, hypothesis and methodology. Sayadyan applied methodology to the case and produced case findings regarding the outcome of the World Bank projects in Armenia.

International Organizations -The World Bank
An international organization can be defined as "an institutional arrangement that combines three elements: a normative framework, a group of member states, and a bureaucracy as administrative core" (Bauer et al. 2012, 28). This definition allows discriminating analytically international organizations from international bureaucracies. An international bureaucracy is then defined as "a public institution which makes decisions concerning specific international problems on the basis of international law standards and international agreements, resolving those problems by implementing special measures" (Biermann et al. 2009, 37).
International organizations and their bureaucracies are considered key actors, directing patterns of global change , Biermann et al. 2009). Among these organizations the World Bank is considered as perhaps the most influential actor bringing about certain types of changes in developing countries (Singh 2009, Marschinski andBehrle 2009). In order to operate the World Bank uses what is known as blueprints that include the Bank´s general policy models. These blueprints are then implemented in many different countries with the expectation that they will produce the same results ).
In dealing with forest policy the World Bank has developed three forest policy initiatives: the 1978 forestry policy, the 1991 and the 2002 forests strategies although the first loans for forestry projects date back to 1949 . While neoliberal and interventionist elements, characteristic of the World Bank interventions in all issue areas, have always been present throughout the history of the World Bank intervention in forest policy, the different forest strategies showed some formal changes . The 1978 forestry policy sought to fulfil the needs of the rural poor while considering the environment. However criticisms claimed that instead of helping the poor it blamed them for deforestation, prioritizing forest industries and plantation forests . The 1991 forests strategy responded to these criticisms with a move towards a more conservationist approach emphasizing even more on poverty alleviation by means of new modes of market-based conservation mechanisms. During this time half of the World Bank´s loans went to boreal and temperate forests, including the former communist countries in Eastern Europe and the former Soviet Union . With the 2002 forests strategy the World Bank promoted private sector investment in forests, new environmental markets and voluntary private sector regulation .
Previous research on policy change and bureaucratic reforms focused mainly on domestic factors as explanatory variables. However, international organizations like the World Bank are also important actors leading to domestic policy change 2009, Burns andGiessen 2016, Burns et al. unpublished). In this thesis we try to analyze how the World Bank as an international organization influences domestic forest policy change.

Transnational Organizations -Forest Certification
After the Rio Summit in 1992, with the increased participation of non-state actors, a new form of governance based on voluntary agreements was created. This new form of private governance has been referred to as transnational regimes. The main difference between international and transnational regimes is that in the latter, non-state actors generate the set of norms and rules instead of states 6 (Dingwerth 2005, Pattberg 2012). As well as international regimes, transnational regimes operate within a wide range of issue areas that have multiplied in the last 20 years (Dingwerth and Pattberg 2009, Pattberg 2012. The lack of credibility of intergovernmental organizations has been indicated as an explanation for transnational regimes' success by emphasizing on inclusiveness, transparency, accountability, and deliberativeness (Dingwerth and Pattberg 2009, Klinke 2009).
Through certification of forest products a new form of political steering has been promoted based on voluntary agreements with the increased participation of non-state actors. In the political science body of literature, forest certification has been so far described as a non-state, market-driven form of private governance developed by non-state actors and building on private institutions , Pattberg 2007). In international policy, it is currently referred to as transnational regimes (Pattberg 2012) and more broadly discussed under the concept of forest governance (Agrawal et al. 2008, Arts 2014.
These perspectives in political science highlight aspects relating to the role of private actors and private institutions and the way they gain authority vis-à-vis state failure to solve forest problems, and they are accompanied by a perceived decline in power of forest related public actors and institutions (Ho 2006, Wibowo and. This view, however, leaves largely underrepresented the role of the state in e.g., creating, enabling, or obstructing conditions for forest certification in general and individual certification schemes in particular. It further downplays the role of the multiple and often competing state bureaucracies in certification politics, which may be crucial as both resources of power and as coalition partners in the establishment process of a particular scheme vis-à-vis their competitors. This thesis claims that, despite what is often claimed, state actors play a crucial role in forest certification and related politics , Bartley 2014) by either obstructing , Sahide et al. 2015 or supporting competing certification initiatives (Hysing 2009, Bell andHindmoor 2012).

Pathways of influence used by international and transnational organizations
Current governing arrangements are no longer represented by single-issue regimes with states as only actors, but rather by a mixture of mechanisms including legal, non-legal, governmental and nongovernmental arrangements . The way these complex international regimes may exert their influence at the domestic level can be categorized into "four pathways of influence, each with its own causal logic: international rules (the traditional focus of regime effectiveness literature); international norms and discourse; creation of or interventions in markets; and direct access to domestic policy processes" (Bernstein and Cashore 2012, 587). The international rules pathway refers to binding international treaties, trade agreements or policies of powerful international organizations; the international norms and discourse pathway refers to non-binding norms and discourse by means of information, symbolism and accountability; the creation of or interventions in markets pathway refers to the use of markets to make domestic policy changes; the direct access pathway refers to the influence by means of direct funding, education, training, assistance and capacity-building Cashore 2000, 2012).
Previous studies on the implementation of the international forest regime complex at the national level have analysed how relevant international influences are on domestic forest policy change in different countries through the four pathways of influence from global governance . While some studies focused on only one pathway of influence, as  who analysed how the international norms and discourse pathway through the concept of sustainable forest management and the resulting Montreal process was implemented in Canada following the interests of domestic actors and their networks; and Cashore and Stone (2012) who analysed the influence of market mechanisms as forest certification and timber legality verification systems in South East Asia. Other studies have tried to explain how international influences occur through the four pathways. Lindstad (2015) analysed international influences on the changes in forest protection and forest-related climate policies in Norway through the four pathways of influence, Wibowo (2015) analysed the domestic consequences of international influences in REDD+, one map and forest certification politics in Indonesia. Arts et al. (2013) analysed influences from the global and European Union level to domestic policy processes in Europe observing that the direct access pathway was not very relevant in European countries.
In this thesis the theorem of pathways of influence is used to analyse how international and transnational organizations promote forest policy change at the domestic level. Especially, the direct access (Burns and Giessen 2016, Burns et al. unpublished) and the creation of or intervention in markets  pathways are used. While  propose that international norms and discourse is a stand-alone pathway, this thesis proposes this is not necessarily a distinct pathway of influence as the other three but rather a pre-stage preparing the way for any of the other three pathways of influence to occur (Burns and Giessen 2016, Burns et al. unpublished). This goes in line with Arts and Buizer (2009) who claim that sustainability and governance discourses are institutionalised in global forest policy over time. As an example, discourse at international level introducing the concept of "sustainable forest management (SFM)" gave way to the creation of markets pathway by means of forest certification schemes to promote SFM (cf. ).

Domestic actors and bureaucratic politics theory
In order for international and transnational regimes to become influential and to impact on domestic policies, a specific actor has to take agency (Biermann et al. 2009). These actors can be either state or non-state. Domestic state actors are defined as domestic bureaucracies. A bureaucracy can be defined as "an agency that has been set up by governments or other public actors with some degree of permanence and coherence and beyond formal direct control of single national governments" (Biermann et al. 2009, 37). Bureaucratic politics theory claims that bureaucracies compete with each other for resources, staff and responsibility for policy domains. Resulting public policy is strongly shaped by this competition between bureaucracies with different preferences, abilities and power capabilities (Krott 2005. In their struggle for power, bureaucracies act in coalition with other public and private actors in fields where their interests (partly) overlap, trying to establish institutions in support of this shared interest. Being rather stable over time, this combination of shared interests and institutions is referred to as a policy sector, e.g., forestry or agriculture . In joining forces, policy sectors attempt to influence public policy based on the overlapping portions of their interests.

8
In forest related international negotiations domestic bureaucracies compete to represent their countries in international forest negotiations , Edwards and Kleinschmit 2013). According to  the bureaucracies which may have an active role in the international forest regime complex and its implementation are: agriculture (including forestry), environment, hybrid between agriculture and environment, foreign affairs, and economics and trade. Based on their tasks and responsibilities, the main line of conflicts runs between conservation-oriented bureaucracies as the ministries of environment and the utilitarian bureaucracies as the ministries of agriculture and forestry . The fragmentation of the international forest regime complex leads to the possibility of different bureaucracies to be in charge of different issues within the regime complex .  used bureaucratic politics theory to show an increase in responsibility of utilitarian bureaucracies over more conservation-oriented bureaucracies in different issues within the international forest regime complex.
Following the core theses of bureaucratic politics theory this thesis theorizes that forest-related bureaucracies use international and transnational regimes as means to increase their power in their permanent competition for resources, staff and responsibility for policy domains.

Power theory
According to previous studies, the power of the different political actors is one of the most important factors explaining forest governance (Agrawal et al. 2008, Brockhaus et al. 2012. Actorcentred power can be defined as "a social relationship in which actor A alters the behaviour of actor B without recognising B´s will" (Krott et al. 2014, 37). Actors have different power sources that they use to produce policy change and increase their influence in the policy making process. These power sources comprise three elements: coercion, incentives and non-verified information . Coercion is defined as altering the behavior of an actor by force, (dis)incentives are defined as advantages and disadvantages introduced to alter the behavior of an actor, and dominant information is defined as unverifiable information that makes the decision process more difficult for actors ).
This actor-centred power theory has been developed using community forestry as an illustrative case . While the theory has already been applied in many community forestry studies across many countries (Devkota 2010, Maryudi et al. 2012, Yufani Movuh and Schusser 2012, Schusser 2013, Schusser et al. 2015, lately it has also been used in studies on the implementation of the international forest regime complex at national level (Brockhaus et al. 2014, Mohammed and Inoue 2014. In this thesis we use the actor-centred power theory combined with bureaucratic politics theory. Bureaucracies, as actors have different sources of power that they use to increase their influence in the forest policy making process. Following this theory the main sources of bureaucracies' power based on coercion are their legal mandates (including decision rights and sanction mechanisms), technical resources (enabling bureaucracies to maintain a policy process) and alliances between bureaucracies and external actors putting pressure behind the preferences of a bureaucracy. The main sources of power based on incentives are transfer funds, creating a link to other issues and funds for setting incentives for others to follow their preferences and alliances between bureaucracies and external actors putting resources behind the preferences of a bureaucracy. The main sources of power based on selected infor-9 mation are the bureaucracies' expertise and active selection of information as well as interest-guided use of beliefs and values ).

Federalism and forests
Federalism can be defined as "a sort of association or league of sovereign states with separate, selfsustaining centres of power and prestige" (Ellefson et al. 2006, 653). This division in power between higher and lower levels of government leads to competition for influence on the policy making process not only between national bureaucracies but also between national and subnational bureaucracies. Domestic bureaucracies may use international issues to strengthen their positions vis-à-vis competing agencies, both at the national as well as the provincial level, by taking up issues, discourses and norms to legitimize their roles as well as by finding domestic or international coalition partners . In federal systems of governance, international issues shift the power balance towards national bureaucracies, even in issues that are actually within the provincial domain (Hofmann 2002. As a response to this shift in power balance, subnational bureaucracies try to get involved in international issues as well. How they will pursue this depends on how responsibilities regarding a certain issue are vertically distributed between the national and subnational governments (Lopreite 2014) but national bureaucracies will remain in a better position. In cases where subnational bureaucracies have the competency over certain issues they might be stronger than in issues of national competences (Lopreite 2014). Federal constitutions granting the subnational level with full regulatory power can obstruct negotiations and implementation of international regimes since national government might not be able to bind subnational governments (Hudson 2012).
While in international forest governance decentralization mechanisms have been previously studied (e.g. Larson 2003, Palmer and Engel 2007, Nathan and Boon 2012) the effects of federalism have so far not been analyzed with only few exceptions (Hofmann 2002, Ellefson et al. 2006, Hudson and Weinthal 2009, Hudson 2012. The United States is an example of the importance of federalism in the implementation of any international forest agreement, since the federal government would be unable to implement any forest management practices due to constitutional constrains (Hudson andWeinthal 2009, Hudson 2012).
This thesis takes as point of departure this important gap in forest policy research and tries to explain how federalism might explain the role different actors play in the forest governance process, especially regarding competition between national and subnational bureaucracies and coalition building with international and transnational organizations Against this theoretical background this thesis proposes the following hypotheses: Hypothesis 1. Elements of the International Forest Regime Complex are made relevant at the domestic level by specific actors. In particular it is the interests of the most powerful actors that determine which forest issue elements become relevant at the domestic level.
Hypothesis 2. International organizations like the World Bank influence national forest policy by the provision of funding and capacity building. In addition to providing resources international organizations gain power by building coalitions with national bureaucracies in order to be able to influence domestic policy.
Hypothesis 3. With the formal goal of enhancing the contribution of the forests to public goods, the World Bank supports market-based conversion of the national forest sector and pushes institutional structures towards deregulation. However, informally, the administrative reform leads to a weak bureaucratic structure unable to implement new policies which are aimed to protect public goods in the forests against side effects of the market driven uses.
Hypothesis 4. Although it is often claimed that transnational regimes network only private actors, in practice state bureaucracies play an important role. In forest certification, international actors seek coalitions with different state bureaucracies.
Hypothesis 5. In federal countries, international and transnational forest regimes influence domestic forest policy by changing the power balance of the domestic networks. International allies help nationallevel bureaucracies to gain relative power over the provincial-level bureaucracies.

Case description
In order to obtain empirical rich evidence for testing the hypothesis two cases with strong political dynamics in the forest sector and strong intervention by the World Bank and with additionally good field access were needed. Following these criteria Argentina and Armenia were selected as case studies. In Argentina the author has good access to the field and rich personal experience. In Armenia cooperation with an experienced senior scientist visiting our Institute in Göttingen where he got familiar with political theory and methods in social sciences provided good field access.

Argentina
Argentina has a continental area of 2,791,810 km 2 , with a length from north to south of 3,700 km between 22° and 55° of Latitude South. This extension determines a wide climatic variation, from subtropical conditions in the north with mean annual temperatures above 20 °C to temperate conditions in the south with mean annual temperatures below 10 °C (SAyDS 2010a). In the country there are around 31.4 million hectares of natural forests divided into six ecological regions (SAyDS 2005). Moreover, there are 1.2 million hectares of cultivated forests, 80% of which are located in the northeast of the country, in the Mesopotamian region (MAGyP 2013). Argentina adopted a federal republican representative form of government. Except for national parks, which are under the jurisdiction of the national government, forests toile within the political responsibility of the provinces and are subject to provincial laws under the umbrella of national laws (Article 124 of the National Constitution 1994). At a national level, the Secretariat of Environment and Sustainable Development is responsible for natural forests, while the Ministry of Agriculture, Livestock and Fisheries is responsible for forest plantations. Forests are mainly owned privately: 99.7% of plantation forests and 93.5% of natural forests are privately owned (Burns and Giessen 2014).
In the early 1990s, in the middle of an economic crisis, a strong coalition of (agricultural and profitoriented) bureaucracies and associations of private landowners (in forestry as well as agriculture) emerged, advocating for more profitable and productive land uses with subsequent supporting policies and administrative frameworks . At this time, international actors as the FAO and the World Bank played a crucial role supporting the domestic coalition of pro-agriculture public and private actors. As a consequence, the Argentine National Forestry Institute (IFONA), a strong and comprehensive forest bureaucracy, was dismantled into four weaker bureaucracies . By forming coalitions with the World Bank, some of the newly created bureaucracies were able to increase their power . The administrative reform and policy changes carried out with funding and a high level of political support provided by the World Bank led to increased investments in the forestry sector, mainly by international firms, especially from Chile (between 1992 and 2013, 66% of the total investments in the forestry sector was done by Chilean companies with three Chilean forestry firms dominating Argentina´s forest holdings and wood exports) (Espach 2005, Burns and. As part of the political changes taking place in the beginning of the 1990s the National Constitution was reformed in 1994. With this reform, province´s sovereignty over their natural resources was stated clear-ly giving the provinces the right to legislate on forest issues within their territory. This makes the subnational bureaucracies involved in the forest sector important actors and potential coalition partners. However, with the new government that took over after the 2001 crisis, the national government increased their power vis-à-vis the provinces by means of new sources of revenue that were not shared between the federal and provincial governments, decreasing provinces´ share of overall revenue and increasing provincial governments´ dependence on the federal government (Levitsky and Murillo 2008).
The above mentioned characteristics make Argentina an interesting case to analyze coalition options between strong private actors as landowners and national and subnational state bureaucracies. The World Bank intervention in the forest policy reform provides a good case to analyze coalition building between domestic and international actors. Between 2001 and 2002, Argentina began the development of forest certification processes, both FSC and PEFC. These processes continued until the year 2014 with the development of standards, providing an appropriate case for studying the role of different state actors, both national and subnational, and the coalitions made, in the development of forest certification schemes.

