Gradual and Transformative Institutional Change: the influence of stakeholders and advocacy coalitions at the Belo Monte Hydroelectric power plant

This work analyzes the performance of stakeholders and advocacy coalitions in processes of gradual institutional change based on the case study of the Belo Monte hydroelectric power plant. The methodology adopts content analysis, using shorthand notes of hearings held in the National Congress, and other publicly available documents. In-depth semi-structured interviews were also conducted, collecting data from individuals involved in the history of the power plant. The study allowed to characterize stakeholders pro or against the enterprise, observing their degree of influence and the role played. In addition, the work used the Advocacy Coalition Framework to verify how stakeholders formed three advocacy coalitions and exerted the strategies of institutional change ‘displacement,’ ‘conversion,’ and ‘layering’ following the gradual and transformative institutional change model. Empirical evidence indicates that the composition of coalition members may determine the type of institutional change and that institutional changes can also result from learning among coalitions, as the incorporation of socio-environmental considerations that have altered the original design of the hydroelectric power plant.


Introduction
A prolific approach in public policy studies seeks to understand the influence of stakeholders in policies' agenda-setting, planning, and implementation.The research based on the New Institutionalism is relevant because it introduces the concept of 'institution' as an element that guides the formation of strategies to address problems while regulating the behavior of the actors involved.Thus, the institutions defined by Mahoney and Thelen (2010) as the set of rules, norms, laws, regulations, and procedures that define the policy, are capable of providing stability, order, and cooperation between the actors.Public policies would then be formulated from (institutional) arrangements with varying degrees of interdependence and commitment among countless types of actors (Almeida and Gomes, 2019).
On the other hand, it is possible to see the occurrence of processes of institutional change caused by actors dissatisfied with the policy.Jenkins-Smith and Sabatier (1993) argue that public policy reflects the hegemonic coalition's beliefs and values.Therefore, whether seeking greater access to resources or a format that brings the policies closer to their beliefs and values, dissatisfied agents engage in processes of gradual and transformative change (Thelen, 2009;Mahoney;Thelen, 2010).According to Gomes, Liddle, and Gomes (2010), the policy's stakeholders are guided by their values, preferences, and expectations, organized in advocacy coalitions, and acting politically in an institutional context marked by disputes and ambiguities (Viera; Gomes, 2014).As Bispo andGomes (2018, p. 1264) point out, for each strategic decision, organizations generally face a diverse group of stakeholders with varied and often conflicting interests and objectives.Vieira and Gomes (2014) carried out a literature review on the influence of stakeholders in public policies and on the processes of institutional change.They proposed an integrative model based on the application of stakeholder analysis models with the Advocacy Coalition Framework (ACF) to describe and explain how the processes of gradual and transformative institutional change take place, adding new elements to the central approach of Thelen (2009) and Mahoney and Thelen (2010).Stakeholder analysis models enable the identification and characterization of the actors influenced and capable of influencing the policy.The ACF, on the other hand, provides an analysis of how such stakeholders are grouped into coalitions based on shared beliefs, improving stakeholder' mapping, in addition to presenting proposals on the dynamics of interaction and learning of coalitions and political changes.Finally, the gradual and transformative institutional change model (GTIC) characterizes the institutional change strategies and the agents' behavior.That said, this paper seeks to answer the following question: how did the process of gradual and transformative institutional change take place based on the action of stakeholders and advocacy coalitions in the planning and implementation of the Belo Monte hydroelectric power plant?Stake (2000) argues that an instrumental case study can promote insights by facilitating the understanding of a phenomenon and refining the theory.In over 43 years of history, since the Hydroelectric Inventory Studies of the Xingu River Basin in 1975, several actors disputed the viability of a power plant in that region.The government launched the bidding for the construction and operation of the plant only in 2010, which was won by Consórcio Norte Energia SA.Thus, the case of Belo Monte seems to support the theoretical premises that underlie the integrative model proposed by Vieira and Gomes (2014).In this sense, the main justification for this study is the empirical application -through a case study -of the propositions of Vieira and Gomes (2014) on processes of gradual and transformative institutional change in public policies.As will be discussed below, it was possible to observe the relevance of the integrative theoretical model, as well as its specific contributions to the theoretical reference models.
The paper is structured as follows: the next section presents the stakeholder analysis models and the elements of the ACF employed in this study.Section 3 describes the GTIC, followed by section 4, which discusses the research methods.Sections 5, 6, and 7 present the results discussed according to the approach proposed by Vieira and Gomes (2014).The final considerations are elaborated in the last section.

