Power relations in the co-creation of water policy in Bolivia: beyond the tyranny of participation

There is a tendency to assume that the construction of a public policy is a linear and straightforward process, which follows a logical sequence of first identifying a need, then formulating a policy, implementing it, and finally concluding with its evaluation (Brock et al., 2001, p. 13). However, in reality, it is a complex process influenced by power relations between different actors (Brock et al., 2001). Ignoring power relations may limit reaching truly active and meaningful engagements of different stakeholders that are arguably critical in the process of (co)creation of a public policy. As a result, empowerment of vulnerable and marginalized groups does not necessarily occur (Brouwer et al., 2013).

To avoid that some stakeholders dominate the process while others end up being ‘abused, overruled or excluded’ (Brouwer et al., 2013, p. 13), it is crucial to recognize and deal with power imbalances from the beginning of, and throughout, the whole process of (co)creation of a public policy (Brouwer et al., 2013). Neglecting power imbalances might actually support more powerful stakeholders rather than benefit the most vulnerable and marginalized (Schiffer, 2007), an outcome often referred to as the tyranny of the participative approach (Cooke & Kothari, 2001; Boelens & Hoogendam, 2002).

Also, in the formulation of public policies related to natural resources, power imbalances may be present. In reference to water rights, López et al. (2019) observe this issue has been analysed rather ‘narrowly’, either from a legal, technical, or an economic standpoint, even though water rights are embedded in complex socioecological systems. Similarly, Boelens et al. (2007) claim that a water policy can be improved by analysing and identifying power relations that reinforce both official and customary water rights at a local, national, and international level. Water governance is an interesting arena to explore both power and collaboration because of the complexity of the interactions between different socio-political structure (Brisbois & de Loë, 2016).

Power is defined by Brouwer et al. (2016, p. 73) as ‘the ability of actors to achieve their goals’. This dynamic interpretation of power allows for different expressions and forms of power, which Green (2016), building on Follet's (1918) appreciation of power, refers to as power ‘within’, power ‘with’, power ‘to’, and power ‘over’. From these expressions of power, powerover usually has a negative connotation of domination and/or control exercised by one individual, group, or organization over another one (Brouwer et al., 2016). This expression of power is seen as a potential ‘obstacle’ for agency and collective empowerment because it ‘undercuts the ability of agents to actualize their own desires’ (Pratt, 2011, p. 82).

A more equitable and empowering change may be more viable when these expressions of power and their interrelations are acknowledged (Gaventa, 2021). For the powerless to make a demand, first they need to ‘develop a sense of self-confidence and a belief in their own rights’ (powerwithin them) to then get organized as a group and work together (powerwith) (Green, 2016, p. 33). Finally, power transforms into the powerto act and define ‘their own futures’ (Gaventa, 2021, p. 5). Many scholars refer to powerwith and powerto as the agency (Miller et al., 2006; Gaventa et al., 2011; Whaley & Weatherhead, 2015).

The dynamic character of power is not only reflected in space and time, but also in different levels of governance (Green, 2016). Lukes (1974) argued that power needs to be explored and studied outside of decision-making spheres, and that more attention should be paid to other aspects such as real and subjective interests, and observable and latent conflict. Building on Lukes' work, Gaventa (2006, p. 25) claims that power must be studied and understood ‘in relation to how spaces for engagement are created, and to the levels of power (from local to global), in which they occur’. These forms, spaces, and levels of power have been combined in an analytical approach referred to as the PowerCube, which serves as the analytical framework of our research.

The PowerCube (Gaventa, 2005) emerged from the need to examine explicitly and graphically the interrelations of three aspects of power: forms, levels, and spaces (Figure 1). The model identifies three forms of power: visible, hidden, and invisible (Gaventa, 2005). These forms of power were defined by Gaventa (2005, p. 15):

• Visible power: Observable decision-making. This level includes the visible and definable aspects of political power – the formal rules, structures, authorities, institutions, and procedures of decision-making […]

• Invisible power: Shaping meaning and what is acceptable. […] shapes the psychological and ideological boundaries of participation. Significant problems and issues are not only kept from the decision-making table, but also from the minds and consciousness of the different players involved, even those directly affected by the problem […]

• Hidden power: Setting the political agenda. Certain powerful people and institutions maintain their influence by controlling who gets to the decision-making table and what gets on the agenda. These dynamics operate on many levels to exclude and devalue the concerns and representation of other less powerful groups […]

It is important to mention that hidden power also is about who is influencing the process outside the public eye because they have hidden interests (Hunjan & Pettit, 2011). Moreover, it is ‘about how people affected negatively by [the process] may challenge it, to make their voices more visible’ (Gaventa et al., 2011, p. 11).