Armenia
The Republic of Armenia, located in the southern Caucasus, has a total area of 29,800 km 2 within 11 provinces. In 1828, the country was incorporated into Russia and later, in 1920, into the United Soviet Socialist Republics (USSR). Armenia achieved its independence when the USSR was disintegrated in 1991, in the middle of a war against Azerbaijan over Nagorno-Karabakh that lasted between 1988 and 1994. As a result of this war, and together with a devastating earthquake in 1988, the country ended up in severe economic and environmental damage, leading to political violence together with transport, economic and energy blockades . This situation caused great pressure on the forests and other natural resources resulting in soil erosion, high deforestation rates and biodiversity loss . Due to its location as well as its topographic complexity, forests in Armenia have a high percentage of plant and animal endemism, and are included in the Caucasus and Irano-Anatolian biodiversity hotspots . By the year 1992, the total forest cover of Armenia was almost 460,000 ha, with about 11.2% of the country being wooded vegetation . Since the Soviet period all forests are state owned and managed only for conservation and protection, according to the 1978 USSR Forestry Code which divided all forests into three categories, classifying all Armenian forests into Group I (protected forests) allowing only sanitary and reproductive cuts . After independence, a new Armenian Forestry Code was published in 1994 based on the 1978 Soviet Forestry Code, again declaring all forests as state owned and reserved for protection and conservation . In 1995, a new bureaucracy, Armforest State Forest Enterprise (Hayantar), in charge of forest management, was created under the Ministry of Nature Protection . At the same time, a new national forest policy aiming to make economic use of the forest was proposed, but this was never approved . In 1998, the Forest Research Experimental Center (FREC) was created, also under the Ministry of Nature Protection, with the aim of forest monitoring and inventorying. However, coordination between these two forest-related bureaucracies was never achieved . This lack of coordination and capacity of the forest bureaucracies, together with a high demand of forest products caused by the lack of imports from Russia, led to an increase in illegal logging of the forests . At this time, the economy began to recover, with stabilization and structural reforms driving a liberalization of the economy, small and medium privatizations and the formulation of a legal administrative framework for market economy. The Ministry of Nature Protection then invited the World Bank to assist them in the establishment of a new forest policy under the principles of sustainable forest management. This invitation led to the formulation of the "Armenian natural resources management and poverty alleviation" project with funding from both the World Bank and the Global Environment Facility (GEF). Summing up Armenia is a perfect case with dynamic forest policy and strong intervention by the World Bank.

Data collection
All articles constituting this thesis focused on two main sources of data: written documents and semistructured interviews with experts. Written sources were distinguished between public and expert sources. The distinction between these two types of sources lies on the degree of publicity. Public sources were defined as all material accessible to the whole population including newspapers and websites. Expert sources were defined as all material focused on a selected audience from the forest sector. The selection of the interviewees was based on their knowledge of the forest sector as well as on ease of field access. In order to include diversity of viewpoints and interests, especial attention was given to interview both private and state actors from conservation and production oriented sectors. Interviews were carried out in Spanish in the case of Argentina and English in the case of Armenia. Additional interviews were carried out in Armenian by a local collaborator. Interviews lasted between half an hour and two hours. If it was not possible to meet with the interviewees in person, they were contacted by phone calls or emails. In all cases, interviewees were granted confidentiality in order to ensure full disclosure . Data was used for the triangulation between documentary sources and all of the interviews, ensuring its validity and reliability, mainly using interviews as an initial indicator of the main issues and positions (Singer 2009. If there were contradictions or differences in interpretation, priority was given to written records .
In total approximately 60 interviews were carried out, 50 in Argentina and 10 in Armenia between October 2012 and September 2014. In addition around 150 documents from both public and expert sources were analyzed in the same time period.

Process tracing
The method of process tracing "attempts to trace the links between possible causes and observed outcomes" (George and Bennett 2005, 6), focusing on sequential processes within a particular case. By providing theory-based historical explanations for a case in which each significant step toward the outcome is explained by referring to a theory, process tracing becomes a powerful method of inference (George and Bennett 2005, 30). Each process tracing account proposes evidence to be found in case the account is true. While some evidence could explain more than one alternative there is evidence that can only be found in case the proposed theory is true, when more of the latter type of evidence is found the stronger the process tracing account (Bennett and Elman 2006). Process tracing has been widely used as a fundamental tool of qualitative analysis in international relations and public policy analysis (Collier 2011). In the specific case of environmental and forest policy analysis process tracing has been used in transnational environmental governance to study the mechanisms of influence of forest and fisheries certification  as well as the emergence and institutionalization of legality verification as a means to address global forest degradation (Cashore and Stone 2014).
In this thesis process tracing was used to understand how international and transnational organizations influence policy change in the domestic level through direct access and market mechanisms. To do so the cases of the World Bank-funded forest projects and forest policymaking in Argentina from 1991 to the present, the World Bank-funded natural resources management and poverty alleviation projects and forest policy making in Armenia from independence to the present and the development of two competing forest certification schemes in Argentina from 2001, when forest certification was first put in the political agenda in the country, to the present were selected. In all cases data was collected from different sources and analyzed by means of qualitative content analysis.

Qualitative content analysis
Qualitative content analysis is a flexible method for analyzing text data (Hsieh and Shannon 2005. This method focuses on the language as a means of communicating with attention to the content of a text (Hsieh and Shannon 2005). This method has been widely used in political sciences to analyze documents and interviews since it is the most appropriate method for revealing in an objective way the most significant text, word or symbol from a large volume of text . The large quantities of text that is analyzed requires that is organized into fewer content categories. This translation from data into categories is defined as coding (Fabra-Crespo and Rojas-Briales 2015) Qualitative content analysis has been widely used in forest policy analysis. Some recent examples of qualitative content analysis in forest policy analysis are studies on media reporting about forest issues as well as biodiversity (Sadath et al. 2013, Fabra-Crespo andRojas-Briales 2015), studies on REDD+ and drivers of deforestation (Wehkamp et al. 2015, Fischer et al. 2016) and studies on the implementation of the international forest regime (Wibowo andGiessen 2015, Sahide et al. 2015;. In this thesis qualitative content analysis was used in all constitutive articles to analyze all data. In order to identify the most relevant international forest issues in Argentina, this method was used in a two stage approach, distinguishing between public and expert sources in both stages. For the public sources, the first stage consisted of an open search in a broadly distributed national newspaper (La Nación) that was selected for eliciting the relevant issues. The terms "international", "forest" and "forestry" were searched with the search engine of the newspaper between the years 2008 -2012. For the expert sources, an open search on professional journals in the field of forestry (Argentina Forestal, Producción Forestal) was done. The second stage, after main actors and issues were identified, included a more focused search. For public sources, websites and position papers of the different actors towards the identi-fied issues were searched. For expert sources, interviews to people with broad knowledge of the forest sector in Argentina were made. To each one of them a question was made: Which international forest policy issues were relevant in Argentina between the years 2008 -2012? Interviews were used as an additional source to verify the data from other. In the analysis, forest-relevant actors were coded according to Krott (2005) into forest users (forest owners, forest workers and employees, general population), associations and parties (forestry associations, environmental NGOs) and government and administration bureaucracies.
In the cases of the World Bank intervention in forest and natural resources policy change in Armenia a documented search of publicly available World Bank proposal documents and project completion reports was done as a first step. Using these documents the executive bureaucracies involved in the projects were identified. Websites of these bureaucracies were then reviewed to obtain published outputs of the projects. Subsequently semi-structured interviews with experts were conducted about the role and influence of international actors and domestic state and non-state actors in the process. Qualitative content analysis was then used to analyze all documents and interviews. The information obtained from the different data sources was then coded and grouped into four main categories, based on the research objectives. These categories were (i) administrative structure, (ii) policy content, (iii) project formulation, and (iv) project implementation.
Qualitative content analysis was also used to analyze all documents and interviews for revealing the role of state agencies in the institution and the setup and operation of two forest certification schemes (FSC and PEFC) and their standards in Argentina.
Which issues become relevant in different countries depends on the domestic actors and their interests. In order to become influential, a political issue needs to be put on the agenda (Howlett and Ramesh 1995). In the politics surrounding the issue, a number of actors concerned will then take positions based on their interests (Krott 2005. This was observed in Argentina, where  identified the main actors and their positions towards the main forest issues. As a result they found that based on the amount of issues each actor had a public position on, environmental NGOs were the most active actors. However, based on the role played in the implementation of the issues, the authors found that the Secretariat of Environment and Sustainable Development was the most relevant actor by being involved in the implementation of all issues detected . This result shows the significance of main bureaucracies in dealing with international forest issues in Argentina as suggested by the bureaucratic politics theory . According to this theory, only a single actor has a dominant role in a particular sector, however,  found that in some issues like desertification and forests and forest certification both, the Secretariat of Environment and Sustainable Development and the Ministry of Agriculture, Livestock and Fisheries seem to work together. Similar results were found in Indonesia  and in Bangladesh (Rahman and Giessen 2014). However, although based on their public positions these two bureaucracies showed similar interests, when ongoing politics were analyzed a different situation was observed Giessen 2016, Burns et al. 2016).
In the case of forest certification, while in a first stage both, the Secretariat of Environment and Sustainable Development and the Ministry of Agriculture, Livestock and Fisheries supported the creation of FSC national standards through informational instruments as well as the provision of staff being involved in the negotiation of the standards, their positions changed as soon as the development of a competing system, PEFC, started . By refusing to take part in the association that administers PEFC in Argentina, the Secretariat of Environment and Sustainable Development shows their lack of support for this system in the country. Opposite behavior was observed from the Ministry of Agriculture, Livestock and Fisheries who not only takes part in the governing body of PEFC in Argentina but also financed the endorsement of the national standard to PEFC . The support to the devel-opment of a national standard for PEFC provided by the Ministry of Agriculture, Livestock and Fisheries in coalition with the three national forest associations, i.e. the Argentine Forest Association (AFoA), the Argentina Federation of Wood and Allied Industries (FAIMA) and the Cellulose and Paper Manufacturers Association (AFCP) led to the success of the process, which in 2014 acquired the mutual recognition of PEFC. On the other hand, the FSC initiative with the support from environmental NGOs and the Secretariat of Environment and Sustainable Development has not yet succeeded in the development of a national standard ). This case supports our hypothesis that while the opposing public positions of different actors towards an issue determine which issues are made relevant in a country, it is the interest of the most powerful actors that determine which issues are actively implemented and how.
A similar situation was observed in the case of forest biodiversity in Argentina identified as one of the main forest issue elements with support from all environmental NGOs as well as from the Secretariat of Environment and Sustainable Development . This support led to the approval of the Convention on Biological Diversity (CBD) by a national law and the establishment of a committee within the Secretariat of Environment and Sustainable Development to follow up the convention developing a national strategy about biological diversity. However, Argentina has not approved the Cartagena Protocol on Biosafety. An explanation for this decision can be found in the fact that Argentina is one of the main world producers of genetically modified organisms, ranking second in the world in 2006. Therefore, most private agricultural sectors, opposed the ratification of this Protocol . In this case, the same as with the case of forest certification, it was the interests of the most powerful actors, in this case the private agricultural sector in coalition with the Ministry of Agriculture which determine which issues are implemented at the national level.
While the public position towards an issue and the conflicting interests of different actors make an issue relevant and visible empirically, the issues which are not made relevant and the reason why they are not made relevant cannot be detected by analyzing the position of the different actors. In the case of Argentina this was observed by  when identifying the most relevant forest issue elements in the country. The analysis carried out by the authors did not identify deforestation and soy bean expansion as an international forest related issue despite its importance (Pengue 2005, Leguizamón 2013. The explicit blocking of certain issues, especially by powerful actors, from the international forest regime complex at the domestic level should also be analyzed in order to understand the implementation of the regime. In the analysis of forest-relevant actors and their positions towards different issues from the international forest regime complex in Argentina actors were classified according to Krott (2005) into forest users (forest owners, forest workers and employees, general population), associations and parties (forestry associations, environmental NGOs) and government and administration bureaucracies (Burns and Giessen 2014). However, this analysis and classification of actors was blind to the role of powerful international organizations as the World Bank. Although their analysis identified the support by international organizations in forest-related initiatives as an issue, the World Bank was not identified as an actor. However,  showed the important role this international organization had in forest policy change in the country.

The World Bank and domestic forest policy
In a study that aimed to analyze the influence the World Bank exerted on domestic forest policymaking and bureaucratic reform in Argentina,  found that by means of funding and capacity building, the World Bank, as an international organization, assisted the coalition of agricultural and profit-oriented bureaucracies and landowner associations (of forestry and agriculture) that emerged after the economic crisis in the country at the beginning of the 1990s. The intervention of the World Bank in the reform of forest policy in Argentina figured prominently in the form of new forest laws that, as a whole, benefited plantation forests and opened land for agricultural production . This supports our hypothesis that international organizations like the World Bank try to influence forest policy by the provision of funding and capacity building to produce domestic policy change. However, in order to become influential at the national level such international organizations as the World Bank need an invitation from a domestic bureaucracy due to sovereignty principles. In the case of Argentina, national forest bureaucracies who have been weakened in relation to strong and emerging agricultural bureaucracies sought assistance through funding from the World Bank. As a result of this assistance the bureaucracies of both native forests and plantation forests were able to increase their power, at least in the short term.
After the Argentine National Forestry Institute (IFONA), a comprehensive and autonomous bureaucracy responsible for all aspects of forests at a national level was dismantled accused of inefficiency and corruption, the Directorate for Native Forests, within the Secretariat of Environment, became the weakest forest bureaucracy in Argentina but gained power later. By forming a coalition with the World Bank, and with the support of environmental NGOs the Directorate wrote a new forest law (Law 26331). This Law granted this bureaucracy permanent funding from the national budget, which allowed it to double its staff after two years of implementation. In addition, this bureaucracy produced the first native forest inventory, in coalition with the World Bank, creating a new unit within the Directorate that would be in charge of the inventory and of monitoring the forest cover. The information produced with this inventory was of great utility for this bureaucracy at the time of implementation of the new law and of great importance for the agricultural policy sector, since it strived to gather information on forests that could be transformed into agricultural uses. These results are in line with the theory proposed by  that states that influence from international bureaucracies through the provision of funding resources could shift the power balance in domestic policy processes, benefitting marginalized or disempowered bureaucracies. Weak bureaucracies, seeking to increase their power, invite and allow international bureaucracies to influence the policy-making process.  also propose that, by helping domestic bureaucracies enforce or implement their own laws by means of direct access, international organizations are more likely to succeed than by passing new legislation. Our results show that, in the Argentinian case, the World Bank was an ally in the formulation and implementation of new laws. However, these laws were in accordance with the requirements of domestic agriculture bureaucracies and of broader agricultural interests, as  suggest.
Because domestic bureaucracies like the Secretariat of Agriculture and the Secretariat of Natural Resources were in permanent competition, they invited the World Bank with the aim of increasing their domestic power at the cost of their domestic competitors. This invitation led to the establishment of influence through direct access of the World Bank on domestic forest policy and administration . This supports our hypothesis that in addition to providing resources, international organizations like the World Bank need to build coalitions with national bureaucracies in order to be able to produce policy changes. By providing funding and capacity building they assist the existing domestic coalitions of bureaucracies and associations in their constant competition for power, or help create new ones.
In a similar study conducted in Armenia, Burns et al. (unpublished) found that World Bank interventions aiming to reduce rural poverty by means of improving natural resources management also led to policy and administrative changes. As a result of the reform promoted by the World Bank through direct access by means of incentives power in this country the forest administration was restructured and new legislation creating private forests approved (Burns et al. unpublished). This came together with a change in forest management, away from pure conservation goals and towards production use of the forests. Along this change, the forest administration, which used to be under the Ministry of Nature Protection, was moved to the Ministry of Agriculture. As observed in the case of Argentina, the forest administration in Armenia increased its budget. However, this administration saw a reduction in policy domain, since control activities were given to the Ministry of Nature Protection, who also kept forest inventory and planning responsibilities. As a consequence, the forest bureaucratic structure reform promoted by the World Bank led to the fragmentation of the forest policy domain with weakened forest bureaucracies but with increased budgets, at least during the duration of the project (Burns et al. unpublished). These results are in line with the findings from the forest sector reform in Argentina, where funding from the World Bank led to the dismantling of an autonomous forest bureaucracy into four different units within different ministries . Similar findings were observed with the forest sector reform in Mali where also as a result of the World Bank intervention the National Forest Administration lost its power, losing responsibilities and budget. In this case, the forest bureaucracy was not dismantled and fragmented like in Argentina and Armenia but weakened by the World Bank promoting the devolution of forest management to the local level and to a newly created bureaucracy funded by the World Bank, responsible for creating efficient markets .
The World Bank, as an international organization, has assisted more than 50 developing countries in the development of their environmental, legal, and institutional capacities (Marschinski and Behrle 2009). Similar patterns of change as the ones observed in Argentina and Armenia were observed in Mali (Gautier et al. 2013), Cameroon (Singer 2009), Brazil (Lele et al. 2000, and Lithuania , among others. In all these cases, the World Bank compensated bureaucracies for their loss in power by means of incentive power.
These results support our hypothesis that for the World Bank to become influential in domestic forest policy they need to build coalitions with national bureaucracies by providing funding and capacity building that national bureaucracies use to get more power resources in their constant competition with other bureaucracies.