Stakeholder analysis models and the Advocacy Coalition Framework
In the integrative model proposed by Vieira and Gomes (2014), stakeholder analysis characterizes the actors, observing their willingness to cooperate or compete, a degree of influence they can exercise and roles they play (Savage et al., 1991;Mitchell, Agle, and Wood, 1997;Gomes, Liddle and Gomes, 2010).Savage et al. (1991) indicate a matrix that combines the stakeholders' threat and cooperation likelihood.The matrix is classified into four types of stakeholders: supportive, nonsupportive, mixed blessing (with ambiguous behavior), and, finally, marginal stakeholders (without the ability to exert influence).Mitchell et al. (1997) elaborated a model of analysis of stakeholders' power and influence, based on three attributes: the type of power exercised (coercive, normative or utilitarian), the legitimacy of their relationship with the organization (their interests need to be recognized as legitimate) and the urgency of their demands (the degree to which the demand needs immediate attention).Regarding the first attribute, Hardy (1996) proposes that power is also based on the possession of resources, access to decision making, and control of the status quo through the manipulation of meanings with the public.Thus, Mitchell et al. (1997) categorize the stakeholders according to the possession of the attributes in (1) dormant, (2) discretionary, (3) demanding, (4) dominant, (5) dependent, (6) dangerous, and, finally, the (7) definitive stakeholders, which are those possessing all three attributes, and are the main target of any organization's attention.
In a study observing local governments, Gomes et al. (2010) proposed a model to analyze stakeholder influence.Based on the way they behave and influence, the authors suggest the existence of five types of stakeholders: 'regulators,' 'collaborators,' 'agenda setters,' 'controllers,' and 'legitimizers.' 'Regulators' influence the budget.'Collaborators' are external or internal agents that contribute to the implementation of public policies and services (as in the case of public-private partnerships).The 'agenda setters' belong to the federal public administration and define the central objectives that will guide the actions of local governments.'Controllers' employ formal or informal control mechanisms over local government practices.Finally, 'legitimizers' are citizens who actively influence the decisions made by local governments.
According to Weible (2006), stakeholder analysis must include the following elements: what are the stakeholders' values/beliefs?Who controls critical resources?Who do they form coalitions with?What institutional strategies and arenas do they use to achieve their goals?The answer to these questions allows exploring the perceptions of these agents regarding the severity, causes, and purposes of a given problem, the distribution of resources among coalitions, and the institutional arenas available to influence policy.On the other hand, the Advocacy Coalition Framework (ACF) can contribute to stakeholder analysis because it considers a political subsystem as a unit of analysis.It stipulates the substantive and territorial limits of a political issue and who must be part of the analysis, the structure of individual beliefs and motivations, identifying the stakeholder resources, and available institutional arenas.Weible, Sabatier, and McQueen (2009, p.122) state the premises of the ACF, based on Sabatier and Jenkins-Smith (1999): "(i) a central role of scientific and technical information in policy processes; (ii) a time perspective of 10 years or more to understand policy change; (iii) policy subsystems as the primary unit of analysis; (iv) a broad set of subsystem actors that not only include more than the traditional iron triangles' members but also officials from all levels of government, consultants, scientists, and members of the media; and (v) a perspective that policies and programs are best thought of as translations of beliefs".
It is possible, therefore, to establish a hierarchical system of beliefs of the actors.At the top are the deep core beliefs, that is, those more comprehensive, stable, and predominantly normative.At the intermediate level are the policy core beliefs, which are moderate in scope and expand the substantive and geographical scope of the political subsystem.Their specificity makes them ideal for forming coalitions and coordinating activities among members.They are resistant to change, but they adjust based on new experiences and information.Finally, at the base of the hierarchical system, are the secondary beliefs, which are substantially and geographically more restricted in scope, and are also empirically grounded.Compared to the others, these beliefs are more likely to change over time (Jenkins-Smith, Nohrstedt, Weible and Sabatier, 2014).As this study will demonstrate later, the disputes surrounding the Belo Monte hydroelectric power plant appear to have been motivated by the different beliefs and values of the stakeholders regarding the pertinence and conception of the enterprise.Stakeholders can cluster into four advocacy coalitions according to two basic criteria: sharing normative and causal beliefs and engaging in a nontrivial degree of coordination of activities over time.Each coalition has a specific set of resources and chooses strategies to achieve its political objectives (Sabatier and Weible, 2007).
There are two ways of bringing about policy change in a political subsystem.The first refers to events external to the subsystem, i.e., socioeconomic changes, public opinion, government coalitions, and other subsystems.These events can alter the distribution of resources, affecting the power of coalitions and changing beliefs.The second way is through policy-oriented learning, defined as "relatively enduring alterations of thought or behavioral intentions that result from experience and/or new information that are concerned with the attainment or revision of policy objectives" (Sabatier and Jenkins-Smith, 1999, p. 123).In addition to political change through external events and learning, it is suggested that change is possible based on agreements between two or more coalitions (Jenkins-Smith et al., 2014).Also, when considering that the institutions are the players responsible for setting the public policies, it is noted that changes will also occur through competitive processes and movements of gradual and transformative institutional change, led by non-hegemonic coalitions.
Regarding the Brazilian academic production using the ACF, the review by Capelari, Araújo, and Calmon (2015) identified only 19 master theses and doctoral dissertations published between 2002 and 2013.The model has been applied more frequently in the field of energy and environmental policies and has replicated results found in international studies.The authors suggest the possibility of coupling two or more models when observing a national policy, complementing the ACF, in addition to exploring a specific gap pointed out by Sabatier and Weible (2007) regarding the analysis of coalition resources, which converges with the approach adopted in this study.

The Gradual and Transformative Institutional Change Model
The Gradual and Transformative Institutional Change Model (GTIC) (Streeck and Thelen, 2005;Thelen, 2009;Mahoney and Thelen, 2010) is based on the following premises: (1) the political system is composed of multiple actors with diverse interests and uneven distribution of power; (2) the actors understand the institutions in different ways; (3) the preferences and interests of each actor can be ambiguous; (4) the agency has unintended consequences; and (5) the actors permeate multiple complex institutional environments.The authors argue that institutions act as instruments to distribute resources with implications of power so that institutional arrangements start to reflect the interests of coalitions.In this perspective, even continuity depends on constant mobilization, political support, and an active effort to resolve institutional ambiguities.The ambiguity of interests and interpretations, as well as the balance of power and obedience to institutions, create spaces for institutional change.According to Mahoney and Thelen (2010), "where we expect incremental change to emerge is precisely in "gaps" or "soft spots" between the rule and its interpretation of the rule and its enforcement" (p.14).In this hiatus, manifestations of coalitions can provide the necessary change so that there is an institutional rearrangement that better accommodates the distribution of resources.Thelen (2009) proposes that changes are determined by the characteristics of the political context (veto power), by the characteristics of the institution itself (interpretation and its application), and by the type of behavior of the dominant agent of change.Therefore, the author mentions four different strategies of institutional change: displacement, layering, drift, and conversion.Displacement refers to change as a consequence of removing existing rules and introducing new ones.In the change by layering, new rules (or amendments and revisions) are added to the old ones, changing the impact of the original rules on the behavior of individuals.Drift refers to change due to the maintenance of rules, even though their impact changes because of contextual shifts.This situation may occur due to negligence or failure to adapt the rule to keep showing the impact intended originally.Finally, in the case of conversion, the rules formally remain the same but are interpreted and applied in other ways, in order to explore ambiguity.Mahoney and Thelen (2010) also establish a classification for the behavior of agents: Insurrectionaries, Symbionts, Subversives, and Opportunists.Insurrectionaries reject the status quo, try to change the institution, and do not always obey the rules.Symbionts are divided into two subtypes: Parasitic, who exploit the institution for their own gains, undermining it in the long run, and Mutualistic who act in the same way, but without compromising the institution.Subversives do not act directly to bring about change, they seem to support it, but they wait for the opportunity when they can be opponents and propose new rules without extinguishing the old ones.Finally, there are the 'opportunists,' who have ambiguous preferences regarding the continuity of the institution and do not actively seek to preserve it.Given the costs, they also do not move to change it.When they act for change, they adopt the conversion strategy.
That said, it is worth emphasizing that the power relations among public policy stakeholders are conditioned by the attributes that characterize each of them (Mitchell et al., 1997), by the role they play in public administration (Gomes et al., 2010), and the way they are clustered in advocacy coalitions.A coalition may have more or less attributes and resources to expand its power, according to how it is formed.This condition will establish whether it will be hegemonic or in search of hegemony.In addition, the institutional structure of a policy will distribute resources and define the status of stakeholders (and, as a result, the status of the coalitions) in disputes over hegemony.