The PowerCube can be used to explore further the expressions of power mentioned above (Gaventa et al., 2011). For example, powerwithin, with, and to can be linked to hidden and invisible forms of power. Brouwer et al. (2016, p. 81) ‘[…] can be exercised from below in the form of resistance and as expressions of power to, power with, or power within. Some citizen's groups may be able to mobilise their own forms of hidden or invisible power as a strategy for empowerment and social change’.

The forms of power can take place in different spaces or arenas of engagement (Gaventa, 2006):

Closed spaces: Spaces where only certain ‘elite’ actors make decisions without the inclusion, consultation nor involvement of ‘the people’ (p. 26).

Invited spaces: Spaces that rise as an attempt to counteract closed spaces by inciting ‘the people’ to participate by state and non-state organizations (p. 26).

Claimed or created spaces: Spaces created by ‘less powerful actors for or against the power holders or created more autonomously by them’ to deal with common needs or concerns (p. 27). Gaventa (2021, p. 11) adds that in claimed spaces these actors ‘can shape their own agenda or express their own voices more autonomously’.

Finally, power takes place at different levels of power, usually pre-defined as local (sub-national governments, councils, and associations), national (governments, political parties, and other nation/state authorities) or global levels (formal and informal decision-making beyond nation state) (Gaventa, 2020). However, other levels of power can be defined according to the context and scale of each individual case.

Although the PowerCube has been used in natural resources and water analysis, most studies focus on only one form of power (Brisbois & de Loë, 2016; Mehta, 2016; Roth et al., 2017; Wamuchiru, 2017; Karpouzoglou et al., 2019; Etiegni et al., 2020; Tantoh et al., 2020; Thompson et al., 2020) or do not explore other dimensions of the model (Rodriquez de Francisco & Boelens, 2014). Although Whaley & Weatherhead (2015) fully implement the Power Cube in a study of water governance in England, they do not study the influence of complex interrelations of various aspects such as interests, socio-political structures, and knowledge in power relations. While in the Bolivian context, Jacobi & Llanque (2018) did include all forms of power in their analysis, their study focused on agro-industrial and indigenous food systems. It seems that the PowerCube has not yet been fully implemented in Bolivia in the context of natural resources and water analysis. Therefore, studying the making of a water rights policy from a broader perspective unveils insights for the (co)creation of more inclusive, legitimate, and empowering public policies. This paper tackles this need by analysing how power relations are influenced by interests, socio-political structures, and knowledge in the context of the (co)creation process of a water policy. Moreover, this paper aims at evaluating if the PowerCube in all its aspects allows for such an integrated analysis of water policy, using a case study from Bolivia.

The rural municipality of Tiraque, located in Cochabamba – Bolivia, is rich in water resources, with more than 20 lagoons, rivers, and springs belonging to two basins whose flows remain constant throughout the year (Figure 2) (PTDI, 2016). Tiraque is a suitable arena to study power relations in the construction of a public water policy because it is known in Bolivia for external and internal conflicts about access to and control of water (Cossio et al., 2010). Besides these, there are other problems related to water (i.e., distribution, pollution, over-exploitation of groundwater, and limited access to irrigation water) (Rodriguez, 2020). Because of these problems, the need to protect water sources and water recharge areas was identified and included in a municipal planning tool as a measure to secure access to sufficient quantity and quality of water, triggering the need for a specific municipal policy. The design of this water policy was led by a local non-governmental organization (NGO), who followed a consultation process established by local social organizations. A draft of the policy was presented to the Municipal Council in October 2019 and the final decision was communicated at the beginning of 2020. The main purpose of this policy is:

The policy was presented to the Municipal Council for its approval. However, it was contested by some sectors who were afraid of losing not only their water rights but also their land management rights, resulting in the rejection of the policy. The Tiraque irrigators, through their social organizations, wanted proof that the policy would not affect their current rights and asked for the development of pilot projects. Currently, the NGO is implementing one pilot project to fulfil this requirement.