The World Bank leading to deregulation and weak bureaucratic structures
The influence of the World Bank in Argentina and Armenia through direct access by funding projects aiming at poverty alleviation and sustainable natural resources management pushed the forest sector towards deregulation (Burns and Giessen 2016, Burns et al. unpublished). In both cases funding by the World Bank started after neoliberal reforms were taken in the political economy of the countries. Neoliberalism is one of the most influential discourses of our time, which gained strength as a normative discourse in the 1980s after the US and UK adopted neoliberal policies ). The World Bank and the International Monetary Fund have promoted neoliberal ideas like trade liberalization, privatization and deregulation in many developing countries ). In Argentina, in the 1990s, the new neoliberal discourse paved the way for the coalition of agricultural interests, together with the World Bank, to influence forest policy and administration structure, through direct access . A same process was observed in Armenia where, following a big economic crisis which started after achieving independence when the United Soviet Socialist Republics (USSR) was disintegrated in 1991, structural reforms were applied driving a liberalization of the economy, small and medium privatizations and the formulation of a legal administrative framework for market economy. These market-friendly economic reforms, as observed in Argentina, received important support from the World Bank (Burns et al. unpublished). However, in post-socialist countries, as the case of Armenia the implementation of these reforms was especially challenging due to the previous system of complete state control, with limited private property and centralized administrative systems (Otto and Chobotová 2013). During the transformation process that started after the Cold War, World Bank lending to postsoviet countries went to those undertaking big Structural Adjustment Programmes aiming to eradicate poverty by promoting sustainable social and political changes together with economic reforms (Davies and Woodward 2014), as was observed in the case of Armenia (Burns et al. unpublished). Such structural adjustment lending provided economic incentives to economic sectors, such as forestry and agriculture, to restructure government activity, shift activities to the private sector and downsize bureaucracies (Davies and Woodward 2014, 254).
Following a framework of neoliberal conservation governance, promoting market-based mechanisms that aim to motivate local users to address environmental concerns by means of incentives, the World Bank promoted the intensification of natural resources production both in the case of Argentina and Armenia. Although seen as a fundamental step towards sustainability, these neoliberal conservation policies have been criticized for informally promoting decentralization, privatization and commodification that allow corporations and international actors to increase their influence over local natural resources (Levine 2002, Fletcher 2010. As a result of the World Bank intervention, both in Argentina and Armenia, new forest laws were approved. These new laws implied a change towards production use of the forests and away from conservation. In the case of Armenia, the new Forest Code established that forests and forest lands can be under state, community and private ownership. This change allowed privatization of lands in cases where land users would afforest or reforest. Until this change, according to the previous Forest Code, land was all state owned. The Code also introduced the category of forests of production significance where wood production activities are allowed, with the approval of forest management plans. This new category of forests produced a changed towards an economic use of the forest from a pure conservation aim (Burns et al unpublished). Legislation and regulations dealing with biodiversity conservation in protected areas were also changed as a result of the influence of the World Bank. In this case a new law on Specially Protected Natural Areas was adopted allowing protected area administrations to generate and retain income (Burns et al. unpublished). The same changes were observed in Argentina where a new forest law (National Law 26331) that formally promotes the conservation of native forests through land-use planning was approved as result of the intervention of the World Bank . This Law establishes three conservation categories for the native forests: Category I (red) includes high conservation value areas where no forest management is allowed; category II (yellow) is made up of medium conservation value areas subject to forest management with an approved sustainable forest management plan; and category III (green) comprises areas of low conservation value that can be converted into other land uses, like agriculture (similar . While it provides ample leeway for promoting agriculture at the cost of native forests, the law requires the planning of provincial native forestland use to be based on a participatory process that includes actors from outside the forest sector. Hence, strong policy sectors, like agriculture, were able to exert great influence on the results of the categorizations. As a result, this law opened the way to land use changes in the green areas, where native forests could now be converted into any one of other land uses, like agriculture, which provide provinces higher returns than the compensation received through conservation in red areas (Burns and Giessen 2016).
By funding projects with the formal aims of promoting sustainable natural resources management under the framework of neoliberal conservation governance the World Bank also promoted an administrative reform of the forest sector.  found that in the case of Argentina, this reform of the forest sector led to the fragmentation and weakening of national forest bureaucracies in relation to strong and emerging agricultural bureaucracies. The dismantling of IFONA which led to the separation of native and plantation forests, including plantation forests in the Secretariat of Agriculture, is in line with the empowerment of the agriculture sector. Although the newly created Directorate for Native Forests was able to increase its power due to its coalition with the World Bank, it still represents a reduction in scale, from an autonomous forest bureaucracy to one unit within a secretariat of environment. In Armenia, similar results were observed by Burns et al. (unpublished). Although in this case the forestry administration was not dismantled as in Argentina, it was moved from the Ministry of Nature Protection to the Ministry of Agriculture. This change was in line with the changes in forest management objectives towards production forests. Following the structural change carried out with support from the World Bank, the responsibility for control was taken away from the forest administration and given to the Ministry of Nature Protection, except for illegal logging controls that were kept under the responsibility of the forest administration. Control functions were then given to three organizations, the Forest State Monitoring Centre (FSMC), the newly created State Forest Service (SFS) and the State Environment Inspectorate (SEI), all within the Ministry of Nature Protection. This overlap between different administrations weakens the control activities (Elmqvist and Rylander 2010). These results are in line with findings from the forest sector reform in Mali, where funding from the World Bank also resulted in a weakening of forest bureaucracies ).
Our results support our hypothesis that despite the formal goal of sustainability and forest conservation, the World Bank supports market-based conversion of the national forest sector and pushes institutional structures towards deregulation. As a result there is a decline in power of forest bureaucracies leading to a weak bureaucratic structure unable to implement the new policy means like planning which should protect public goods in the forest and counter the effects of the market driven uses of the forest.

Transnational regimes and state bureaucracies
Forest certification started as a transnational regime, with the conservation-oriented private sector as the main driver. By means of the national implementation of the schemes, with the active presence of national conservation-oriented actors and little participation from national bureaucracies, these transnational regimes became influential at the national level.  observed this in the case of FSC in Argentina, where a national environmental NGO housed FSC and acted as coordinator of the development of the standards. However, despite what is often claimed , Pattberg 2005, Auld et al. 2008, state bureaucracies play a decisive and active role in forest certification processes. In Argentina,  found that forest certification through FSC was first put into the political agenda by the Secretariat of Agriculture through organizing a workshop about the potential implementation of FSC in the country. During the development of FSC national standards, both, the Secretariat of Environment and Sustainable Development and Secretariat of Agriculture, Livestock, and Fisheries participated as members of the environmental chamber. This shows that national forest bureaucracies supported FSC in Argentina, at least in the beginning, through both informational instruments and the provision of staff. By involvement in the negotiation of standards, national bureaucracies could try to influence the final standards according to their interests. However, national standards for FSC were never achieved, showing the great competition faced by this system (Espach 2006. As a response to the development of a conservation-oriented forest certification scheme of high standards like FSC, production-oriented private sectors created more industry-friendly systems that could compete with, and later even replace conservation-oriented systems. By forming coalitions with national state production-oriented bureaucracies, these competing schemes gained strength. In Argentina,  found that the formulation of national standards for PEFC endorsement, an initiative carried out by the forestry associations in coalition with the Ministry of Agriculture and the National Standardization bureaucracy, presented strong competition to FSC. In this case, the Ministry of Agriculture, by becoming allies with the forestry associations through the provision of informational instruments and technical expertise in standard development, and later by incentive instruments, such as paying for the PEFC endorsement procedure, supported the creation of a more industry-friendly scheme that could compete with, and preferably even replace FSC in the country. A similar situation was observed in Indonesia, where the Ministry of Forestry supported forest certification by funding a national initiative (LEI) competing with FSC (Bartley 2010, Wibowo et al. unpublished). One can argue that private landowners have more power in countries like Argentina and Indonesia with weak state capacities, however the same pattern was observed in countries with strong state capacity like Finland and Germany. In Finland, PEFC, with the support of forest owner organizations, out-competed FSC, with 95% of the forests in the country certified under PEFC only one year after the approval of the standard (Kestikalo et al. 2009). In Germany, PEFC was created by the forest sector in coalition with other landowners and managers across Europe as a response to the ENGO driven FSC. In this case PEFC had a strong support from almost all private and state landowners . As suggested by Öberg et al. (2011) corporatism does not depend on whether institutions are weak or strong but rather whether they are willing or not to engage in mutual exchanges.
These results support our hypothesis that other than often claim state bureaucracies play an important role in the implementation of transnational regimes at the national level. In line with these results fur-ther studies even suggest that state bureaucracies are now reclaiming their role in private governance by means of legality certification processes as well as other mandatory certification mechanisms (Sahide et. al. 2015, Burt et al. 2016).
In forest certification processes international actors seek coalitions with different state bureaucracies.  found that in Argentina both the European Union and the World Bank supported different bureaucracies in the formulation of national standards for PEFC endorsement, an initiative carried out by the forestry associations in coalition with the Ministry of Agriculture that emerged as strong competition for FSC, almost replaced it. By seeking coalitions with international actors, the Ministry of Agriculture (formerly the Secretariat) in coalition with the EU and the National Standardization bureaucracy in coalition with the World Bank, tried to enhance their power in the competition for policy domain on standard development of forest certification schemes. However, the forestry associations supported the National Standardization bureaucracy, which would not be permanently in charge, giving administrative power back to the landowners' sector. In the competition between these two bureaucracies, the Ministry of Agriculture could not achieve the policy domain of writing the national standards, and support from the EU was dropped. As concluded by , in situations where new promising venues for political influence open, private actors prefer ad hoc coalitions with weaker bureaucracies that would not be permanent and with less technical knowledge.

Federalism and International and Transnational forest regimes
The political system of federal countries like Argentina provides multiple institutional access points for policy change that international and transnational regimes try to use in order to influence the domestic level, consequently changing the power balance of the domestic networks . Domestic actors then can use these regimes to shift the power balance in their favor. However, how important federalism is depends on how responsibilities regarding certain issues are vertically distributed between the national and the subnational governments (Lopreite 2014). In cases where the issue at stake is within provincial competencies, the national level will try to use the international sphere to increase their responsibility (Scharpf 2006, Wibowo and. This is the case with forests in Argentina, whose responsibility toils within the provincial level . By making coalitions with the World Bank, national forest bureaucracies were able to approve a new national forest law providing for "minimum standards" regulation, meaning that every province must reorganize its provincial laws in order to meet minimum requirements . By the approval of this law, the Directorate for Native Forests increased their power in relation to the provincial level. As pointed out by Hofmann (2002), national bureaucracies benefit by gaining influence through the implementation of the International Forest Regime Complex (Hofmann 2002).
In the case of developing forest certification schemes in Argentina,  found that national bureaucracies, like the Ministry of Agriculture, tried to increase their power over the provinces by leading the national initiative of an industry-friendly scheme like PEFC with the support of private landowners. In this case, despite the provinces retaining the power over forest issues, national bureaucracies by means of transnational regimes increased their power. The provincial bureaucracies, on the other hand, try to retain their power in forest issues but at the same time support their stable coalitions with private actors in their territories. In the case of forest certification in Argentina this was seen both in the development of FSC and PEFC. In the case of FSC, they tried to obstruct the development of regional standards, especially in the regions were forest landowners are stronger, like the Mesopotamia region. At the same time they joined the development of the PEFC standard, being part of the CERFOAR association . In cases where the jurisdiction over an issue lies at the national level, the subnational level can still use agenda setting and bills initiation to influence policy change (Lopreite 2014).

Strategic options for stake holders
Based on the results from this dissertation the following strategic options are suggested to the main national forest relevant state bureaucracies in Argentina. These suggestions could also be applied to national state bureaucracies in other federal countries.
The main problem national state bureaucracies dealing with forests in Argentina face is that, except for national parks, which are under the jurisdiction of the national government, forests toile within the political responsibility of the provinces and are subject to provincial laws under the umbrella of national laws (Article 124 of the National Constitution 1994). This makes the subnational bureaucracies involved in the forest sector important actors competing with national state bureaucracies as well as potential coalition partners. Moreover, forests are mainly owned privately: 99.7% of plantation forests and 93.5% of natural forests are privately owned  making the private forest landowners another important actor in forest related issues.

Directorate for Forestry Production, Ministry of Agriculture
In facing the above mentioned problem, the Directorate for Forestry Production responsible for forest plantations, timber processing industries and agroforestry, within the Ministry of Agriculture can build coalitions with international and transnational organizations in order to change the power balance towards national level.
The World Bank, as an international organization promoting private sector investment in forests as well as new environmental markets and voluntary private sector regulation by means of neoliberal elements ) is an excellent coalition partner for the Ministry of Agriculture and specifically for the Directorate for Forestry Production, supporting their stable coalitions with private forest landowners. A coalition with the World Bank will not only increase their power by means of funding but also by a high level of political support and credibility to foreign investors. An increase in investments in plantation forests and related industries would lead to an increase in the representation of the forest sector in the national GDP and hence an increase in power of the Directorate for Forestry Production vis-à-vis other national state actors.
National state bureaucracies can also use international issues to strengthen their positions vis-à-vis competing agencies, both at the national as well as the provincial level, by taking up issues, discourses and norms to legitimize their roles as well as by finding domestic or international coalition partners (Bernstein and Cashore,2012;Giessen 2013, Wibowo and. As an example the Directorate for Forestry Production can use the emerging bioeconomy discourse (Pülzl et al. 2014, Kleinschmit et al. 2014) as well as recent commitments of reduction of emissions made by the country in the climate change convention (UNFCCC) to promote bioenergy with wood biomass from plantation forests as a source of renewable energy and hence increase their responsibilities in policy domain.
The climate change convention (UNFCCC) and the Convention to Combat Desertification (UNCCD) are examples of international agreements the Directorate for Forestry Production can use to increase their power. Even though the Directorate for Forestry Production is not the application authority of these agreements they can benefit from them, for example, by promoting plantation forests as important car-bon sinks and developing clean development mechanism (CDM) projects with plantation forests in areas where normally plantation forests are not established. As well they can promote plantation forests as means of rehabilitation of degraded lands and fighting desertification.