Method
This work seeks to characterize the behavior of stakeholders and advocacy coalitions in the processes of gradual and transformative institutional change related to the Belo Monte hydroelectric power plant.According to Thelen (2009), studies on this type of institutional change use a historical analysis of events with a horizon of at least ten years.This is a premise also observed in the Advocacy Coalition Framework (ACF), according to Jenkins-Stmith et al. (2014).Thus, the analysis of the power plant's history focused on the period 1999 to 2013, which represented the resumption of the project on the agenda, the bidding for the concession, and the beginning of the works.The literature review showed that the empirical approaches of gradual and transformative institutional change models, ACF, and stakeholder analysis models are mainly based on case studies (Gomes et al., 2010;Sabatier and Jenkins-Smith, 1993;Munro, 1993;Mawhinney, 1993;Araújo, 2007;Vianna, 2011;Faletti, 2010).
Data collection took place through documentary research and in-depth semi-structured interviews with individuals linked to the history of the power plant.An extensive search for official documents, such as specific laws, minutes of meetings, assemblies, and public consultations, as well as technical reports and other publicly accessible files, was conducted.After preliminary examination of the documents, a script was drawn up for conducting an exploratory interview with one of the Eletronorte technicians who was involved with Belo Monte between 1999 and 2002.From this interview, a definitive version of the instrument was designed, in addition to a preliminary list of key actors to participate.The interviews were scheduled, and the snowball technique was used to select new interviewees.Table 1 shows the interview questions and the list of the interviewees' occupations and workplaces.During October and November 2013, 16 interviews were conducted in the cities of Brasília (DF), Belém do Pará, and Altamira (PA) with an average duration of 58 minutes.The number of respondents was determined by the saturation principle, which, according to Gaskell (2008), assumes there is a limited number of opinions on a given topic within a specific social environment.The interviews were transcribed and underwent content analysis, as instructed by Bauer (2008).A spreadsheet was created to record the interview excerpts considered most relevant, separated by the respondent and by the categories stipulated a priori, based on the stakeholder analysis models adopted in the theoretical framework.The instrument was created to encourage respondents to naturally mention stakeholders during the interview (questions 1, 2, and 4).It is believed that this strategy is more relevant than asking the interviewee directly who the power plant's stakeholders are.With the analysis of these data, it was possible to identify and classify the stakeholders, considering the categories proposed by the models.The examination of the interview's content (questions 2, 3, and 5) also contributed to elaborate the map of the relatively stable system parameters and the actors' belief structure, according to the premises of the ACF.
After data collection, the study focused on the analysis of the documents obtained.However, these documents were insufficient for the investigation of the case study according to the ACF and the gradual and transformative institutional change model (GTICM).Therefore, it was necessary to return to the field in search of shorthand notes from public hearings held in the Chamber of Deputies and the Federal Senate.The Relationship, Research and Information Coordination of the Chamber of Deputies made links available via email through which the plenary pronouncements and hearings held by the house committees related to the electricity sector or Belo Monte could be accessed, as specified in Table 2.This provided a second search for documents that supported the development of the study.
With the help of specialized software, data were organized, and the categories of analysis were structured.In total, the shorthand notes of 35 hearings that took place between 1999 and 2013 were examined, corresponding to approximately 91 hours.The key actors were identified, representatives of the countless stakeholders previously mapped, and a database was elaborated with the significant passages of their speeches in light of the theoretical framework.
Therefore, the stages of construction of the case study were: (1) study of Belo Monte's history, (2) identification and classification of stakeholders, (3) identification of coalitions, (4) description of the coalitions' behavior, over the determined period, (5) identification of institutions and institutional positioning of coalitions, and (6) analysis of the process of institutional change.In their book Policy Change and Learning, Sabatier and Jenkins-Smith (1993) brought together several case studies with empirical applications of ACF.The methodological decisions made in this research reflect the practices observed in the case studies portrayed in the book, in the authors' suggestions, and the perception when studying the most recent research developed in Brazil.Among them, the pioneering study by Jenkins-Smith and Sabatier (1993) on Lake Tahoe, Munro's (1993) work on water policy and transpositions in California, Araújo's (2007) national study on biodiversity and forests, the work of Vianna (2011) on the transposition of the São Francisco River and the study by Vicente, Calmon, and Araújo (2017) on the urban spatial planning policy in the Federal District.
The in-depth interviews and consulted documents were based on the definition of the beliefs of the advocacy coalitions.Also, the relatively stable system parameters were defined through the historical study of the enterprise, that is, the aspects of the problem and the context shared by the actors limiting their beliefs, resources, and strategies (Jenkins-Smith and Sabatier, 1993).After developing the map of the relatively stable system parameters, the document analysis code was elaborated, in which the beliefs of the actors indicated in Figure V are structured.The deep core beliefs are the same as those identified by Araújo (2007), while the rest were elaborated during this study.Then, in an analysis and interpretation process, fragments of the speeches were categorized according to the belief structure.A triangulation of the data collected from different sources is highlighted: official and unofficial documents, in-depth interviews, and shorthand notes from public hearings.

Analysis of stakeholders of the Belo Monte hydroelectric power plant
Initially, 84 stakeholders linked to Belo Monte were identified.However, after analyzing and interpreting the data based on the three stakeholder analysis models, a list of the most influential actors was obtained -called the 'definitive stakeholders' -, discriminating them according to their attitude towards the enterprise and the role they can play, as shown in Table 3.In this case, there was a clear polarization, and also some stakeholders who do not seem to have a final opinion about the power plant.In addition, some entities had controversial cooperation with the enterprise, contrary to what might be expected, such as FUNAI, MMA, and IBAMA.It is worth mentioning that the model by Mitchell et al. (1997) provides a certain dynamism to the analysis.It considers the stakeholder's status as the result of the perception of the reference organization, and that their condition depends on the interactions with other stakeholders or the influence of external factors, which will indicate the possession of one or more determining attributes.However, the stakeholder salience can also change based on their approach or alliance with others.This is the case with Greenpeace and ISA, classified as dominant stakeholders (with power and legitimacy, but without urgent demand), which, when cooperating with local associations, puts pressure on the enterprise.The next step was to apply the five-sided model of stakeholder influence developed by Gomes et al. (2010).First, some of the categories were simplified to become more flexible to suit the case.Differently from the context of local government administration, the case studied is marked by the installation of an enterprise resulting from a public-private partnership, financed to a large extent by the state-owned Brazilian Development Bank (BNDES), in an area of influence involving eleven different municipalities.As for Collaborators, the study considered the existence of actors with the technical or financial capacity to implement the enterprise and who envision a favorable context, justifying their engagement (large contractors, for example).On the other hand, some agents may act in a different role from what they were originally assigned.For various reasons, an entity with a profile to be a controller, for example, may act as a collaborator, such as observed in the case studied.The study revealed that interviewees and participants in public hearings considered IBAMA and other state and federal environmental licensing bodies as acting in the interests of those advocating the implementation of the power plant.At one public hearing, it was stated that the Ministry of the Environment, under pressure from the Ministry of Mines and Energy, forced more flexible conduct by IBAMA in the Belo Monte power plant's licensing process.
In addition to monitoring the use of the available budget, stakeholders classified as 'controllers' according to the five-sided model ensure that the institutions are respected.An emblematic example refers to the understanding of how public hearings should be conducted with indigenous communities.Due to the lack of regulation of article 231 of the federal constitution, there was a divergence between the Federal Public Ministry (MPF) of the state of Pará and Norte Energia S.A.This issue was taken to the arbitration of the Brazilian Supreme Court.Regarding the stakeholders called as 'legitimizers,' this role should be broadly played by the countless organizations that can represent the wishes of the population, such as religious organizations, associations, and class entities.In the case studied, this role proved to be so relevant as to influence the schedule of feasibility and construction studies for the Belo Monte hydroelectric plant.As for the stakeholders classified as 'regulators' and 'agenda setters,' they kept their original understanding as portrayed in the five-sided model of stakeholders influence.
According to Gomes et al. (2010), all stakeholders seem to have the ability to influence the organization's goals.For Mitchell et al. (1997), on the other hand, some of them -such as the discretionary stakeholders -do not have this capacity.Some of the actors observed in the case study were not easily allocated in these categories.They were considered in another group called 'sentinels,' who are those watching and waiting for the developments in a veiled fashion.They have no legitimacy to exercise influence and do not participate (at least explicitly) as collaborators, even though they have political interests.Examples of Sentinels in the case of the Belo Monte plant are the foreign ore buyers, electricity distribution companies, manufacturers of turbines and other components for hydroelectric and mining plants.On the other hand, what other contributions can the approach combining such stakeholder analysis models bring to the Advocacy Coalition Framework (ACF)?Is it possible to characterize the types of power exercised by each coalition, or even suggest which coalitions would be more influential on a political subsystem, given the composition of its members?Is it possible to indicate the strategies that can be adopted by coalitions when considering the categories of the five-sided model of stakeholder influence?When recognizing the roles of coalition members, evidence emerges about the arenas and strategies employed by coalitions to bring about institutional change.