Different power relations, socio-political structures, interests, and knowledge systems interplay throughout the process of water policy-making, culminating in the social rejection of the policy, and the setting of a new agenda in 2020. Given this context, the development of a municipal policy for the protection of water sources and water recharge areas is an arena for internal/external struggles and conflicts, in response to new arrangements of the territory and water management and uses, and for the inclusion and exclusion of actors related to water control (Boelens & Hoogendam, 2002). It can also be a space for confrontation and rejection of external stakeholders, such as NGOs or international cooperation.

Prior to the Land Reform in 1953, most of the land and water resources were in control of wealthy landowners (mostly Spanish descendants) (Antequera, 2018). In return for shelter and food, estates were cultivated by colonos (indigenous workers). Outside the estates, small surfaces of land were owned by piqueros (mestizo and free indigenous families) (Antequera, 2018). To maintain the water supply for food production, indigenous colonos working for estates owners had to build intakes, canals, and dams (Gerbrandy & Hoogendam, 2002). As a result, until 1953, water rights were distributed among all estates' owners who invested in the construction of hydraulic infrastructure, which could also be located outside their estates (López et al., 2019).

After the Land Reform, estates disappeared and were transformed into peasant communities, and peasant unions started to emerge (Bustamante et al., 2019). López et al. (2019) define the period of 1950–1978 as a transition to community control of irrigation water. Through this period, colonos and newly established communities claimed the water rights that belonged to their former estate owners, even though such resources could be located outside the community's boundaries (López et al., 2019). A new set of ‘habits and customs’ (‘usos y costumbres’ in Spanish) was established regarding water management and uses (López et al., 2019). Water demand for agricultural production increased after the intensification and expansion of agricultural land by colonos. Water rights demands were raised by the piqueros. As a result, new hydraulic infrastructure was built and water rights were given according to families' investments (economic and labour) in these constructions (López et al., 2019, p. 206). Therefore, families who did not participate in this process were not granted water rights.

The period between 1978 and the early 1990s was marked by a major presence of the State with the re-organization of the peasant irrigation systems (López et al., 2019). In the late 1970s, the agrarian unionism gained strength with the creation of the National Confederation of Peasant Workers of Bolivia (CSUTCB for its acronym in Spanish) and the National Federation of Peasant Women of Bolivia ‘Bartolina Sisa’ (Costas et al., 2005). These peasant organizations have developed a great capacity to mobilize their grassroot members and/or to physically blockade the country to pressure the government to attend to their demands or to manifest their opposition to certain policies.

The CSUTCB, the largest organization in the country, counts over 3 million agricultural affiliates (CSUTCB, 2019). It has a hierarchical structure, whose base units are the communities represented by agricultural unions. Associated unions form sub-centrals, which in turn are grouped into cantonales, which grouped together make up central provincial representations, in turn grouped into nine departmental federations. Finally, the latter group collectively makes up the National Confederation.

The Central Provincials level is relevant for the real mobilization of grassroots members since at this level it is possible to coordinate with the smallest and lowest levels of organization (cantonal, sub-central, and agricultural unions) (Costas et al., 2005). For example, if a Central Provincial disagrees with a mobilization or blockade defined at a national level, it is difficult for the sub-centrales (the next level of organization) to act against the Central Provincial's decision. Therefore, the Central Provincials are the ‘fundamental organic nucleus’, since no decision to mobilize by the Confederation will be executed without the endorsement of the Central Provincial (Costas et al., 2005, p. 141). Once this endorsement is given, mobilizations will begin from the community level.

The organizational structure of Tiraque's peasants is well developed. The highest level of social organization is the Central Provincial Peasant Workers Union of Tiraque. From this level down to the agricultural unions, organizations are legally established and have organized and consolidated structures. According to a municipal planning tool, a total of 147 rural communities and seven neighbourhood councils are organized in sub-centrals (PTDI, 2016). At the same time, the Central of Indigenous Peasant Women ‘Bartolina Sisa’ has 1,200 affiliates (PTDI, 2016). Specifically in relation to water management, there are the Federation of Indigenous Agricultural Irrigators of Cochabamba (FRIAC for its acronym in Spanish), the Tiraque Drinking Water Committees, and the Tiraque Irrigation and Services Association (ARST for its acronym in Spanish). The FRIAC, composed of eight irrigation associations, was created in 2008 as a strategic instrument during a water conflict with a neighbouring municipality (Rocha-López, 2020).