Directorate for Native Forests, Secretariat of Environment and Sustainable Development
As well as the Directorate for Forestry Production, the Directorate for Native Forests can build coalitions with international and transnational organizations in order to change the power balance towards national level.
The World Bank and the Global Environment Facility (GEF) can also be important coalition partners for the Directorate for Native Forests through funding of forest biodiversity conservation projects and poor communities living in the forests, in line with the World Bank´s forest strategies. This coalition with such strong actors will increase the power of the Directorate for Native Forests. Such was the case with previous projects with funding from the World Bank and GEF that led to a new national forest law which grants this bureaucracy permanent funding from the national budget, allowing it to double its staff after two years of implementation.
Most international forest agreements are focused on native forests, competence of the Directorate for Native Forests. By using these agreements this bureaucracy can increase its policy domain and hence its power. Such is the case of the climate change convention and especially REDD+. By developing a REDD+ national strategy with potential international funding this bureaucracy can increase its power vis-à-vis other state and non-state actors competing for forest land.
Transnational regimes can also be uses by national state bureaucracies to increase their power. Such is the case of forest certification. While in Argentina so far FSC, a conservation-oriented system, has been almost replaced by PEFC, a more industry-friendly scheme developed with the support of the forestry associations in coalition with the Ministry of Agriculture and the National Standardization bureaucracy a coalition between the Directorate for Native Forests and international and domestic environmental NGOs could increase the influence of FSC in the country. Taking an active role in the development of national standards for FSC could provide the Directorate for Native Forests some influence in the management of plantation forests, increasing their policy domain.

Introduction
Since the Rio Summit in 1992, a growing number of international policies and regimes have evolved aiming to govern forests globally. This "set of implicit or explicit principles, norms, rules and decision-making procedures around which actors' expectations converge in a given area of international relations" (Krasner 1982, 186) can be described as an international forest regime complex Arts and Babili 2012;. Such a regime complex, for influencing domestic policies at national levels, depends on international and especially domestic coalition partners . These can include international as well as domestic state and nonstate actors. Special attention, however, must be paid to domestic bureaucracies, as they have to invite international actors due to sovereignty principles and in the domestic policymaking process they are instrumental in developing policies including also regulatory instruments. A research program on international forest policy that examines the pathways and politics through which international institutions influence such domestic policy change is currently being developed (Arts and Babili 2012; Bernstein and Cashore 2012). The analysis of the interaction among international regimes, their actors, and domestic bureaucracies is especially crucial McDermott, Cashore, and Kanowski 2010). It is in this interaction that the politics of competing bureaucracies in the acquisition of new tasks and responsibilities for implementation occur.
During the authoritarian regime that ended in 1983 in Argentina, the foreign debt multiplied sevenfold, increasing the importance of international creditors as political actors in the country (Carranza 2005). By the end of the 1980s, Argentina was in the middle of a severe economic crisis triggered by hyperinflation, when most of the powerful international banks in the country provoked a run on the local currency (Carranza 2005). As a solution to this crisis, the neoliberal economic reforms demanded by international and domestic actors were applied (Carranza 2005;), leading to a new agro-export policy. As in other Latin American countries, the World Bank played an important role in these reforms, increasing its presence and influence after the crisis ). In the forest policy context, domestic agricultural bureaucracies invited the World Bank to assist in the development of the forestry sector, based on its potential for economic growth . This article argues that by inviting the World Bank, domestic bureaucracies sought funding from an international partner in order to increase their power vis-à-vis competing administrations, and in this way formed coalitions with the World Bank.
Considering this background, this article argues that in the midst the economic crisis of the early 1990s, a strong coalition of (agricultural and profit-oriented) bureaucracies and associations of private landowners (in forestry as well as agriculture) emerged, advocating national development based primarily on more profitable and productive land use. This led to a predisposition in favor of changes toward more profitable land uses and the respective supporting policies and administrative frameworks. These eventually resulted in dismantling the Argentine National Forestry Institute (IFONA), a rather strong and comprehensive forest bureaucracy, into weaker units, so as not to obstruct transformations to more profitable land uses such as timber plantations and soy production. In order to facilitate this process further, the domestic coalition was then enhanced by the presence of such international actors as the Food and Agriculture Organization (FAO) and, most prominently, the World Bank. This was in order to provide more support for these neoliberal developments by making concrete recommendations for policy development. Consequently, we argue, the World Bank played a crucial role upon joining and supporting the domestic coalition of pro-agriculture public and private actors.
Hence, the aim of this study is to analyze the influence of the World Bank on domestic forest policymaking, undermining forest bureaucracies and policy with a bias toward the agricultural sector in Argentina, by exploiting a historical cleavage of neoliberal domestic politics combined with bureaucratic competition among government agencies searching for broader missions and budgets. In order to perform this analysis we employ bureaucratic politics theory from policy analysis , combined with the theorem of direct access from global governance theory . In so doing, this article poses the following main research question: How did a pro-agriculture coalition of international bureaucracies and strong national bureaucracies, together with their domestic allies, influence a full restructuring of administrative structures and policies affecting forestry in Argentina?

Theoretical Approach
The means by which international regimes, such as the international forest regime complex, may exert their influence at the domestic level can be categorized into "four pathways of influence, each with its own causal logic: international rules; international norms and discourse; creation of or interventions in markets; and direct access to domestic policy processes" (Bernstein and Cashore 2012, 587). It is "the interaction of mechanisms and processes, sometimes along multiple pathways, that create[s] collective influence" (Bernstein and Cashore 2012, 603). However, the type of interlinkages is so far undertheorized, leading to one possible relation among pathways, that is, that discourses do not directly influence domestic policy making but rather prepare the way for other pathways to occur.
Taking the pathways individually,  consider direct access as being the potentially strongest one; hence, this is the focus of this article. Direct access refers to the influence by means of direct funding, education, training, assistance, and capacity building . This pathway is linked to providing incentives to domestic policy actors, which, according to the theory of incentives (i.e., the principalagent model), can cause an actor's interest to change direction, and therefore functions as a critical power resource . Although it represents a very important form of influence, this path can be described as one of "infiltration" in the domestic policymaking process . It is seen as intrusion and raises issues of sovereignty . As a consequence, direct access is most likely to occur with the support of specific domestic state actors who bring critical legitimacy to the process, which is essential in overcoming sovereignty concerns . In exchange, those domestic state actors who invite outside actors gain additional sources of power, such as legitimacy and financial resources . In our analysis, these domestic state actors are defined as domestic bureaucracies.
A bureaucracy can be defined as "an agency that has been set up by governments or other public actors with some degree of permanence and coherence and beyond formal direct control of single national governments" (Biermann et al. 2009, 37). Bureaucratic politics theory claims that bureaucracies compete with each other for resources, staff, and responsibility for policy domains. This competition between bureaucracies with different preferences, abilities, and power capabilities shapes the resulting policy to a great degree . In their struggle for power, bureaucracies act in coalition with other public and private actors in fields where their interests (partly) overlap, trying to establish institutions in support of this shared interest. Being rather stable over time, this combination of shared interests and institutions is referred to as a policy sector, for example, as forestry or agriculture . In joining forces, policy sectors attempt to influence public policy based on the overlapping portions of their interests. In order to influence domestic policies, international regimes' bureaucracies are likely to seek coalitions with national bureaucracies. These coalitions can sometimes give international bureaucracies the opportunity of having direct access to domestic policy change.

Case Description
At the end of the 1980s, the newly elected President Menem designated the former Vice-President of Bunge & Born (the largest multinational firm in Argentina, specialized in food processing and grain and oilseed trade) as Finance Minister, as soon as he took office 1 (Carranza 2005). This designation was one out of many ways of showing his intent to implement the neoliberal reforms demanded by international and domestic actors (Carranza 2005;). These reforms were carried out under the Convertibility Plan, which aimed to end hyperinflation and promote growth in Argentina. In 1991, following the neoliberal discourse, the Deregulation Decree (Table 1) was signed, turning the political economy toward a neoliberal policy, tied to an agro-export policy 2 (Carranza 2005;Leguizamón 2013). This decree aimed to promote trade by ending protective regulation of the domestic economy, dismantling all bureaucracies that were in charge of these activities and simplifying the tax system (Leguizamón 2013). In this process, the Argentine National Forestry Institute (IFONA), a comprehensive and autonomous bureaucracy responsible for all aspects of forests at a national level, accused of inefficiency and corruption (Interview 1, Interview 2, Table 2), was dismantled, dividing its functions and staff between four bureaucracies: (i) the newly created Directorate for Native Forests, responsible for native forests and contained within the also newly created Secretariat of Natural Resources and Human Environment; (ii) the Directorate for Forestry Production, responsible for forest plantations, timber processing industries, and agroforestry, and part of the Secretariat of Agriculture, Livestock and Fisheries; (iii) the National Parks Administration (APN), a semiautonomous agency within the Secretariat of Natural Resources, responsible for protected areas at the federal level; and (iv) the National Institute for Agricultural Research (INTA), a semiautonomous research and extension entity that includes a forestry research unit within the Secretariat of Agriculture (World Bank 1996) (Figure 1).
These market-friendly economic reforms were praised by international financial institutions and institutions of global governance, like the G-8 (Carranza 2005). This led to an increased presence and influence from the World Bank in the country, with the World Bank becoming a very important actor in the domestic reforms by diversifying its funding and by means of capacity building through knowledge .
The new agro-export policy was geared toward a few commodities for the export market (Leguizamón 2013)   1990 and 1999, benefitting mainly big international oilseed firms (Teubal 2011). The five biggest companies accounted for 38.7% of the oilseed exports in 1990, whereas the proportion increased to 57.9% in 199857.9% in , and to 80% in 200257.9% in (Teubal 2011. At that time, foreign companies appropriated large farms, and had bought around 16.6 million ha by the year 2005 (Pengue 2005). These changes in the bureaucratic structure of forestry and the intervention of the World Bank through the funding of different projects in a sector competing with the agricultural sector make this an appropriate case for the study of the direct access pathway of influence exerted by international bureaucracies, domestic bureaucratic politics, and policy change.

Empirical Methods
To understand how the World Bank influences policy change through direct access, we conducted a qualitative case study of World Bank-funded forest projects and forest policymaking in Argentina from 1991 to the present. To do so, we employed process tracing . The method of process tracing "attempts to trace the links between possible causes and observed outcomes" (George and Bennett 2005, 6), focusing on sequential processes within a particular historical case. By providing theory-based historical explanations for a case, in which each significant step toward the outcome is explained by making reference to a theory, process tracing becomes a powerful method of inference (George and Bennett 2005, 30). We collected data from different sources with the aim of determining whether the direct access by international bureaucracies, together with competing national bureaucracies and national non-governmental organizations, can be related to the restructuring of the national forestry administration and forest policy change. As a first step, we completed a document search of publicly available sources to gather World Bank proposal documents and project completion reports from 1991 to the present. Using these documents, we identified executive bureaucracies involved in the projects. We then reviewed the websites of these bureaucracies to obtain published outputs of the projects. We subsequently conducted semistructured interviews with experts about the role and influence of international actors and domestic state and non-state actors in the process. Interviews lasted between a half hour and 2 hours. If it was not possible to meet with the interviewees in person, we used phone calls or e-mails. In all cases interviewees were promised confidentiality in order to ensure full disclosure . In total, we conducted nine interviews during October 2013 and April 2014 (Table 2). Seven interviewees came from domestic state agencies, one from domestic nonstate organizations, and one from the Food and Agriculture Organization. We used all the data collected for triangulation between documentary sources and all the interviews, providing valid and reliable data .
We used a qualitative content analysis method to analyze all documents and interviews . We coded the information obtained from the different data sources and grouped it into four main categories, based on the research objective. These categories were (i) administrative structure, (ii) policy content, (iii) project formulation, and (iv) project implementation. The outcomes (categories i and ii) and the process (categories iii and iv) were analyzed using these four categories.

Results
In Argentina, the World Bank, as an international bureaucracy, by means of funding and capacity building, assisted the coalition of agricultural and profit-oriented bureaucracies and landowner associations (of forestry and agriculture) that emerged after the economic crisis at the beginning of the 1990s. By forming coalitions with the World Bank, certain bureaucratic units were able to increase their staff and budget and produce valuable information, thereby increasing their power. This agrees with our theory, in that international organizations can operate through the provision of funding and capacity building to assist the existing domestic coalitions of bureaucracies and associations in their constant competition for power, or to help create new ones. Because bureaucracies like the Secretariat of Agriculture and the Secretariat of Natural Resources were in permanent competition, they invited the World Bank with the aim of increasing their domestic power at the cost of their domestic competitors. This invitation led to the establishment of influence through direct access of the World Bank, as an international organization, on domestic forest policy and administration.

The Politics of the World Bank Intervention Process
At the end of the 1980s, IFONA invited the World Bank to fund a project to promote the forestry sector (Interview 1, Table 2). Negotiations began within IFONA, the World Bank, and FAO, and the "Forestry Sector and Environmental Protection Project" was conceived (World Bank 2006). However, the Deregulation Decree of 1991 dismantled IFONA as a comprehensive forest bureaucracy before the project started (Table 1), with the support of the Secretariat of Agriculture (Interview 1, Table 2). At the beginning, forest associations (e.g., AfoA, FAIMA, AFCP) did not assume a public position regarding the dismantling of IFONA (Interview 5, Table 2), although informal statements supported the measure, pointing out its inefficiency and corruption (Interview 1, Table 2). IFONA's functions and staff were then divided among four newly created bureaucracies (Interview 1, 2, 3 and 4, Table  2) that had to compete with each other for staff, resources and policy domain. A coalition with the World Bank, which involved funding accompanied by a high level of political support, would allow specific bureaucratic units to increase their staffs and budgets and to produce valuable information, thereby increasing their power.
After IFONA was dismantled, and with the challenge of promoting economic growth, the forestry sector was considered to hold the promise of potential for growth, mainly through plantation forests and by yielding land for soya production. The government, through the Secretariat of Agriculture, requested the World Bank to conduct a review of the sector in 1992 with the aim of narrowing the gap between existing and potential levels of production (Interview 7, Table 2, Table 3), which showed the forestry sector's increased productivity goals. As a result of this review, the national government, together with the World Bank and FAO, determined that the World Bank could provide significant assistance to "develop and implement an appropriate role of the national bureaucracies in forestry and conservation as part of their support for economy-wide policies and actions necessary for the improvement of Argentina's forestry potential" (World Bank 2006, 5). This led to the development of the first two forest projects, "Forestry Development Project" and "Native Forests and Protected Areas Project," with funding from the World Bank, which praised the trade liberalization taking place at that time and stated that future policies should continue in this direction (World Bank 1995;. At this time, the FAO-World Bank Cooperative Program, which involved technical experts from both the World Bank and FAO, played an important role through direct access by means of capacity building and by providing technical experts to assist in the project's formulation and supervision (Interview 7, Table 2; World Bank 2007). FAO technical experts, by providing technical expertise, preselect information and alternatives for political action, and by this preselection they increase their influence in the development of the projects. At this stage, foreign firms were identified, mainly from Chile, that were interested in investing in the Argentine forestry sector.
In 1996 and 1997 the first two projects were approved. They were implemented by different bureaucracies, and had as their main goal the revision of forest laws and regulatory framework, maintaining and deepening the reform processes taking place in Argentina at that time by means of persuasion, incentives (World Bank 1995;, and the generation of information in the form of forest inventories ( Table 3). At that time, the World Bank's forest strategy stipulated that funding of the forestry sector was "conditional on government commitment to sustainable and conservation-oriented forestry," which included an increase in production in plantation forests, shifting from native forests (World Bank 1991;. This strategy adopted a more preservationist and tougher approach to forest businesses, in comparison with the previous strategy from 1978 . Prior to the projects' completion, a third World Bank forestry project was proposed and later approved in 2008, the "Argentina Sustainable Natural Resource Management Project." This was to be implemented by three different bureaucracies, the Secretariat of Environment and Sustainable Development (SAyDS), the Ministry of Agriculture, Livestock and Fisheries (MAGyP) and the National Parks Administration (APN), with separate accounts and independent objectives, with a total funding from the World Bank of US$60 million (Table 3).