Advocacy coalitions of the Belo Monte hydroelectric power plant
This section presents the relatively stable system parameters (Table 4) and the belief structure that enabled the coalitions' identification (Tables 5 and 6).Economic development increases the demand for electricity.The country's energy matrix favors hydroelectricity, and its greatest generation potential is in the Amazon.

Sociocultural Values and Social Structure
(1) the Xingu region has been the target of a failed policy of colonization by military governments, (2) the local population has a history of fights for basic rights (health, for example) by strengthening social movements, (3) previous infrastructure projects in the Amazon have also generated controversy, such as 'Transamazônica' highway; (4) the Catholic Church has a strong presence in the region, (5) the fight for land has already caused countless deaths, the Amazon has a symbolic dimension that sensitizes a large part of Brazilian society and foreign entities, (7) the Xingu River is considered an ecological sanctuary and is sacred to indigenous peoples.

Institutional Context
National Water Resources Policy, Law 9433/97; the National Environment Policy; the Conama resolutions; the Electric Energy Market Model (Law 8987/95; Law 9074/95); specific environmental licensing rules and; the Xingu Sustainable Regional Development Plan Source: elaborated by the author.
Table 5 shows the three potentially existing coalitions that make up the political subsystem examined in this study.Once specified which beliefs are shared by each one, the distribution of previously identified stakeholders is analyzed, and it is possible to discuss, given their composition, the stakeholder coalitions' salience and roles.

Policy Core Beliefs
1-Hydroelectric use of the Xingu River and environmental preservation: (a) The work will cause irreparable negative effects on the environment, (b) the restrictions imposed on the project minimize the impacts.2-Public benefited from the project: (a) The project will not benefit the local population, from Pará, or even from the northern region, (b) the work will benefit both the center-south and the local population, the state, and the northern region, in addition to electro-intensive industries, (c) only electro-intensive industries with foreign capital will benefit.3-Enterprises in the Amazon: (a) The Amazon is a sanctuary and should not suffer such interventions, (b) the Amazon is a source of wealth and the country should explore it, (c) the Amazon is a source of wealth and the country should explore it as long as the environment is preserved.4-The energy demand/supply ratio in Brazil: (a) Brazil needs to expand the energy supply, (b) Brazil would already benefit from the modernization of the existing system, (c) Brazil does not need to increase its energy supply, but to reduce its demand.5-The uses of the Xingu River: (a) The Xingu River must be exploited to produce energy to its full potential, (b) the Xingu River must be explored with other uses, such as tourism, transport, and fishing.6-The Brazilian energy matrix: (a) The matrix must prioritize major energy uses, (b) the matrix must prioritize alternative sources of renewable energy, (c) alternative sources of renewable energy must be balanced with large uses.7-The value of Belo Monte: (a) Belo Monte generates value for the country when the energy supply increases but does not generate value for the region, (b) Belo Monte generates value for the country and the region, (c) Belo Monte does not generate any value for the country or the region.8-The state of Pará as an energy producer: (a) Pará has the vocation of being an exporter of electricity to other industrial centers, (b) Pará must develop industries and benefit from the energy it produces.9-Regional insertion measures: (a) Regional insertion measures can benefit the local population, (b) regional insertion measures are a fallacy and a tool to legitimize the enterprise.10-Participatory decision-making and systems of social accountability: (a) Broad participation as a key element of the policy, including governmental and non-governmental actors of all types, b) emphasis on the participation of the governmental authority and representatives of selected groups from organized civil society.