Losers and Winners of the World Bank Intervention Outcomes
Policy. The "Forestry Development Project," implemented by the Directorate for Forestry Production, led to the creation of Law 24857 on Fiscal Stability for Forest Activity, Law 25080 on Investment for Forest Plantations, and Law 25509 on User Rights (World Bank 2006; Table 3). These laws had the aim of facilitating investments in plantation forests without having to own the corresponding land (Table 1). Unlike previous laws, these laws do not require the replantation of trees after trees are cut (Interview 1, Table 2). This meant that previously forested areas could now be converted to other land uses, like agriculture, which was not possible before. These changes increased investments in the forestry sector, reaching US$2.9 billion between 1992 and 2013. Chilean companies made 66% of these total investments, while the Directorate for Forestry Production invested 13%, by means of incentives. 3 The area of plantation forests increased from 760,000 ha in 1987 (Denegri and Aguerre 1996)  An outcome of the "Native Forests and Protected Areas Project," implemented by the Secretariat of Natural Resources and Human Environment, through the Directorate for Native Forests and the National Parks Administration, was the drafting of new legislation on native forests (World Bank 2007). This led to the eventual approval of National Law 26331, which promotes the conservation of native forests through land-use planning (World Bank 2007; Table 1, Table 3). This law provides for "minimum standards" regulation, meaning that every province must reorganize its provincial laws in order to meet minimum federal requirements . In addition, this law establishes three conservation categories for the native forests: Category I (red) includes high conservation value areas where no forest management is allowed; category II (yellow) is made up of medium conservation value areas subject to forest management with an approved sustainable forest management plan; and category III (green) comprises areas of low conservation value that can be converted into other land uses, like agriculture (similar Gartland 2012; Romero 2012).
The responsibility for the classification of the forests into these three categories lies with each corresponding province, which has to produce a map of its native forests, classifying their parts into the three conservation categories. Because the law allows the provinces a high level of discretion in its interpretation, the economic, social, and political context of each province will influence the way it implements this . While the law provides ample leeway for promoting agriculture at the cost of native forests, it requires the planning of provincial native forestland use to be based on a participatory process that includes actors from outside the forest sector. Hence, strong policy sectors, like agriculture, were able to exert great influence on the results of the categorizations. 4 This law opened the way to land use changes in the green areas, where native forests could now be converted into any one of other land uses, like agriculture, which provide provinces higher returns than the compensation received through conservation in red areas. By January 2013, 21 of the 23 provinces had finished and approved their provincial laws, including the corresponding map requested by the national law (Interview 6, Table 2). Of these 21 provinces, four were still waiting for national approval from the Directorate for Native Forests at that time.
The law also establishes a fund (Fondo Nacional para el Enriquecimiento y la Conservación de los Bosques Nativos) to compensate landowners (private or public) who preserve the native forests according to their conservation status . The fund is to receive 2% of the agricultural export tax of primary and secondary agricultural products of each previous year, but no less than 0.3% of the general annual budget of the national government. According to the law, in 2012, 1.8 billion Argentinian pesos (US$400 million) should have been destined to the fund for forest conservation (70% for management plans and 30% for the national and provincial bureaucracies). However, only 250 million Argentinian pesos were actually given to this fund, approximately one-seventh of the established amount, showing the central government's commitment to this law (Interview 6, Table 2). In the period during which the bill was stalled in the Senate, environmental nongovernmental organizations (NGOs) and the major productive forestry sector organizations rallied to support it.
Another important outcome of the project was the realization of the First National Inventory of Native Forests (World Bank 2011; Table 3). The inventory has provided key data for the technical justification for the modification of the native forests law and an important instrument for its implementation (World Bank 2011). The ambitious definition of "native forest" used for the inventory, for example, excluded from the inventory a great part of the Monte ecoregion, where trees do not reach the altitude defined by the inventory. Therefore, these forests were not included in the area subject to categorization. In addition, the inventory would establish the patterns for future land use in these regions (World Bank 2011). The information produced by this inventory is also of great importance for the agricultural policy sector, since it strived to gather information on forests that could be transformed into agricultural uses.
In the year 2010, National Law 25080 was modified by Law 26432 to include, in one of its articles, that plantations are subject to the incentives only if their location is in accordance with what is established in the provincial maps of Law 26331 (Interview 9, Table 2). This means that, since the year 2010, the provinces have had to have their laws approved in order to get the incentives. This change in a law from the Ministry of Agriculture, as a consequence of a law from the Directorate for Native Forests, coincides with the theory proposed by , which states that influence from international bureaucracies, through the provision of funding resources, could shift the power balance in domestic policy processes, benefitting marginalized or disempowered bureaucracies.
Administrative Reform. During the implementation of the "Native Forests and Protected Areas Project," a restructuring of the environmental national bureaucracies moved the National Parks Administration to the Ministry of Tourism (Figure 1). At that time, the Secretariat of Natural Resources was renamed as the Secretariat of Environment and Sustainable Development (SAyDS) (World Bank 2007).
As a result of the "Native Forests and Protected Areas Project," the National Forest Monitoring Unit (UMSEF) was created (World Bank 2007; Table 3) and was then absorbed by the Directorate for Native Forests, which now finances and supports it as an official forest monitoring service (World Bank 2011). Law 26331, implemented as a result of this project, strengthened the Directorate for Native Forests via the creation of the fund. This law established that 30% of the fund goes to the national and provincial bureaucracies, allowing this bureaucracy to more than double its staff in two years, from 20 to 50 employees (Interview 3, Table 2), increasing its power, at least in the short term.
In 2009, during the implementation of the Argentina Sustainable Natural Resource Management project, the Secretariat of Agriculture, until then within the Ministry of Economics, became the Ministry of Agriculture, Livestock and Fisheries (Table 1 and Figure 1). The bureaucratic status of the agricultural administration, acquiring independence from a strong bureaucracy like the Ministry of Economics, shows the increase in power achieved.

Discussion and Conclusion
Previous research on bureaucratic reform in Latin America focused mainly on domestic factors as explanatory variables, with only few studies analyzing the role of external influences ). Our study supplements this research by focusing mainly on international influences, like the influence of the World Bank, as a complementary explanatory variable in the reform of the bureaucracy and policy of the forest sector that took place in Argentina. Our study shows that in Argentina, national forest bureaucracies have been weakened in relation to strong and emerging agricultural bureaucracies. By seeking assistance through funding from the World Bank, the bureaucracies of both native forests and plantation forests were able to increase their power, at least in the short term. However, agricultural bureaucracies were assisted by the World Bank at the same time and in greater amounts, receiving a total funding of US$620 million between 1991 and 2015, as compared to the US$102 million that the forestry sector received for the same period, half of which was also administered by the Secretariat of Agriculture. 5 This search for assistance gave the World Bank the opportunity to gain direct access to the creation of policies that in the end promoted plantation forests and opened land for agricultural production. These results are in line with findings from the forest sector reform in Mali, where funding from the World Bank also resulted in a weakening of forest bureaucracies , calling for a more systematic research on the declining power of forest bureaucracies due to an overarching neoliberal approach in international forest policy ).
After IFONA was dismantled, the Directorate for Native Forests, within the Secretariat of Environment, became the weakest forest bureaucracy. By forming a coalition with the World Bank, and with the support of environmental NGOs, the Directorate wrote a new law (Law 26331). This law granted this bureaucracy permanent funding from the national budget, which allowed it to double its staff after 2 years of implementation. In addition, this bureaucracy produced the first native forest inventory, in coalition with the World Bank, creating a new unit within the Directorate that would be in charge of the inventory and of monitoring the forest cover. The information produced with this inventory was of great utility for this bureaucracy at the time of implementation of the new law and of great importance for the agricultural policy sector, since it strived to gather information on forests that could be transformed into agricultural uses. These results are in line with the theory proposed by  that states that influence from international bureaucracies through the provision of funding resources could shift the power balance in domestic policy processes, benefitting marginalized or disempowered bureaucracies. Weak bureaucracies, seeking to increase their power, invite and allow international bureaucracies to influence the policymaking process.  also propose that by helping domestic bureaucracies enforce or implement their own laws by means of direct access, international bureaucracies are more likely to succeed than by passing new legislation. Our results show that in the Argentinian case, the World Bank was an ally in the formulation and implementation of new laws. However, these laws were in accordance with the requirements of domestic agriculture bureaucracies and of broader agricultural interests, as  suggest.
The World Bank, as an international bureaucracy, has assisted more than 50 developing countries in the development of their environmental, legal, and institutional capacities (Marschinski and Behrle 2009). The direct access pathway is not the only source of the World Bank's influence; there is also the fact that engagement by the World Bank provides credibility to foreign investors (Marschinski and Behrle 2009). In Argentina, international investments in the forestry sector reached US$2.9 billion in the period 1992-2013. Similar funding to that provided by the World Bank, but from a different actor, would probably not have resulted in the same degree of direct access. Further studies should focus on the relationship between the direct access pathway and the power of the actor using this pathway. This is because such actor-specific power functions as leverage for any means of direct access (e.g., capacity building, funding). This leads to the assumption that the same magnitude of direct access can cause quite different changes in domestic policymaking due to information means, trust and reputation, additional potential funds, other (dis-)incentives, or even coercive actions.
The invitation made by a domestic actor is of great importance in the process of direct access. Previous studies have shown that domestic actors and their networks can influence intergovernmental negotiations to secure international norms that support their interests . In Argentina, the agricultural sector is a strong policy sector composed of several associations (e.g., SRA, CRA) that supported the neoliberal discourse initiated in 1991 (Leon and Rossi 2003). Agricultural landowners, together with the forestry sector, supported all neoliberal measures, as well as the incentive policies encouraging plantation forests. The dismantling of IFONA, which led to the separation of native and plantation forests, including plantation forests in the Secretariat of Agriculture, is in line with the empowerment of the agriculture sector. Although the Directorate for Native Forests was able to increase its power due to its coalition with the World Bank, it still represents a reduction in scale, from an autonomous forest bureaucracy to one unit within a secretariat of environment.
Neoliberalism is one of the most influential discourses of our time, and gained strength as a normative discourse in the 1980s after the United States and United Kingdom adopted neoliberal policies ). The World Bank and the International Monetary Fund have promoted neoliberal ideas like trade liberalization, privatization, and deregulation in developing countries ). In the 1990s the new neoliberal discourse in Argentina paved the way for a coalition of agricultural interests, together with the World Bank, to influence forest policy and administration structure, through direct access. This result suggests that the pathway of international norms and discourse proposed by , rather than a stand-alone pathway, may be interpreted as a pre-stage preparing the way for the other pathways to appear. In our results, the discourse pathway prepared the way for the direct access pathway. Future studies should focus on the interlinkages of the different pathways, especially on the relationship of the discourse pathway with the other three pathways of influence (Arts and Babili 2012;. Forest certification has been so far described as a non-state, market-driven form of private governance developed by non-state actors. However, bureaucracies can influence the effects of these programs in many ways, using their competencies to either enhance or restrict non-state certification schemes. The objective of our work is to analyze the role state agencies played in the forest certification standard development processes in Argentina, including the related coalition building strategies. Our hypotheses are that forest certification is not an activity among merely private actors, but that state agencies do play a small but necessary role in creating favorable frame conditions, and in standard formulation and scheme establishment. Coalitions built around production-oriented state agencies and strong political sectors have a stronger influence in the formulation of standards of forest certification schemes than coalitions between conservation-oriented bureaucracies and environmental NGOs. In order to understand the process of standard formulation for FSC and PEFC in Argentina, data on the different schemes were collected from public statements from stakeholder organizations, written summaries of debates of schemes from both public meetings and meetings by the groups writing the standards, and statements from the agencies in charge of writing the standards. The results show that in Argentina, FSC was temporarily useful to the forestry sector and landowners as a quick solution and hence was supported even by utilitarian state agencies. In light of an evolving operational PEFC scheme four years later, this support by state agencies was dropped and shifted towards the more landowner-friendly PEFC. After this dynamic in the coalition building process, FSC was on a decline in Argentina.

Introduction
After the failed intergovernmental efforts to agree on a legally binding convention on forests at the Rio Summit in 1992, environmental nongovernmental organizations, led by the World Wildlife Fund (WWF) in coalition with other non-state actors, founded the Forest Stewardship Council (FSC) in 1993 as an instrument to promote sustainable forest management through certification (Tosun, 2012). As a response to this initiative, forest-based industries and private landowner associations, in coalition with productionoriented bureaucracies in many producer countries, created a more "industry-friendly" competitor scheme for forest certification. They eventually joined in a common organization founded in 1998, which later became the Programme for the Endorsement of Forest Certification (PEFC) (Dietrich and Tosun, 2013;. In the case of Argentina, FSC was established in 2001, supported even by private landowners, but has been followed by a recent initiative to transition to PEFC.
With this increased participation of non-state actors, through forest certification a new form of political steering has been promoted based on voluntary agreements. In the political science body of literature, forest certification has been so far described as a non-state, market-driven form of private governance developed by non-state actors and building on private institutions Pattberg, 2007). In international policy, it is currently referred to as transnational regimes (Pattberg, 2012) and more broadly discussed under the concept of forest governance (Agrawal et al., 2008;Art, 2014;.
These perspectives in political science highlight aspects relating to the role of private actors and private institutions and the way they gain authority vis-à-vis state failure to solve forest http://dx.doi.org/10.1016/j.landusepol.2015.12.005 0264-8377/© 2015 Elsevier Ltd. All rights reserved. problems, and they are accompanied by a perceived decline in power of forest related public actors and institutions (Ho, 2006;. This view, however, leaves largely underrepresented the role of the state in e.g., creating, enabling, or obstructing conditions for forest certification in general and individual certification schemes in particular. It further downplays the role of the multiple and often competing state bureaucracies in certification politics, which may be crucial as both resources of power and as coalition partners in the establishment process of a particular scheme vis-à-vis their competitors. This important role of specific state actors is reflected in theoretical innovations, leading away from forest governance but pointing towards a forest governmentality approach for political certification studies (Art, 2014). Recent empirical findings support this governmentality perspective and even suggest a crucial role for state agencies in forest certification and related politics Bartley, 2014;. According to this literature, state agencies can obstruct Sahide et al. 2015b) or support certification initiatives (Hysing, 2009;Bell and Hindmoor, 2012).
The aim of this study is to analyze against this background the so far under-researched role state agencies played in the consecutive FSC and PEFC standard development processes in Argentina, including the related coalition building strategies.