Secondary Beliefs
1-The viability of the enterprise: (a) The enterprise is not technically and ecologically viable, (b) the venture is viable from all points of view.2-Carrying out environmental impact studies: (a) environmental impact studies are consistent.They were done impartially and autonomously, (b) environmental impact studies are inconsistent.They were biased, serving the interests of those who formed them.3-The actions representing the contribution of the entrepreneur: (a) The entrepreneur must invest in state services that have been neglected throughout history, (b) the entrepreneur must stick to the points specified (conditions) by the licensing agency, (c) the entrepreneur must be concerned with the financial return of the work, and making a profit is their social responsibility.It is essential to note the presence and relevance of the MPF, as its action balances the dispute against the dominant coalition.In addition, given its nature, the arena chosen for the confrontation and the strategies used led to the judicialization of the project.It was possible to observe that the MPF of Pará filed at least twelve lawsuits to paralyze the feasibility studies or the works of the power plant.The Materialistic Political Coalition is made up of actors favorable to Belo Monte and seems to have a clear alignment in its deep core beliefs.However, when considering the policy core beliefs, some stakeholders' positions differ, such as their opinion regarding the state of Pará's attitude as a producer of electricity, and the perception about the exploitation of natural resources in the Amazon.The stakeholders' position also differ when considering the secondary beliefs (the actions representing the entrepreneur's contribution, access to information about the power plant, and public hearings).This coalition envisages financial gains for the region, for Pará or for itself, as in the case of local business people who believed in the emergence of a virtuous economic cycle, in which they would provide equipment and services for the work.Also, politicians are still fighting for more financial compensation for Pará, such as the application of the state tax ICMS on the energy produced.Other members of this coalition act as Collaborators and Legitimizers: Camargo Corrêa, CNEC WorleyParsons, Consórcio Construtor de Belo Monte, Engevix, Petro and Funcef Funds, Leme Tractebel -GDF SUEZ, Light, NeoEnergia, Odebrecht, Queiroz Galvão, Vale do Rio Doce, ACIAPA, Federation of Industries of Pará and Forte Xingu (local entrepreneurs).The analysis of each coalition's composition showed that the mixed blessing stakeholders have not yet been considered (Savage, Nix, Whitehead and Blair, 1991).The Belo Monte Consortium, politicians and local leaders, and the Government of Pará stand out as potential members of the materialistic political coalition.At first, these actors are not fully favorable to the enterprise and use their influence to bargain and act dubiously without clear principles or beliefs.The other stakeholders in this category are regulatory agencies and agencies of the judiciary branch.
Unlike the MPF of Pará, which seems to be a member of the idealistic political coalition, they seem to be neutral actors in the subsystem.The role of the Brazilian National Congress is emphasized.Although interviewees mentioned the members of the congress generically (considering them as actors favoring the enterprise), it was possible to observe in the shorthand notes the existence of politicians aligned with the three coalitions.For example, the majority of politicians from the North region act in favor of Belo Monte, because they want more resources for the development of the region.
The hegemonic coalition is the technocratic political coalition and, therefore, the Belo Monte hydroelectric power plant project is a reflection of its beliefs.However, considering its history and the trajectory of the energy sector, it is clear the learning process that the dominant coalition went through.Perhaps the most striking example is related to the policy core beliefs "Enterprises in the Amazon" and "Measures for Regional Insertion."Since the project was publicly rejected at the 1st Meeting of Indigenous Peoples of the Amazon in the late 1980s, its original configuration has undergone several changes.The main one was the reduction of the flooded forest area from 1225km² to 516km², with the objective of not directly affecting indigenous lands.This was possible due to the use of a system of channels that configure Belo Monte as a "run-of-river hydroelectric power plant."However, this means that when the water flow decreases, less energy is generated.Another significant change was the incorporation of the Belo Monte Region Sustainable Development Plan (PDSBM) and the Regional Insertion Plan (PIR).Nevertheless, it is worth mentioning that, since the Ministry of Mines and Energy's National Electric Energy Plan 1987/2010, there was concern about issues related to the environment and the regional insertion of the projects.
It was also observed that, if the dominant coalition manages to change the belief structure of another coalition (or its members), the stakeholder salience can change, i.e., an agent's legitimacy can be expanded or its demands considered more urgent by the dominant coalition.If the change is profound enough, an agent can change its role in the subsystem.In the case studied, two associations of local entrepreneurs were empowered with prestige and the ability to manipulate meanings based on interactions within the materialistic political coalition.The rapprochement with the technocratic political coalition influenced stakeholder salience and the role they played, acting as 'legitimizers' definitive stakeholders.
Based on the ACF, the definition of the political subsystem is based on the stakeholders' actions.Therefore, the study proposes that both the volatility of the roles and stakeholder salience depend on the relatively stable system parameters.Thus, for example, the symbolic dimension of the Amazon and the fact that the most significant Brazilian hydroelectric potential is located precisely in that region amplifies the stakeholder salience of riverine populations and indigenous peoples.These actors have interests considered legitimate, urgent demands, and the power to influence the definition of policy goals for the region.Likewise, the institutional context legitimizes the roles played by stakeholders.In the case presented here, the environmental licensing bodies had a leading role, with greater stakeholder salience as controllers or collaborators.

Coalitions and their strategies for gradual and transformative institutional change
This section specifies the main movements of coalitions influencing the process of implementation of the Belo Monte power plant.Then, the integrative approach proposed by Vieira and Gomes ( 2014) is completed, adding the gradual and transformative institutional change (GTIC) proposed by Mahoney and Thelen (2010) to characterize the behaviors and strategies of each coalition.

Table 7 Coalition strategies
Technocratic Political Coalition 1-In 1989, Eletronorte admits that it no longer uses indigenous names and announces that Kararaô would be called Belo Monte.
2-In 1994, a new project is presented with the objective of reducing the resistance of environmentalists and foreign investors.Significant changes are made, such as the reduction of the plant's reservoir from 1225 km 2 to 400 km 2 and the non-flooding of the Paquiçamba Indigenous Area.
3-In 1995, the National Privatization Program modified the legislation on public services to condition concessions to bidding processes and remove large consumers from the concessionaires' commercial monopoly, ensuring free access to the transmission and distribution systems.4-In 1995, two laws came into force: Law 8987/95, which created conditions to encourage the participation of private capital and competition in the construction of new projects, and Law 9074/95, which brought legal bases for large energy consumers interested in purchasing energy regardless of the generating company in the respective region.5-In 1996, the Brazilian government created the regulatory agency National Electric Energy Agency (ANEEL), and the National Electricity System Operator (ONS), the Wholesale Energy Market (MAE) and the Independent Energy Producer (PIE), privileging expansion of the sector through private capital.
6-Also, in 1996, the Ministry of Mines and Energy (MME) and the state-owned enterprise Eletrobrás promoted a bid to choose consulting companies to prepare the Brazilian Electricity Sector Restructuring Project.7-In May 1998, Law 9648/98 was enacted, aiming at the restructuring of Eletrobrás and its subsidiaries.8-In 1999, the Expansion Planning Coordinating Committee (CCPE) was created to lead the planning of the expansion of the system, taking over the tasks of the former Electrical Systems Planning Coordinating Group.9-In 2000, the Pluriannual Plan 2000/2003 is presented incorporating Belo Monte as a strategic work to increase the energy supply in the country and as a structuring project for the Development Axis -Madeira/Amazonas.10-In May 2001, during a time of crisis and electricity rationing, the MME announced an emergency plan worth USD 30 billion to expand the energy supply.The government planned to build 15 hydroelectric power plants, including Belo Monte.
11-2001 was also marked by the issuing of the Provisional Measure MP 2152-2, also known as "MP do Apagão" (MP of the blackout), which, among other things, determines that Conama establishes simplified licensing for projects in the electricity sector with low environmental impact.
12-Resolution n. 02, of the National Council of Energy Policy (CNPE), enacted on September 17, 2001, defines Belo Monte as a project of strategic interest in planning hydroelectricity expansion until 2010.It recommends that studies on the environmental impact and multiple uses of reservoir waters be carried out with the participation of the ministries of the mines and energy (MME), environment (MMA), planning (MPOG), development (MDIC), the regulatory agency of the electricity sector (ANEEL), and the regulatory agency of the water sector (ANA).
13-In March 2002, CNPE created a working group to prepare a viability plan for Belo Monte.14-In 2002, Eletronorte organized, at the University of Brasilia, the seminar "Economic Dynamism and Environmental Conservation: a Challenge for Belo Monte... and for everyone," aiming to subject the Belo Monte Region's Sustainable Development Plan (PDSBM) and the Regional Insertion Plan (PIR) to public evaluation.15-In 2003, Luiz Pinguelli Rosa assumed the presidency of Eletrobrás, stating that the Belo Monte construction project would be discussed and solutions for economic and social development around the dam would be on the agenda.