Theoretical propositions
Since the Rio Summit in 1992, a growing number of international regimes addressing forests have evolved into what can be described as an international forest regime complex ). An international regime can be defined as "a set of implicit or explicit principles, norms, rules and procedures around which actorsí expectations converge in a given area of international relations" . These regimes attempt to influence domestic policies at national levels. At the same time, with the increased participation of non-state actors, a new form of governance based on voluntary agreements was created. This new form of private governance has been referred to as transnational regimes. The main difference between international and transnational regimes is that in the latter, non-state actors generate the set of norms and rules instead of states (Pattberg, 2012).
For both international and transnational regimes to become relevant at the national level there has to be a local actor advocating for it and forming favorable conditions. Domestic actors may take advantage of new international norms and policy discourses in order to break into or re-shape domestic actor networks Sahide et al., 2015a). Previous studies have shown that international regimes may lead to changes in the relation of power within domestic networks, strengthening certain groups of actors while weakening others (Grugel and Peruzzotti, 2010;. Transnational regimes might have the same impact at the national level with domestic actors seeking coalitions with international actors in their pursuit for favorable policies. In this pursuit private actors may act individually (e.g., lobbying, personal contacts) but also collectively (e.g., through coordinated actions by their associations or by forming coalitions) (Tosun, 2013). In the case of coalitions with state bureaucracies this is referred to as corporatism, defined as "a mutually beneficial exchange between interest groups and government where some actors control something that others desire" (Öberg et al., 2011, 365). In this exchange state bureaucracies seek more responsibility in policy domain while controlling public expenditure and legislative power  and private actors seek to influence public policy that is favorable to them while at the same time providing consent, approval or active support of certain bureaucracies and their policies (Öberg et al., 2011). In federal countries, like Argentina, with strong subnational institutions, stratification into provinces provides multiple coalition options for both national as well as actors from transnational regimes expanding the arenas of contestation. This exchange between private and public actors is necessary even in private governance where state bureaucracies are claimed to play a small but necessary role (Hysing, 2009;. According to the bureaucratic politics theory, those national and subnational state bureaucracies being involved in such multilevel politics, compete with each other for resources, staff, and the responsibility for policy domains (Krott, 2005;. In order to increase their power, national and subnational bureaucracies are likely to seek coalition with other national and international actors (both public and private) having compatible interests, making up a policy sector . Policy sectors "focus on a specific area of public policy (forests, water, agriculture, etc.) and include all the groups, organizations, and institutional rules pertaining to that arena of policy making and implementation" (Shannon and Schmidt, 2002). In joining forces of their actors in a corporatist fashion, policy sectors attempt to influence public policy as coalitions. These coalitions can be of two types: (a) stable over time forming policy sectors; (b) ad hoc when new promising venues for political influence open (Rommetvedt et al., 2012).
Domestic bureaucracies may use international issues to strengthen their positions vis-à-vis competing agencies, both at the national as well as the provincial level, by taking up issues, discourses and norms to legitimize their roles as well as by finding domestic or international coalition partners . In federal countries, international issues shift the power balance towards national bureaucracies, even in issues that are actually within the provincial domain (Hofmann, 2002;. As a response to this shift in power balance, subnational bureaucracies try to get involved in international issues as well. How they will pursue this depends on how responsibilities regarding a certain issue are vertically distributed between the national and subnational governments (Lopreite, 2014) but national bureaucracies will remain in a better position. In cases where subnational bureaucracies have the competency over certain issues they might be stronger than in issues of national competences (Lopreite, 2014).
Both international and transnational regimes operate in a wide range of issue areas (Pattberg, 2012). Forest certification has been described as a transnational regime where non-state actors, by means of markets, establish international institutions made of norms and rules Pattberg, 2007;. Although forest certification schemes are considered as internationally formulated, the national level plays an important role in the formulation of national standards and their implementation (Keskitalo et al., 2009). FSC has one set of principles and criteria defined by FSC-International; however, the actual standards and criteria are defined by regional working groups coordinated by national chapters (Espach, 2006). PEFC, on the other hand, is based on the standards developed by each nation, which are then evaluated by PEFC International under certain minimum requirements (Tosun, 2012;. This process led to the development of two different coalitions, one advocating for conservation practices and the other with production-oriented interests.
Against this theoretical background of national and subnational level public bureaucracies trying to increase their influence over private governance initiatives, we formulate the following propositions: 1. Forest certification and state actors: Although it is often claimed that transnational regimes are an activity among merely private actors, state bureaucracies do play a small but decisive role. In forest certification, state bureaucracies play an important role as coalition partners bringing resources, in both FSC and PEFC.

Coalitions and weak bureaucracies:
a. Competing national bureaucracies seek coalitions with domestic as well as international actors to enhance their power vis-à-vis other bureaucracies and private actors. In the case of forest certification national bureaucracies seek coalitions with international actors as the EU, WB, FSC international and PEFC international to enhance their power in the standard development of specific forest certification schemes, which private conservation and production-oriented actors try to obstruct. 3. Domestic private actors form stable coalitions with state bureaucracies building up a policy sector. However, in cases of new venues for political influence they would prefer ad hoc coalitions. In the case of establishing certification schemes, private production-oriented actors, such as private landowners, prefer employing rather weak bureaucracies that are (i) not permanently in charge of standard development and (ii) less knowledgeable about technological details of forest certification.
• Federalism: Despite provinces being strong actors in forest issues, national bureaucracies gain relative power due to international influences as forest certification processes.

Case description
In Argentina, there are around 31.4 million hectares of natural forests divided into six ecological regions (SAyDS, 2005). Moreover, there are 1.2 million hectares of cultivated forests, 80% of which are located in the northeast of the country, in the Mesopotamian region (MAGyP, 2013). Argentina adopted a federal republican representative form of government. Except for national parks, which are under the jurisdiction of the national government, forests toile within the political responsibility of the provinces and are subject to provincial laws under the umbrella of national laws (Article 124 of the National Constitution 1994). At a national level, the Secretariat of Environment and Sustainable Development is responsible for natural forests, while the Ministry of Agriculture, Livestock and Fisheries is responsible for forest plantations. Forests are mainly owned privately: 99.7% of plantation forests and 93.5% of natural forests are privately owned . Hence, this provides an interesting case to analyze coalition options between strong private actors as landowners and national and subnational state bureaucracies.
In the early 1990s, in the middle of an economic crisis, a strong coalition of (agricultural and profit-oriented) bureaucracies and associations of private landowners (in forestry as well as agriculture) emerged, advocating for more profitable and productive land uses with subsequent supporting policies and administrative frameworks . At this time, international actors as the FAO and the World Bank played a crucial role supporting the domestic coalition of pro-agriculture public and private actors. As a consequence, the Argentine National Forestry Institute (IFONA), a strong and comprehensive forest bureaucracy, was dismantled into four weaker bureaucracies . By forming coalitions with the World Bank, some of the newly created bureaucracies were able to increase their power . The administrative reform and policy changes carried out with funding and a high level of political support provided by the World Bank led to increased investments in the forestry sector, mainly by international firms, especially from Chile (between 1992 and 2013, 66% of the total investments in the forestry sector was done by Chilean companies with three Chilean forestry firms dominating Argentinaís forest holdings and wood exports) (Espach, 2005;. Thus, being a good case to analyze coalition building between domestic and international actors. As part of the political changes taking place in the beginning of the 1990s the National Constitution was reformed in 1994. With this reform, provinceís sovereignty over their natural resources was stated clearly giving the provinces the right to legislate on forest issues within their territory. This makes the subnational bureaucracies involved in the forest sector important actors and potential coalition partners. However, with the new government that took over after the 2001 crisis, the national government increased their power vis-à-vis the provinces by means of new sources of revenue that were not shared between the federal and provincial governments, decreasing provincesí share of overall revenue and increasing provincial governmentsí dependence on the federal government (Levitsky and Murillo, 2008). This case of strong federalism with strong links between the national and provincial level provides a good case to analyze the role of national and subnational level as coalition partners in the development of forest certification schemes.
Between 2001 and 2002, Argentina began the development of forest certification processes, both FSC and PEFC. These processes continued until the year 2014 with the development of standards, providing an appropriate case for studying the role of different state actors, both national and subnational, and the coalitions made, in the development of forest certification schemes.

Empirical methods
In order to analyze the role of different state and private actors and their coalitions in the development of forest certification schemes, we used a qualitative case study approach employing process tracing , which "attempts to trace the links between possible causes and observed outcomes" (George and Bennett, 2005, 6), focusing on sequential processes within a particular case. By providing theory-based historical explanations for a case in which each significant step toward the outcome is explained by referring to a theory, process tracing becomes a powerful method of inference (George and Bennett, 2005, 30). Data from different sources were collected with the aim of determining the role of state bureaucracies in the development of two competing forest certification schemes, FSC and PEFC. As a first step, a literature search was done from professional journals and websites of certification bodies, environmental NGOs, forestry associations, and related bureaucracies, in order to gather data on the certification process and identify key actors. Subsequently, semi-structured interviews were conducted with experts having broad knowledge of private and state certification processes. Interviews lasted between half an hour and two hours. If it was not possible to meet with the interviewees in person, they were contacted by phone calls or emails. In all cases, interviewees were granted confidentiality in order to ensure full disclosure . In total, five interviews were conducted between October 2013 and August 2014. Data was used for the triangulation between documentary sources and all of the interviews, ensuring its validity and reliability . If there were contradictions or differences in interpretation, priority was given to written records . Qualitative content analysis was used to analyze all documents and interviews  for revealing the role of state agencies in the institution and the setup and operation of certification schemes and standards.

FSC as an interim solution for landowners
FSC was first put on the political agenda in 2001 by the Secretariat of Agriculture by organizing a workshop about a potential FSC in Argentina (FVSA, 2003, Interview 1). However, FSC first arrived in the country at the beginning of 2001, when an international oil company certified their plantation forests in the northeast of Argentina (Sanchez Bonifato, 2001).
By the end of 2001, the Argentinian partner of WWF, Fundación Vida Silvestre Argentina (FVSA), housed FSC in the country by providing office spaces (FVSA, 2003). At the beginning of 2002, different meetings and workshops were held that aimed at creating a working group responsible for drafting national plantation forest management standards. At this time, FSC International recognized the General Director of the FVSA as the contact person of FSC Argentina, the first step in the development of the national initiative (FSC, 2002). By June 2002, a National Working Group was created in a meeting held at the Secretariat of Environment and Sustainable Development (FVSA, 2002). This group had three chambers (social, environmental, and economic) with five members each; half of the members represented both environmental and social NGOs, 25% represented the forest companies, and the other 25% represented different bureaucracies (FVSA, 2003). This group was elected mainly for drafting the national standards for the management of plantation forests with the aim of getting FSC endorsement. Both, the Secretariat of Environment and Sustainable Development and the Secretariat of Agriculture, Livestock, and Fisheries participated in the development of the FSC national standards for plantation forests by being part of the environmental chamber (FVSA, 2003). The biggest transnational forest companies in Argentina, who operate large-scale and rapidly growing plantations for the production of pulp and sawn wood, participated in the development of standards in the economic chamber (FVSA, 2003), although so far none of them has sought certification. However, most of these companies cannot certify under FSC, due to FSC's prohibition of certification for any plantation established after 1994 on land that was previously native forest (Espach, 2006). This regulation is part of FSC's international set of Principles and Criteria and could not have been modified by the national chapter.
At the same time that the National Working Group was created, a Regional Working Group was appointed for developing forest management standards for the Atlantic Forest eco-region (FVSA, 2003). After the elective process of the Regional Working Group, representation conflicts arose (FVSA, 2003). To overcome this situation, the entire National Initiative became greatly involved in the dispute resolution. After many meetings between National Initiative delegates and the local stakeholders involved, including the recently elected members of the Regional Working Group, the parties reached consensus about initiating standard drafting for the natural forest of the eco-region, committing themselves to undergoing a revision of the representatives' election process within a six-month term (FVSA, 2003). However, this never actually happened, even though a consultant was hired for this purpose (FVSA, 2003). This situation reflects the strong opposition FSC certification faced in that part of the country, mainly from big Chilean companies that are the main investors in fast-growing species plantations in the area (Espach, 2006).
Both the national and regional initiatives were funded by WWF Switzerland and WWF International through their program Forest for Life (FVSA, 2003). A few months after the election of the Atlantic forest Regional Working Group, in coalition with two other environmental NGOs (Greenpeace and Proyungas) and with funding until the year 2006 from a private company in the oil and gas sector (Gasoducto Nor Andino Argentina SA), several meetings were held to launch the regional initiative of the Yungas ecoregion (FVSA, 2003). However, both efforts and funding for these three initiatives were insufficient for achieving the stated goals (Interview 1). Two other Ecoregional initiatives were proposed (Chaco and Bosques Andino-patagonicos), but without funding (FVSA, 2003).
The National Working Group divided the writing of the standards between the three chambers, assigning specific principles and criteria to each of them. Each chamber took a different approach, including the designation of a writing committee (Environmental Chamber), working in plenary sessions (Economic Chamber), and organizing workshops for stakeholders' input (Social Chamber) (FVSA, 2003). Both the economic and environmental chambers made progress throughout 2003; however, the social chamber made almost no progress at all. As a result, the National Working Group requested the National Initiative Coordinator to compile, edit, and complete the standard draft. This meant compiling all of the work done by the chambers, and completing the issues not covered by any of the chambers (FVSA, 2003). In 2005, the National Initiative of FSC began public consultation on the first draft of the standards for sustainable forest management of forest plantations (Cerfoar, 2013a). However, this standard was never approved. After the consultation, negotiations over starting a pilot test began, but this also was not accomplished (Interview 1). After the failure of this initiative, FSC ceased to have a national focal point in Argentina, and FVSA ceased to be a member of FSC (FSC, 2014). However, in 2010, a new civil society with members from forestry associations and environmental NGOs was formed with the aim of drafting national standards according to the new set of principles and criteria of the FSC (Interviews 2 and 3), and it has been the national focal point in Argentina since the end of 2014.

PEFC as a subsequent attempt to replace the ambitious FSC scheme
In the year 2002, PEFC, originally conceived as a European system, started admitting non-European members (Cerfoar, 2013a). At this time, the national organism responsible for writing all norms in Argentina, along with the ISO national body, IRAM, began developing the technical norms of voluntary application meant to be the basis of the national system of forest certification (Cerfoar, 2013a).
In 2004, in union with Brazil, Chile, and Uruguay, IRAM signed a regional technical co-operation treaty, with funding from the World Bank, called "Market access and integration through technical standardization" (Cerfoar, 2013a). At the same time, the Secretariat of Agriculture began negotiations with the European Union to obtain funding for the creation of a National Program for the Certification of Forest Plantations. This project established the National Institute of Agricultural Technology (INTA), an autarchic organism dependent on the Secretariat of Agriculture, in charge of writing the standards of sustainable forest management for forest plantations (Escobar, 2006, Interview 3 and 5). In the beginning, both bureaucracies worked together in the writing for this project; however, after the first meetings, the Secretariat of Agriculture presented the project on its own (Cerfoar, 2013a). This project was approved by the European Union at the end of 2004. Both bureaucracies competed for responsibility in policy domain, aiming to be in charge of the development of national standards. By seeking coalitions with international organizations like the World Bank and the European Union, both bureaucracies increased their power, by means of funding, in the role on standard development. Moreover, IRAM had support from the private sector through the main three forest associations (Argentine Forest Association-AFoA, the Argentina Federation of Wood and Allied Industries-FAIMA, and the Cellulose and Paper Manufacturers Association-AFCP), who provided this initiative with office spaces.
At the end of 2005, IRAM began to raise the topic of gaining recognition by PEFC, pointing out the need to include public bureaucracies and the private sector in the process. At that time, both bureaucracies (IRAM and INTA) were developing projects of similar conditions, developing a technical norm of voluntary application. IRAM stated the intention of establishing a dialogue in order to join forces (Cerfoar, 2013a). By the middle of 2006, members of the project run by the Secretariat of Agriculture with funding from the EU started joining the meetings organized by IRAM, having made the decision to work together (Cerfoar, 2013a). At that time, the Direction of Forestry from the Secretariat of Agriculture became an official member in the writing of the standards with voice and vote privileges (Cerfoar, 2013a). However, IRAM is the national organization in charge of writing national norms, so INTA's intention to write the norms themselves was not possible, and after some time co-operation stopped. In 2007, the EU dropped funding to the Secretariat of Agriculture and the project was cancelled (Argentina Forestal, 2007). With the project finished, representatives of the Secretariat of Agriculture resigned their positions in the working group responsible for writing the norms. Until this point, it had been decided that the Secretariat of Agriculture was going to be responsible for starting the process of mutual recognition with PEFC (Cerfoar, 2013a). As a solution to the lack of a bureaucracy taking the lead in getting mutual recognition with PEFC, CERFOAR, a civil non-profit association, was formed in 2009 by the three forestry associations (AFoA, FAIMA, and AFCP) to manage the Argentine Forest Certification System. A year later, during the 14th PEFC General Assembly in Rio de Janeiro, Brazil, CERFOAR was accepted as a PEFC National Governing Body (PEFC, 2013). This situation favored the forest associations, who became founding members of the national governing body with full administrative power and the ability to invite any actor of their preference to become member of this body. Once the association was established, they invited different national bureaucracies and their technical institutes to be part of the development and implementation of the national certification system and take part in the decisions by joining the association administering CERFOAR (Cerfoar, 2013b). During the beginning of 2010, the secretary of CERFOAR met with the Director of Forestry from the now Ministry of Agriculture, Livestock, and Fisheries. During this meeting, they talked about the role the Ministry of Agriculture could play in financing the endorsement of CERFOAR with PEFC and worked on defining the commitments, and it was established that the Ministry of Agriculture's legal advisors would study a model of agreement with the association administering CERFOAR (Cerfoar, 2013a). The Secretariat of Environment never replied to the invitation to be part of the association that administers CERFOAR (Interview 2). Nowadays, the following entities are part of the CERFOAR Association: AFoA, FAIMA, and AFCP (as founding members), the National Industry Secretariat, National Institute of Agricultural Technology, National Ministry of Agriculture, Livestock, and Fisheries, Ministry of Ecology and Natural Resources of the Province of Misiones, and the Ministry of Work, Production and Tourism of the Province of Corrientes (Cerfoar, 2013c, Interview 5). In 2014, CERFOAR acquired the mutual recognition of PEFC.
The process of developing standards for sustainable forest management began in 2003 and ended in 2010 with the approval of the standards following widespread national public consultation and three pilot tests (Cerfoar, 2013b). The technical norms meant to be the basis of the national system of forest certification were developed through a Standard Setting Forum coordinated by IRAM in the presence of technical advisors from different organizations (Cerfoar, 2013a). Throughout the process, 97 different experts participated, 35% representing the forestry associations, 21% the Secretariat of Environment, 17% the Secretariat of Agriculture, and 21% representing other national and provincial bureaucracies. Only 3% represented environmental NGOs (Cerfoar, 2013d).