Idealistic Political Coalition
1-In 1988, a symposium was held at the University of Florida (USA) on proper management of tropical forests.It included indigenous leaders from Brazil who report the impacts expected from the plan of energy exploitation in the Amazon as well as the IBRD's intention to finance the hydroelectric project in Xingu.
2-In May of the same year, the MPF filed a public civil action aimed at suspending the studies conducted by Fadesp, alleging irregularities in the process of hiring the organization.The MPF also argued that the work should be licensed by IBAMA and not by the state government of Pará, since the Xingu River also crosses the state of Mato Grosso.
3-In February 1989, the 1 st Meeting of the Indigenous Peoples of the Xingu is held in the city of Altamira.The objective is to draw the attention of society to the construction of the Hydroelectric Complex of Xingu and to protest against the decisions made regarding the Amazon without the indigenous population's participation.4-In May 1989, the 1 st Meeting of Indigenous Peoples of the Volta Grande Region of Rio Xingu takes place, in addition to the Meeting of Communities of Volta Grande do Rio Xingu.In these meetings, the communities confirmed their decision against the construction of the Belo Monte power plant and published manifests asking the authorities for further reflection on the electricity sector.8-On March 28, 2006, the power plant's environmental licensing process is suspended through a legal injunction.The decision prevents the continuity of the works until the National Congress hears the affected peoples.A year later, the same Federal Court in Altamira backs down and dismisses the MPF's request to annul the licensing made by IBAMA.The decision sets a negative precedent by stating that the National Congress can authorize the implantation of hydroelectric power plants in indigenous lands without the need for a specific law or consent from affected peoples.9-In November 2009, a public hearing of the Inter-American Commission on Human Rights takes place in the USA, where the case of Belo Monte is presented as an example of the impact caused by large dams in Latin America.10-In April 2010, the Human Rights and Minorities Commission of the Chamber of Deputies decides to ask Norte Energia S.A. for explanations about accusations made in a public hearing.
11-On December 2, 2010, indigenous and riverine people engage in a demonstration at the Senate, protesting against the power plant after a public hearing at the Senate's Human Rights Commission.At the hearing, the Indigenous peoples accused FUNAI of abandoning them.
12-A letter is also sent to the UN denouncing the violation of the right to free, prior, and informed consent, provided for in Convention 169 of the International Labor Organization (ILO) and the UN Declaration on the Rights of Indigenous Peoples.
Source: elaborated by the author.
Table 7 shows that the composition of coalitions influences the type of action taken.As the technocratic political coalition has always had access to the state's structure, several reforms were carried out in the electricity sector to make the structure more complex and strengthen a political point of view.In addition to the creation of bodies in the public administration (e.g., ONS, ANEEL), laws were also drafted to regulate the electricity market, which resulted in the participation of the private sector (and therefore members of the materialistic political coalition) not only through bids, but also contributing to direct the sector's expansion policy.The bidding model also has a financing system based on the BNDES, further facilitating companies who are members of the coalition to access power.Thus, it is possible to say that the existing institutional context is at the service of the technocratic political and materialistic political coalitions.On the other hand, the so-called idealistic political coalition seems to have no voice.It fights mainly with the assistance of the Public Prosecutor's Office and also has the support of some members of the National Congress.The actions of this coalition consist mainly of social mobilization based on events, demonstrations, and taking issues to court.The materialistic political coalition, however, seems to act more discreetly and faithful to the dominant coalition, intervening in issues and specific moments.Although composed of economically powerful actors, its performance was more visible within the scope of the National Congress.
When analyzing the case based on the GTIC, the actions of the technocratic political and materialistic political coalitions are characterized as promoting institutional change classified as displacement, conversion, and layering.In turn, the idealistic political coalition engaged in changes by conversion and layering, as shown in Figure VIII.The technocratic political coalition works to change through displacement by introducing laws, provisional measures, and resolutions.This coalition's interpretation of article 231 of the Brazilian Federal Constitution, which provides on the exploitation of resources in indigenous lands, is considered as a change by conversion.As for the activities of the materialistic political coalition, the legislation that authorized the implementation of the Belo Monte power plant (Legislative Decree Project 1785/05, approved by the Chamber of Deputies, and PDS 343/05, approved in the Senate) can be considered changes by Layers.Finally, the idealistic political coalition, acting mainly with the support of the MPF, sought to interpret Article 231 of the federal constitution to suit its interests.According to the coalition, the article is ambiguous because it does not regulate how the public hearings should be conducted with indigenous populations, stating that these consultations are insufficient and that they were conducted by IBAMA, when they should have been carried out by the National Congress.The coalition has always sought to explore the ambiguity of the rule, aiming for change through conversion.It is worth mentioning that the coalition advocates the creation of a complementary law that regulates this article of the federal constitution and, therefore, it is possible to assume the emergence of a movement of institutional change by layering.In the case of Belo Monte, there is a hegemonic and another coalition that does not need to act as opponents in search of institutional change, but need to change the institutional context to support the balance of power in place.Therefore, gradual and transformational institutional changes will not always be undertaken by groups in an unfavorable condition, as Mahoney and Thelen (2010) suggest.The framework of the technocratic and the materialistic political coalitions in the typology the authors proposed, suggests that both may have acted either as mutualistic or opportunists.On the other hand, the idealistic political coalition seems to have always acted as an insurrectionary.
In conclusion, the changes in the policy took place as a consequence of pressure by public opinion (motivated by the actions of the idealistic political coalition), learning from the incorporation of new technical information, and, above all, from competitive processes of gradual and transformative institutional change (Jenkins-Smith et al., 2014;Mahoney and Thelen, 2010).However, the changes in Belo Monte seem to be concessions that the dominant coalition had to make for the project to become viable also from a social and environmental point of view.Putting the beliefs of the technocratic political coalition in this perspective (Tables 5 and 6), these concessions reflect a vision marked by the following deep core beliefs: 1D, 2B, 2C, 3C, 4C, and 5D.The following policy core beliefs also appear to guide these concessions: 1B, 3C, 6C, and 9A.
When analyzing the belief structure of the technocratic political and idealistic political coalitions, despite the clear differences, there are points of convergence.As shown in Table 5, in the deep core beliefs, there is agreement on the beliefs 1D, 2C, 4C, and 5D.Regarding the policy core beliefs, there is a convergence around the beliefs 1B, 3C, and 5B.As for the secondary beliefs, there is no convergence between the two opposing coalitions.It is possible to speculate that, throughout Belo Monte's history, both coalitions seem to avoid more radical beliefs.It is suggested that this may be evidence of the political change process indicated in the ACF.Some questions may be asked based on this finding: is it possible to analyze and position political subsystems according to a degree of maturity around the consensus produced?Is it possible that more mature political subsystems produce fewer disputes between coalitions?In situations like these, how do the processes of GTIC take place?Does the increased incidence of learning processes among coalitions generate less competitive processes of GTIC?Finally, it is necessary to observe whether the beliefs identified indeed materialize in the institutions that regulate the political subsystem.An example in the case studied is Resolution n. 02, of September 17, 2001, of the National Energy Policy Council (CNPE), which is a definitive stakeholder classified as a regulator and belonging to the hegemonic coalition.The resolution establishes that the studies on environmental impacts and the multiple uses of the water in the power plant's reservoir should be produced by Eletronorte and monitored by other entities, such as MMA, ANA, ANEEL, and MPOG, to assess the potential of the enterprise in the region's economic and social development.Nevertheless, the idealistic coalition, in its interpretation of Article 231 of the federal constitution, opposes the resolution demanding broad social participation in the process, as a key element of the policies in a democratic and participatory state.Bispo and Gomes (2018) state that "the dynamics of the social environment where public policy is applied is made up of contradictory and structurally linked relationships" (p.1268).This seems to be the case with the Belo Monte hydroelectric power plant.This research described and analyzed how the relations among stakeholders clustered in advocacy coalitions are guided by a set of institutions, which are questioned and confronted, at the same time as they may change and, consequently, bring changes to a policy.