Forest certification and specific state bureaucracies
Forest certification started as a transnational regime, with the conservation-oriented private sector as the main driver. By means of the national implementation of the schemes, with the active presence of national conservation-oriented actors and little participation from national bureaucracies, these transnational regimes became influential at the national level. This can be seen in the case of FSC in Argentina, where a national environmental NGO housed FSC and acted as coordinator of the development of the standards. However, despite what is often claimed Pattberg, 2005;Auld et al., 2008), state bureaucracies play a decisive and active role in forest certification processes. Forest certification through FSC was first put into the political agenda by the Secretariat of Agriculture through organizing a workshop about the potential implementation of FSC in the country. This is in line with our proposition that bureaucracies can support certification schemes by informational means. During the development of FSC national standards, both, the Secretariat of Environment and Sustainable Development and Secretariat of Agriculture, Livestock, and Fisheries participated as members of the environmental chamber. This shows that national forest bureaucracies supported FSC in Argentina, at least in the beginning, through both informational instruments and the provision of staff. By involvement in the negotiation of standards, national bureaucracies could try to influence the final standards according to their interests. However, national standards for FSC were never achieved, showing the great competition faced by this system. Faggi et al. (2014) found that the main motivation for forest companies to achieve FSC certification was legal.
As a response to the development of a conservation-oriented forest certification scheme of high standards like FSC, productionoriented private sectors created more industry-friendly systems that could compete with, and later even replace conservationoriented systems. By forming coalitions with national state production-oriented bureaucracies, these competing schemes gained strength. In Argentina, the formulation of national standards for PEFC endorsement, an initiative carried out by the forestry associations in coalition with the Ministry of Agriculture and the National Standardization bureaucracy, presented strong competition to FSC. In this case, the Ministry of Agriculture, by becoming allies with the forestry associations through the provision of informational instruments and technical expertise in standard development, and later by incentive instruments, such as paying for the PEFC endorsement procedure, supported the creation of a more industry-friendly scheme that could compete with, and preferably even replace FSC in the country. A similar situation was observed in Indonesia, where the Ministry of Forestry supported forest certification by funding a national initiative (LEI) competing with FSC (Bartley, 2010;Wibowo et al. unpublished). One can argue that private landowners have more power in countries like Argentina and Indonesia with weak state capacities, however the same pattern was observed in countries with strong state capacity like Finland and Germany. In Finland, PEFC, with the support of forest owner organizations, out-competed FSC, with 95% of the forests in the country certified under PEFC only one year after the approval of the standard (Keskitalo et al., 2009). In Germany, PEFC was created by the forest sector in coalition with other landowners and managers across Europe as a response to the ENGO driven FSC. In this case PEFC had a strong support from almost all private and state landowners . As suggested by Öberg et al. (2011) corporatism does not depend on whether institutions are weak or strong but rather whether they are willing or not to engage in mutual exchanges.

Coalitions with preferably weak bureaucracies
Competing national bureaucracies seek coalitions with domestic as well as international actors (the EU, WB, FSC international, PEFC international) to enhance their power in standard development of forest certification schemes, which private conservation and production-oriented actors try to obstruct. In Argentina, environmental and social rights NGOs are mainly local in scope. The bigger national environmental NGOs like Greenpeace are more focused on urban problems, and although they are supporters of FSC, arguing that it is the only means of assuring environmental and social standards , they do not consider forest management a priority (Espach, 2005). This situation makes coalitions between national NGOs difficult. Coalitions with international environmental NGOs like WWF allowed the development of the national initiative; however, these coalitions were not strong enough to overcome local resistance from the strong private sector, especially at the regional level.
By forming coalitions with national state production-oriented bureaucracies, competing schemes became strong. In Argentina, the formulation of national standards for PEFC endorsement, an initiative carried out by the forestry associations in coalition with the Ministry of Agriculture that emerged as strong competition for FSC, almost replaced it. In that case, the Ministry of Agriculture, by becoming allies with the forestry associations through the provision of informational instruments and technical expertise in standard development and later by incentive instruments, such as paying for the PEFC endorsement procedure, supported the creation of a more industry-friendly scheme that could compete with, and preferably even replace FSC in the country.
By seeking coalitions with international actors, the Ministry of Agriculture (formerly the Secretariat) and the National Standardization bureaucracy tried to enhance their power in the competition for policy domain on standard development of forest certification schemes. However, the forestry associations supported the National Standardization bureaucracy, which would not be permanently in charge, giving administrative power back to the landowners' sector. In the competition between these two bureaucracies, the Ministry of Agriculture could not achieve the policy domain of writing the national standards, and support from the EU was dropped. This supports our hypothesis that in situations where new promising venues for political influence open, private actors would prefer ad hoc coalitions with weaker bureaucracies that would not be permanent and with less technical knowledge.

Forest certification and federalism
International and transnational regimes can influence the national level, changing the power balance of the domestic networks. Domestic actors can use these regimes to shift the power balance in their favor. In federal countries, the political system provides multiple institutional access points for policy change. However, how important federalism is depends on how responsibilities regarding certain issues are vertically distributed between the national and the subnational governments (Lopreite, 2014). In cases where the issue at stake is within provincial competencies, the national level will try to use the international sphere to increase their responsibility . This is the case with forests in Argentina, whose responsibility toils within the provincial level . In the case of developing forest certification schemes national bureaucracies, like the Ministry of Agriculture, tried to increase their power over the provinces by leading the national initiative of an industryfriendly scheme like PEFC with the support of private landowners. In this case, despite the provinces retaining the power over forest issues, national bureaucracies by means of transnational regimes increased their power. The provincial bureaucracies, on the other hand, try to retain their power in forest issues but at the same time support their stable coalitions with private actors in their territories. In the case of forest certification in Argentina this was seen both in the development of FSC and PEFC. In the case of FSC, they tried to obstruct the development of regional standards, especially in the regions were forest landowners are stronger, like the Mesopotamia region. At the same time they joined the development of the PEFC standard, being part of the CERFOAR association. In cases where the jurisdiction over an issue lies at the national level, the subnational level can still use agenda setting and bills initiation to influence policy change (Lopreite, 2014). pathway refers to influence by means of direct funding, education, training, assistance and capacity building Cashore 2000, 2012). This pathway is linked to providing incentives to domestic policy actors, which, according to the theory of incentives, can cause an actor´s interest to change direction, and therefore functions as a critical resource of power Ongolo 2012, Krott et al. 2014). Although it represents a very important form of influence, this path can be described as an "infiltration" in the domestic policy-making process . It is seen as intrusion and raises issues of sovereignty . As a consequence, direct access is most likely to be achieved with the support of specific domestic state actors who bring critical legitimacy to the process, which is essential in overcoming sovereignty concerns . In exchange, those domestic state actors that invite outside actors gain additional sources of power, such as legitimacy and financial resources . These domestic state actors can be defined as domestic bureaucracies.
A bureaucracy can be defined as "a public institution that makes decisions concerning specific problems on the basis of general legal standards, resolving those problems by implementing special measures" (Krott 2005, 126). Every bureaucracy has both formal and informal goals. Formal goals are related to serving the public interest while the informal ones are related to surviving, as well as maximizing power, budget and staff (Krott 2005, Wibowo and. The bureaucratic politics theory claims that bureaucracies compete with each other for resources, staff and responsibility for policy domain. This competition between bureaucracies with different preferences, abilities and power capabilities shapes the resulting policy to a great degree (Krott 2005. Bureaucracies have different sources of power that they use to produce policy change and increase their influence in the policy-making process. These power sources comprise three elements: coercion, incentives and nonverified information . Coercion is defined as altering the behavior of an actor by force . The main sources of bureaucracies' power based on coercion are their legal mandates (including decision rights and sanction mechanisms), technical resources (enabling bureaucracies to maintain a policy process) and alliances between bureaucracies and external actors putting pressure behind the preferences of a bureaucracy. (Dis)incentives are defined as advantages and disadvantages introduced to alter the behavior of an actor . The main sources of power based on incentives are transfer funds, creating a link to other issues and funds for setting incentives for others to follow their preferences and alliances between bureaucracies and external actors putting resources behind the preferences of a bureaucracy. Dominant information is defined as unverifiable information that makes the decision process more difficult for actors . The main sources of power based on selected information are the bureaucracies' expertise and active selection of information as well as interest-guided use of beliefs and values. Bureaucracies act in coalition with other public and private actors with compatible interests making up a policy sector (e.g. Forestry, Agriculture; Rayner et al. 2001, Giessen and). In joining forces, policy sectors attempt to influence public policy based on the overlaps of their actors´ interests.
In order to influence domestic policies, international actors are likely to seek coalition with national bureaucracies. These coalitions can give international bureaucracies the opportunity of having direct access to domestic policy change. At the same time, as coalition partners, international bureaucracies can bring an additional source of power to national bureaucracies in their permanent competition with other bureaucracies for staff, resources and responsibilities .
Considering this theoretical framework our hypotheses are: International organizations, like the World Bank, use their different power elements in order to produce policy change at the national level. By means of incentives power, through loans and grants in the form of community-based natural resources management and poverty alleviation projects, they support market-based conservation, in line with neoliberal conservation governance, pushing institutional structures towards deregulation. By enforcing market models and private ownership, developing policy creates illegality, favoring private sector investment.
By using dominant information power as capacity building during projects´ proposal, design and management, international organizations silence alternative policies by selecting national knowledge and expertise.
The influence of international organizations can operate through informal coercive power to support existing domestic bureaucracies or to help create new ones. In such cases, international actors and institutions are supported by specific bureaucracies at the cost of their domestic competitors with a resulting weakened national administration. Domestic bureaucracies will call upon international bureaucracies as allies in cases useful for them and their sector's societal clientele.
In states with weak capacities, where governments are only partially in control of legitimate means of force and are not in the position to implement political decisions, the lack of autonomy of these bureaucracies limits the influence international actors have at the national level.

Case description
The Republic of Armenia, located in the southern Caucasus, has a total area of 29,800 km 2 within 11 provinces. In 1828, the country was incorporated into Russia and later, in 1920, into the United Soviet Socialist Republics (USSR). Armenia achieved its independence when the USSR was disintegrated in 1991, in the middle of a war against Azerbaijan over Nagorno-Karabakh that lasted between 1988 and 1994. As a result of this war, and together with a devastating earthquake in 1988, the country ended up in severe economic and environmental damage, leading to political violence together with transport, economic and energy blockades . This situation caused great pressure on the forests and other natural resources resulting in soil erosion, high deforestation rates and biodiversity loss . Due to its location as well as its topographic complexity, forests in Armenia have a high percentage of plant and animal endemism, and are included in the Caucasus and Irano-Anatolian biodiversity hotspots . By the year 1992, the total forest cover of Armenia was almost 460,000 ha, with about 11.2% of the country being wooded vegetation   (FAO 1995, Sayadyan and. At the same time, a new national forest policy aiming to make economic use of the forest was proposed, but this was never approved . In 1998, the Forest Research Experimental Center (FREC) was created, also under the Ministry of Nature Protection, with the aim of forest monitoring and inventorying. However, coordination between these two forest-related bureaucracies was never achieved . This lack of coordination and capacity of the forest bureaucracies, together with a high demand of forest products caused by the lack of imports from Russia, led to an increase in illegal logging of the forests (FAO 1995, Sayadyan and. At this time, the economy began to recover, with stabilization and structural reforms driving a liberalization of the economy, small and medium privatizations and the formulation of a legal administrative framework for market economy. The Ministry of Nature Protection then invited the World Bank to assist them in the establishment of a new forest policy under the principles of sustainable forest management. This invitation led to the formulation of the "Armenian natural resources management and poverty alleviation" project with funding from both the World Bank and the Global Environment Facility (GEF) (World Bank 2002).

Empirical methods
To understand how the World Bank influences policy change through direct access, we conducted a qualitative case study of the World Bank-funded forest project and the resulting forest policy-making in Armenia from 2002 to the present. To do so, we employed process tracing . The method of process-tracing "attempts to trace the links between possible causes and observed outcomes" (George and Bennett 2005, 6), focusing on sequential processes within a particular historical case. By providing theory-based historical explanations for a case, in which each significant step towards the outcome is explained by making reference to a theory, process tracing becomes a powerful method of inference (George and Bennett 2005, 30). We collected data from different sources with the aim of determining whether the direct access by international bureaucracies, together with competing national bureaucracies and national non-governmental organizations, can be related to the restructuring of the national forestry administration and forest policy change. As a first step, we completed a document search of publicly available sources to gather World Bank proposal documents and project completion reports from 2002 to the present. Using these documents, we identified executive bureaucracies involved in the project. We then reviewed the websites of these bureaucracies to obtain published outputs of the projects. We subsequently conducted semi-structured interviews with experts about the role and influence of international actors and domestic state and non-state actors in the process. Interviews lasted between half an hour and two hours. If it was not possible to meet with the interviewees in person, we used phone calls or emails. In all cases, interviewees were promised confidentiality in order to ensure full disclosure . We used all the data collected for triangulation between documentary sources and all the interviews, providing valid and reliable data . If there were contradictions or differences in interpretation, priority was given to written records, as suggested by .
We used a qualitative content analysis method to analyze all documents and interviews . We coded the information obtained from the different data sources and grouped it into three main categories, based on the research questions. These categories were: i) administrative structure; ii) policy content and iii) power elements.