Final Considerations
Based on the models of stakeholder analysis adopted, it was possible to show how volatile the stakeholder salience is and the roles they play.Also, it was observed that salience and roles are subject to a dynamic that is understood when considering the context (political subsystem) and moments of interaction.The research identified a limitation of the five-sided model of stakeholder influence proposed by Gomes et al. (2010) when applied in contexts other than local governments, proposing a new category of stakeholder called "Sentinels."Regarding the Advocacy Coalition Framework (ACF), the study examined the coalitions' composition, by characterizing members as stakeholders with different degrees of influence and with different roles within the political subsystem.Thus, the stakeholder analysis conducted previously indicated the arenas, resources, and types of strategies to be adopted.It was observed in the case studied that the technocratic political coalition was hegemonic due to the characteristics of its members: stakeholders with great power, legitimacy, and urgency of demand, exercising roles such as collaborators, regulators, agenda setters, and controllers.The idealist political coalition, acting in a counter-hegemonic struggle, has members with more restricted performance (controllers and legitimizers).The main strategies of the coalition were to judicialize the issues related to Belo Monte and to mobilize public opinion.
The study corroborated Vieira and Gomes (2014), suggesting that the analysis of gradual and transformative institutional change (Mahoney and Thelen, 2010) can be beneficial if accompanied by the use of stakeholders analysis and the ACF.The case study indicates that the behavior in favor of this type of change is not exclusive to actors in an unfavorable situation.Hegemonic groups must fight to maintain an unstable balance of power (Thompson, 1995).In addition, the study identified how certain beliefs about a public policy materialize in institutions, evidenced in Resolution n. 02 of CNPE, enacted on September 17, 2001.
One of the limits of this research is that the stakeholder analysis, identifying and characterizing the coalitions, may simplify the reality since coalitions are analyzed as units.A more refined analysis of each stakeholder can reveal internal differences, for example, representatives of different coalitions in the same organization (as observed in the case of the National Congress).The shorthand notes revealed the existence of commissions in the Chamber of Deputies and in the Senate in favor of the power plant, opposing the commission of Human Rights and Minorities Commission of the Chamber of Deputies, which adheres to the interests of the idealistic political coalition.
Another limitation is that the approach adopted was unable to capture all the nuances of the disputes and their motivations.A case study on Belo Monte using other frameworks may shed light on issues that were not observed in this research.Thus, it would be interesting to examine the actors' relationships, grounded on the literature on governance and public policy networks.For example, a study could use Love and Stout's (2017) work on public encounters between public agents and social actors, and their influence on policy outcomes.These social interactions are structured by a set of formal and informal rules and can be observed in a broader context of governance, during policy planning and implementation.According to Love and Stout (2017), the approach of integrative governance encounters can generate consensus and satisfactory results for a value-based policy such as the search for integration, inclusion, consensus, and belonging.It is an alternative to an approach supported by hierarchical relations of domination in an elitist perspective (Institutional Governance Encounters).
The examination of this study's limitations and contributions results in a series of questions to be addressed in future studies.For instance, it is important to investigate who are the typical stakeholders of hydroelectric plants and verify from multiple cases whether they are different in their belief and value structures.Regarding the specification of coalitions, it is recommended to test the approach suggested by Weible et al. (2009) and Jenkins-Smith et al. (2014) regarding the existence of sub-coalitions.Would the materialistic political be a sub-coalition belonging to the technocratic political coalition?Future research may also follow the model proposed by Gomes et al. (2010), testing it in other contexts, also assessing the relevance of the new category proposed here, the 'Sentinels.' Future studies could include investigating ways to quantify stakeholder salience according to the variables 'power,' 'legitimacy,' and 'urgency of the demand' (Mitchell et al. 1997).Another issue to be explored is the increase in salience given the cooperation between stakeholders; that is, how an actor can benefit from establishing a cooperative relationship with a more influential actor.As for studies regarding the gradual and transformative institutional change, it would be fruitful to discuss in depth the changes undertaken by hegemonic groups.This perspective may reveal other behaviors and strategies.It is also recommended to deepen the studies on how beliefs and values are manifested in the set of institutions that regulate the policy.Finally, historical analysis is suggested to understand the territorial appropriation of the Amazon and to examine further the role of other economic sectors interested in the implementation of Belo Monte, such as the mining sector.
; -Chamber event with the Heinrich Böll Foundation and Rios Vivos Coalition: June 18, 2002; June 19, 2002; June 20, 2002 -Permanent Subcommittee -Alternatives and Energy Solutions for the Amazon Region: June 10, 2003; -Amazon Commission, National Integration and Regional Development: October 15, 2008; -Commission on Human Rights and Minorities: July 7, 2010; -Environment and Sustainable Development Commission: May 17, 2011; -Temporary subcommittee to monitor the execution of the works at the Belo Monte Power Plant: November 5, 2010; Distributive justice: (a) low concern for equity, (b) exclusive focus or prioritization of intra-generational equity, (c) exclusive focus or prioritization of intergenerational equity, (d) intra-generational and intergenerational equity.Manifestation that shows concern for present and future generations.2-Ethical relationship between man and nature: (a) Pure anthropocentric vision, (b) moderate anthropocentric vision.It includes concerns related to man's responsibility to nature, (c) eco-centrism.Emphasis on interaction.Criticizes the instrumental view.Reconciling position between anthropocentrism and biocentrism, (d) biocentrism.Reverence for the intrinsic value of nature.3-Knowledge as a value: (a) Disregard of knowledge as a value, (b) exclusive focus or prioritization on traditional knowledge, (c) exclusive focus or prioritization of technical and scientific knowledge, (d) explicit appreciation of both scientific and traditional knowledge.4-Core view about the state: (a) A clear presence of the state, with a focus on planning and direct action, which goes beyond regulatory actions, with omission or conflict in relation to the attitude, (b) it must be minimal.State action in a limited field.State regulation is permitted in strategic areas, (c) it must be democratic and participatory.Emphasis on the relevance of broad citizen participation in government decisions.5-Essential view on natural resources: (a) pure utilitarianism, (b) exclusive focus or prioritization of sustainable use, (c) exclusive focus or prioritization of preservation, (d) reconciliation between sustainable use and preservation, implying reconciliation between the conservationist and preservationist attitude.
4-Dissemination of information about the work: (a) Access to information is wide, b) access to information is restricted, (c) access is apparently wide, but conveniently open (only what matters is disclosed and discussed).5-Public hearings: (a) The hearings were held legally, (b) the hearings did not respect the conditions established in the law, (c) the law is unclear on how to proceed.6-The nature of public consultations: (a) Consultations must be consultative, (b) consultations must be a decision-making instance.7-The action of IBAMA and Funai: (a) They acted within the law, and impartially, (b) they were instruments in favor of making the venture viable.8-The role of the MPF: (a) The action of the MPF is guided by impartiality, (b) MPF's action is partial, influenced by the values of its representatives.9-Alternative Solutions: (a) Alternative energy sources combined with the modernization of already implemented systems can supply the demand, (b) alternative sources of energy generation are unable to meet demand.Source: elaborated by the author.The Idealistic Political Coalition is mainly composed of environmentalists, social movements and nonprofits, sectors of the Catholic Church, academics, and specialists.Their technical-scientific arguments strengthen the coalition.Local politicians and members of Congress are part of this coalition.The coalition has 'controllers' and 'legitimizers' stakeholders: Federal Public Ministry (MPF) of Pará, AIMA, CIMI, Federation of Agriculture and Livestock Industry of Pará, Fundação Viver, Produzir, Preservar, MAB, Movement for the Development of the Transmazônica and the Xingu (MDTX), Movement Xingu Vivo Para Sempre, and representatives of the riverine populations.