Direct access, incentives power and market-based conservation creating illegality
After independence, in 1991, Armenia´s economy fell into a severe recession. In 1993, the GDP was almost half of the 1990 level. This economic crisis led to the impoverishment of many rural communities, with as much as 55% of the country´s population classified as poor (World Bank 2002). In 1994, after a period of hyperinflation, fiscal crisis and severe economic contraction, a new stabilization and structural reform program was carried out, together with trade and price liberalization, establishment of a legal framework for a market economy, privatization and reduction in government intervention, leading to a real GDP growth in 1998 of 7.2 percent per annum, between the highest values among the Commonwealth of Independent States (CIS) countries (World Bank 2009). These market-friendly economic reforms received important support from the World Bank, who as part of a strategy to reduce poverty and illegal logging, in coalition with the Ministry of Nature Protection, developed the "Natural Resources Management and Poverty Reduction Project". It took three years of preparation from identification to approval, during which time complex objectives and numerous activities were proposed (World Bank 2009). The final project, which started in 2002, consisted of four components: i) community-based watershed management; ii) state forest management; iii) protected areas management and biodiversity conservation; and iv) project management and administration. The project was to be implemented by the Ministry of Nature Protection with a total funding of USD 16 million of which USD 8.3 million was credit from the World Bank through the International Development Association (IDA), USD 5.1 million was a Global Environment Facility (GEF) grant, USD 1.1 million was from a Swedish Government grant through the Swedish International Development Cooperation Agency (SIDA) and USD 1.5 million was a contribution from the Armenian Government (World Bank 2009). The main objective of the project was to reduce rural poverty by means of improving natural resources management while at the same time protecting important natural resources (World Bank 2009). At this point, forest degradation was identified as one of the critical areas of natural resource government concern, with high deforestation rates, ineffective forest policy, legal and institutional frameworks, underfinanced forest institutions and limited control of illegal logging (World Bank 2009). Many attempts have been made to stop or regulate illegal logging since independence; however, the lack of capacity of the forest administration led these attempts to fail. In 1995, an initiative to establish a national forest program, in coalition with the United Nations Food and Agriculture Organization (FAO), began, however only some statements of forest policy were achieved (Forest Policy andStrategy 2004, Sayadyan 2005). As a solution to these problems the project aimed to review and improve forest-related legislation, reduce illegal logging, undertake forest certification and reform the forest administration (World Bank 2009). These objectives were mainly to be implemented with funding from SIDA through the Forest Institution Support Project.
Besides the aim to reduce illegal logging and increase the quality, quantity and productivity of the forest cover, the project aimed to increase incomes of the participating communities, increase crop and livestock productivity, increase community participation in natural resources management decisions and reverse the degradation in pasture vegetation cover (World Bank 2009).
Among the activities incentivized by the project, under the component of Community-Based Watershed Management, each participating village received USD 15,000 for planting a mixture of indigenous tree and fruit species, fencing, and demarcation for protection and natural regeneration of degraded land adding up to 1,500 ha (World Bank 2009). However, after the project finished, the lack of resources to maintain these plantations threatened their survival. In addition, each village received USD 35,000 to perform livelihood improvement activities, including the construction of livestock watering points, restoration of field tracks, fertilization of community pastures, improvement and rehabilitation of village hay meadows, reintroduction of forage legumes into crop rotations, improved wheat and spring barley technology demonstrations, and bee keeping for honey (World Bank 2009).
With the aim of developing community forest management, three consecutive activities were proposed by the project. As a first step, community forestry management plans for former kolkhoz (collectivelyowned) and sovkhoz (state-owned) forest areas were supposed to be developed. Once these plans were approved, forest area management rights and responsibilities were supposed to be legally transferred to local communities or village-level organizations with a later investment in rehabilitation, reforestation, enrichment planting, or other improvements (World Bank 2009). However, the implementation of these activities was not successful. Only 7 out of the 12 planned forestry management plans were prepared and not even a single one of them was approved by the end of the project. As well, no legal transfer of the management rights was done. A similar situation was observed with the forest management plans, with two of the plans getting approval and three still pending. However, even though these plans were made, the implementation phase was not achieved. Both the community forestry management plans and the forest management plans were developed in order to comply with the World Bank´s requirements, but no actual attempt to implement them was made.
In protected areas, management plans were supposed to be developed at the beginning of the Project, however, management plans for the Dilijan and Lake Sevan National Parks were only developed with modern international standards in 2007 when the plans were approved, leaving no time for implementation within the Project´s duration. Moreover, the multi-sectoral consultation nature of the process led to major compromises on conservation aims (World Bank 2009). As a consequence, the quality and effectiveness of these plans, especially from the Lake Sevan National Park, were questionable. The level of the Lake Sevan had been lowered considerably over the previous decades due to water extraction with irrigation and hydropower purposes. In 2001, a program was prepared by the Ministry of Nature Protection to raise the level of the lake by 6 meters over a period of 30 years, with the aim of improving ecological conditions. During the development of the management plan, with funding from the World Bank, half-built and abandoned buildings from the Soviet era that would be flooded by the water level rise were identified and planned to be destroyed. However, before the plan was accepted, the lakeshore property was leased on a long-term period to private investors who were going to expand and modernize these buildings. The total number of buildings that would have been flooded was 150, stopping the flooding plans.
As part of the "Natural Resources Management and Poverty Reduction Project" and with additional funding from the Swedish government through a SIDA grant, an "Illegal Logging Action Plan" was adopted in 2004 (Sayadyan 2005). According to official reports (decision N38, Government of Armenia 2004), unauthorized timber extraction for subsistence carried out by local communities was the main source of illegal logging in the country. The political and economic transition that took place after independence, together with the war against Azerbaijan between the years 1988 -1994 and the disruption of goods and energy supplies from the Soviet Republic, led to a big energy crisis that could have led to an increase in illegal logging for subsistence reasons. However, the disruption of imports from Russia also included timber imports, which until 1991 comprised 95% of the wood processing industry's raw material. This disruption was not reflected in the products of the wood industry, which from 1991 onwards had to obtain its raw material from illegal logging within Armenia, becoming probably the main driver of illegal logging in the country (Sayadyan 2005). In 1998, the Armenian forest administration (Hayantar) agreed to sell 2000 m 3 of high quality oak from Armenian forests during a five year period to the Yerevan Brandy Company (Sayadyan 2005). However, according to the Forest Code (1994), the extraction of healthy oak trees is prohibited and this agreement could only be achieved by illegal logging.
Official market volumes for timber processing in 2005 represented USD 10 million, while the real estimates calculated a total of USD 132 million, more than ten times the official values. On the other hand, official market volumes of fuelwood in the same year represented USD 2.5 million while the real estimates were around USD 7.5 million, three times the official values. In the total real market value of wood, fuelwood represented less than 6% of the total. In terms of wood volume in the same year, official market volumes for timber processing were 6,600 million m 3 while real volumes were estimated around 339,000 million m 3 , 50 times the official values, representing 54% of the total market volume. On the other hand, fuelwood official market volumes were 63,000 m 3 while the real estimates were around 291,000 m 3 , almost five times the official values and representing 46% of the total market volume (Hergnyan et al. 2007). This data shows that although World Bank reports claimed fuelwood to be the main driver of deforestation through illegal logging, timber processing industries account for 60% of the total volume of illegally harvested wood (Hergnyan et al. 2007). While fuelwood consumption is driven mainly by poverty, representing 9% of the total households´ means of cooking and heating in 2006, timber harvesting is driven by profit generation and affected by supply and demand factors (Hergnyan et al. 2007). The wood processing industry in Armenia is comprised of more than 300 small and mediumsized private actors with their main source of wood from illegally harvested sources. While the demand for fuelwood consumption is expected to decrease in the mid-term perspective, the commercial wood industry will continue to grow.
As a result of the "Illegal Logging Action Plan", permits from the Forest Enterprises to obtain fallen wood for fuelwood had been reduced and restricted, citing conservation aims. Hence, few households are able to obtain their wood directly from the forest and are forced to get it from intermediaries. This situation favors certain groups or individuals who are able to create rent opportunities. By the year 2010, more than 64% of households bought fuelwood from an intermediary, a significant increase in comparison to 52% in 2003 (Junge and Fripp 2011). Even though local communities are not the main drivers of illegal logging, they are used as cheap labor force by private wood processing industries.

Dominant information as capacity building leading to policy change
The The project also incentivized the reform of legislation and regulations dealing with biodiversity conservation in protected areas. A new law on Specially Protected Natural Areas in Armenia was adopted in the year 2000, however, its implementation called for new bylaws. Although new regulations on monitoring and land registry were approved as a result, the Ministry of Justice did not approve regulations on land use. This new legislation would allow protected area administrations to generate and retain income (World Bank 2009).
Despite the large volumes of illegally harvested wood, there is no official definition of illegal logging in Armenia (Hergnyan et al. 2007), allowing the World Bank to use their own definition for the purpose of the project. By bringing in external experts as advisors as well as putting them in charge of different components of the project using dominant information power, the World Bank limits the range of alternatives, silencing some propositions and selecting choices in line with their own interests.
Also, as means of capacity building, SIDA provided support through the Swedish University of Agricultural Sciences for long-term cooperation on forestry education with the Armenian Agricultural Academy, establishing a Forestry Department in the Armenian Agricultural Academy in 2003. This cooperation provided equipment, literature and other facilities for forestry education as well as training and consultancy services (Sayadyan 2005).

International organizations and national administration structure
During 1992-2003, the forest management body of Armenia -the 'Hayantar' -underwent numerous restructurings, and as a result of each restructuring lost a part of its authority. As part of the forest sector institutional reforms carried out with funding from the project, 'Hayantar' became an organization within the structure of the Ministry of Agriculture, moving out of the Ministry of Nature Protection (Forest Policy and Strategy 2004). In addition to this change, there was an increase in budgetary allocations to this administration, improving its performance and effectiveness. From 2004 until the end of the project, funding of the forest sector increased tenfold (World Bank 2009). Until 2004, 'Hayantar' was responsible for implementing both forest management and control. Following the structural change carried out with support from the World Bank, the responsibility for control was given to the Ministry of Nature Protection (Governmental Resolution of the Republic of Armenia N° 7, 2004). However, there are three organizations with monitoring and control functions, the Forest State Monitoring Centre (FSMC), the newly created State Forest Service (SFS) and the State Environment Inspectorate (SEI), all within the Ministry of Nature Protection. Moreover, 'Hayantar' also controls illegal logging. This overlap between different administrations weakens the control activities (Elmqvist and Rylander 2010). The Ministry of Nature Protection, through the Forest Research Experimental Center (FREC), is responsible for forest inventory and planning as well as research and training. During the Soviet period, no academic forestry education was provided in Armenia, as a result 'Hayantar' employees have a low level of forest education, with only 4% of the staff members having undergone forestry education. This situation makes the formulation, implementation and control of sustainable forest management plans difficult.

Global influences changing domestic policy
During the formulation stage of the project, international experts hired by the World Bank identified illegal logging, which they claimed was caused mainly by the demand for firewood by local actors (Hergnyan et al. 2007, World Bank 2009, as one of the main drivers of deforestation. Using illegal logging as a discourse pathway, they proposed, as part of the main aims of the project, policy changes to move towards production use of the forests and away from pure conservation. This would allow privatization of the forestland as well as a reform of the forest administration, moving the forest administration from the Ministry of Nature Protection to the Ministry of Agriculture. This is in line with previous studies that suggested that discourse, previously described as a pathway of international influences , Leipold and Winkel 2016, is rather a prestage that prepares the way for the direct access pathway (Hogl et al. 2009, Burns and by means of incentives and dominant information power. The total volume of illegal logging due to fuelwood consumption reported in the World Bank reports differs from the values reported by local experts, exceeding their estimates by almost 55,000 households (Sayadyan 2005, Hergnyan et al. 2007. They assumed that local users were the main drivers of deforestation, and proposed policy changes attempted to improve the situation by promoting a change in the Forest Code, the formulation of a National Forest Policy and Strategy and a National Forest Program together with an Illegal Logging Action Plan (World Bank 2009). However, these changes only restricted local users from getting fallen wood from the forest for fuelwood and forcing them to obtain it from intermediaries.

Global and domestic actors´ power
The role of external and local actors differs along the cycle of the projects. While in the formulation phase, external actors are more influential, it is the local actors in the implementation phase who have more power Hjortsø 2004, Singh 2009). During the preparation, design and management phase, international actors by means of dominant information limit the range of ideas by silencing alternative policies and selecting certain local actors and experts . This can be observed in the case of the 'Armenian natural resources management and poverty alleviation' project, where from a total of 17 experts participating in the formulation and implementation of the project, ten (almost 60%) of the experts came from foreign countries and only seven from Armenia, representing equally the Ministry of Nature Protection, the Ministry of Agriculture and environmental NGOs. A similar case was observed in the forest sector in Lithuania where international organizations, with a political agenda and incentive power derived from funding resources, left limited access to local actors in the identification and preparation stages of the projects, keeping all decision-making authority .
As a result of the World Bank project in Armenia, by means of direct access with incentive power, the forest administration structure and legislation were changed. As part of these changes, the forest administration, which used to be under the Ministry of Nature Protection, moved to the Ministry of Agriculture. Alongside this, the forestry administration body increased its budget; however, it also implied a reduction in policy domain, because control was lost over responsibilities that were given to the Ministry of Nature Protection, who also kept forest inventory and planning responsibilities through FREC. Through these changes, the forest policy domain became fragmented and the forest bureaucracies became less powerful, although their budgets were increased, at least during the duration of the project. This situation is similar to those observed in other countries by previous studies (Gautier et al. 2013, Burns and. In Argentina, funding from the World Bank led to the dismantling of an autonomous forest bureaucracy into four different units within different ministries . In the case of Mali, also as a result of the World Bank intervention, the National Forest Administration lost its power, losing responsibilities and budget. In this case, the forest bureaucracy was not dismantled like in Argentina but weakened by the World Bank promoting the devolution of forest management to the local level and to a newly created bureaucracy funded by the World Bank, responsible for creating efficient markets . In all these cases, the World Bank compensated bureaucracies for their loss in power by means of incentive power.

Imbalance of deregulation, new government organization and new government means
The reform initiated by the World Bank deregulated the forest sector by creating private forests. Simultaneously, the forest administration was restructured and new planning instruments were introduced with the formal aim to strengthen the ecological goals in the forest. In practice, the opposite happened: The deregulation increased harvesting and export quickly due to strong actors and the new plans came never into being because of weak state actors and governance processes in the implementation. The exploitative effects of forest reforms are not unique to Armenia. It seems that the mixture of effective deregulation and ineffective new governance provides a formal concept for legitimization of reforms. Informally, such an approach is doomed to fail -it is detrimental for sustainable forestry but benefits the informal winners who gain additional profits from the forest in these countries.

Weak states and the influence of international actors
Although several laws were formulated and later approved as a result of the 'Armenian natural resources management and poverty alleviation' project, the implementation was never achieved. Although the new Forest Code established the concept of private forests, even now there are no private forest owners in Armenia. The same was observed with both community forestry management plans and forest management plans, developed during the project in order to comply with the World Bank´s requirements; however, until now no actual attempt to implement them has been made. This can be explained by the state´s lack of power as pointed out by . A similar situation was observed in Lithuania, where the prevailing opinion is that the assistance projects in forestry have not had a significant influence on the development of the sector . While international projects have produced many legislative proposals, only a few have been enacted . This is a very different situation from what was observed in Argentina, where projects funded by the World Bank led to the development of a new forest law, which was immediately implemented .

Conclusion
In Armenia, the World Bank, as an international organization and by means of funding and capacity building, assisted the coalition of agricultural and profit-oriented bureaucracies and landowner associations (of forestry and agriculture) that emerged after the independence of the country at the beginning of the 1990s. With a trend towards neoliberal conservation governance, by forming coalitions with the World Bank, certain bureaucratic units were able to increase their staff and budget and produce valuable information, thereby increasing their power. This agrees with our theory, in that international organizations can operate through the provision of funding and capacity building to assist the existing domestic coalitions of bureaucracies and associations in their constant competition for power, or to help create new ones. Because bureaucracies like the Ministry of Nature Protection and the Ministry of Agriculture were in permanent competition, they invited the World Bank with the aim of increasing their domestic power at the cost of their domestic competitors. This invitation led to the establishment of influence through direct access of the World Bank, as an international organization, on domestic natural resource and environment policy and administration under the umbrella of poverty reduction projects. The reform influenced by the World Bank introduced new planning instruments with the formal aim to strengthen the ecological goals in the natural resources management. In practice, the opposite happened: The deregulation increased harvesting and export quickly due to strong actors and the new plans came never into being because of weak state actors and governance processes in the implementation. The mixture of effective deregulation and ineffective new governance provides a formal concept for legitimization of reforms. Informally, such an approach is doomed to fail -it is detrimental for sustainable natural resources management but benefits the informal winners who gain additional profits in these countries while restricting access to poor local users.
In recent decades, globalization and internationalization led to an increase in the number of international regimes attempting to infl uence national behaviour over many different issues. By using the case of the international forest regime complex this thesis seeks to evaluate how an international organization such as the World Bank and private institutions of forest certifi cation infl uence domestic forest policy. The cases of Argentina and Armenia were selected as examples of developing countries open to international infl uences with weak forest sectors that went through a recent administrative restructuring. In so doing this dissertation seeks to answer how do international and transnational organizations infl uence domestic forest policies?
The results show that: the infl uence of the World Bank in Argentina and Armenia pushed the forest sector towards deregulation; state bureaucracies play an important role in the implementation of transnational regimes at the national level; the political system of federal countries provides multiple institutional access points for policy change that international and transnational regimes try to use in order to infl uence the domestic level, consequently changing the power balance of the domestic networks.