Finally
, the Technocratic Political Coalition includes stakeholders who act as collaborators, regulators, agenda setters and Controllers: IBAMA, BNDES, Steering Committee of PAC, CNPE, Ministry of Planning, Office of the Chief of Staff of the Presidency of the Republic, Ministry of Mines and Energy, Energy Research Company, FUNAI, INCRA, Ministry of the Environment, and National Electricity System Operator (ONS).The Belo Monte hydroelectric power plant is certainly a political project that mobilized a large part of the state's structure.
16-In 2004, Luiz Pinguelli Rosa publicly reaffirmed his willingness to build Belo Monte.17-On March 15, 2004, Laws 10847/04 and 10848/04 created a new model for the Electricity Sector, creating the Energy Research Company (EPE) and a new framework of rules for the commercialization of electricity in a profound restructuring of the expansion planning of electrical systems.18-In 2008, Belo Monte becomes a project of the Growth Acceleration Program (PAC).19-In July 2008, the CNPE defined the Hydroelectric Use (AHE) of Belo Monte as the only hydroelectric use to be explored on the Xingu River, through CNPE resolution 06/2008.20-In 2009, FUNAI authorizes the construction of the power plant, affirming to IBAMA that it considers the project viable, with conditions applied.21-On December 10, 2009, the Federal Court suspends licensing and requests further public hearings.On the 11 of December, the law injunction was overthrown, and IBAMA returned to analyze the project.Without the prior license of this agency, the federal government was prevented from carrying out the concession bid, scheduled for December 21.22-In February 2010, the Ministry of the Environment authorizes the construction of the Belo Monte plant with the publication of the environmental license.Materialistic Political Coalition 1-In April 2002, the Mines and Energy Commission request an audience with the President of Brazil's Supreme Court (STF) to address the ban on studies related to Belo Monte.2-On July 6, 2005, Legislative Decree 1785/05, which authorizes the implantation of the power plant, is approved by the Chamber of Deputies.A week later, the Senate also approves the bill, now called PDS 343/05.3-On November 19, 2009, Belo Monte is the subject of a public hearing in the Senate.The Participatory Legislation Human Rights Commission discussed the terms of the study of environmental impact with the MPF.
5-In August 1989, the Movement for the Development of Transamazônica and the Xingu (MDTX), which brings together about 113 social organizations, prepared the document entitled "SOS Xingu: a call to common sense about damming rivers in the Amazon."6-On July 21, 2005, the Instituto Socioambiental, the Coordination of Indigenous Organizations of the Brazilian Amazon (Coiab), Greenpeace, and the Center for the Rights of the Populations of the Carajás region legally questioned the Attorney General's Office against the implementation of the Belo Monte power plant.The organizations claim that Congress authorized the work without listening to the affected communities.7-On August 26, 2005, based on this question, the Attorney General's Office filed an ADIN (Direct Action of Unconstitutionality) with Brazil's Supreme Court against the decree approved by Congress.

Table 1 Interview questions and participants Interview questions and list of participants
enterprise?Who or what are these pressuring forces?How has it been exercised?Is there a link between these sectors?How does it happen?And how do Belo Monte officials and supporters react?5-What are the institutions (rules and laws) that regulate Belo Monte?Is there any aspect that causes ambiguity or impairs the implementation of the project?How do you identify the position of the favorable and contrary actors in relation to these norms/laws?Interviewee H: Manager of Land and Reallocations Affairs at Norte Energia; Interviewee I: President of Associação dos Índios Moradores de Altamira; Interviewee J: Eletronorte engineer; Interviewee K: professor of the Department of Anthropology at UFPA; Interviewee L: Environmental Analyst at Eletronorte; Interviewee M: consultant at Elabore; Interviewee N: businessman from Altamira, representative of the Forte Xingu Movement; Interviewee O: Special Secretary for Infrastructure and Logistics for the Sustainable Development of the Government of Pará; Interviewee P: Eletronorte engineer.Source: elaborated by the author.

Table 3 Classification of the stakeholders of the Belo Monte hydroelectric power plant Definitive Supportive Stakeholders Collaborators
Office of the Chief of Staff; Ministry of Mines and Energy, Presidency of the Republic, Energy Research Company.
ControllersFUNAI; IBAMA; INCRA; Ministry of the Environment; National electricity System Operator (ONS) Legitimizers ACIAPA; Federation of Industries of do Pará; Forte Xingu (local entrepreneurs) Definitive Non Supportive Stakeholders Controllers Federal Public Ministry in Pará Legitimizers Associação dos Índios Moradores de Altamira; Conselho Indigenista Missionário; Federation of Agriculture e Livestock of Pará; Movement of People Affected by Dams; Xingu Movement for the Development at the Trans-Amazon Highway and Xingu; Movement Xingu Vivo Para Sempre Definitive Mixed Blessing Stakeholders Regulators National Congress (Chamber of Deputies and Federal Senate) Controllers ANEEL; National Congress (Chamber of Deputies and Federal Senate); IPHAN STF; TCU; Government of Pará; ICMbio Legitimizers Belo Monte Consortium (representation from the 11 cities); Government of Pará Source: elaborated by the author.

Table 8 The institutional change strategies of each coalition Coalition Strategies for gradual and transformative institutional change (Displacement, Conversion, Layering, and Drift)
Conversion: distinctive interpretation given to article 231 of the federal constitution.MaterialisticLayering: Legislative Decree 1785/05, approved by the Chamber of Deputies, and PDS 343/05, approved by the Senate.IdealisticConversion: distinctive interpretation given to article 231 of the federal constitutionLayering: lobby for the creation of a complementary law that regulates article 231 of the federal constitution.Source: elaborated by the author.