Currently, Tiraque's development is oriented mainly towards agricultural production (potatoes, grains, and fava beans). Agriculture is responsible for nearly 88% of the livelihoods (PTDI, 2016). Families may have access to multiple sources of water and irrigation systems (López et al., 2019). Natural springs are the main sources of the daily water supply to the communities (PTDI, 2016). They are also used for animals and to irrigate different crops. In total, 34 irrigation systems were identified in Tiraque (Viceministerio de Recursos Hídricos y Riego, 2013). Therefore, water control is crucially important in (partly) irrigation-based agricultural systems, and is subject to manifestation of power because of struggles, demands, conflicts, organizational entities, rules, and norms related to water management and use. Throughout history, rural communities have developed usos y costumbres regarding water use, dealing with power relations, and water conflicts over water rights.

In recent years, several qualitative tools were designed and tested successfully to conduct power relations analysis throughout the world (Gaventa et al., 2011; Hunjan & Pettit, 2011; Pettit, 2013). This paper follows a critical qualitative research approach (Merriam & Tisdell, 2015, p. 61) to explore power relations during water policy-making. Different forms (visible, hidden, and invisible), spaces (closed, invited, and created), and levels of power (community, municipal, regional), and their interactions, were analysed using the PowerCube (Gaventa, 2006) as an analytical framework.

Data were collected through a qualitative methods approach in an intensive field work between 2017 and 2020. Primary information was gathered through semi-structured in-depth interviews with key informants, participatory observations, workshops, and transects during and after the construction of the water policy. Information from secondary sources was obtained from an exhaustive bibliographic review.

• Semi-structured interviews (n = 34) revolved around a set of questions across the PowerCube, as defined by Hunjan & Pettit (2011). Some actors were interviewed several times at different stages of the process.

• Field notes (n = 10), field videos (n = 31), and workshop audios (n = 4) were made through participatory observation at meetings with social organization and municipality's staff members, water policy workshops and transects to identify water sources and recharge areas. Likewise, meeting audios (n = 30) with different stakeholders were collected during coordination, analysis, and/or reflection meetings of the process of construction of the water policy.

• Secondary information about the case was reviewed such as state planning tools, maps, scientific papers, policy drafts, technical reports from NGOs and from a university among others.

The variety of methods and sources of information was used to gain a deeper understanding of the research questions and corroboration through data triangulation. On the one hand, data source triangulation allowed to gather multiple perspectives on power relations in the construction of the water policy. On the other hand, methods triangulation allowed to explore the same issue through in-depth interviews, participatory observations, workshops, and literature review.

All data collected were transcribed and a content analysis was developed following Erlingsson & Brysiewicz (2017). The PowerCube analysis was complemented with an exploration of the interrelations with knowledge, interest, and socio-political structures. These categories were chosen in advance based on the set of guiding questions presented by Hunjan & Pettit (2011). Quotes were translated from Spanish to English by the first author and are given in italics. The respondent (R) is given between brackets.

To unravel the interrelations between power, interests, knowledge, and socio-political structures in the process of construction of a public policy in the municipality of Tiraque, we first present a summary of the process of the making of the public policy, including the main events and actors involved. Next, the different forms of power and their relations with interests and organizational structure are analysed. This analysis considers different spaces and levels of power. Results provide important lessons for future initiatives for more inclusive and equitable processes of public policy formulation and research.

Since the early 2000s an international NGO supported local communities with irrigation projects in Tiraque. The need to have a policy for the protection of water sources and water recharge areas was evidenced by the NGO's personnel because water sources were polluted, eroded, and forested with exotic species. This urgency was also identified in a government planification tool. As a result, in 2018, the international NGO hired a local NGO as a consultant to lead the overall process of constructing a public water policy (Figure 3). The local NGO started the process through a series of meetings and workshops with different sectors and social organizations. By February 2019, the local NGO facilitated the creation of an Advocacy Group, and the construction of the policy was included in the agenda of the municipality and that of the Central Provincial. Members of the Advocacy Group were two executives of the Central Provincial of Tiraque, one from the FRIAC, one executive from the ARST, two Municipal Councillors, and one community communicator. This group was to help set the agenda for the construction of the policy.

Between April and August 2019, a series of theoretical and practical workshops resulted in the presentation of a draft in Spanish of the public policy, developed by a legal consultant hired by the local NGO. Next, a series of meetings with municipal technicians as leaders of social organizations was held at the Municipal Council to review the draft and to present it for its approval. The presidential campaign of 2019 played a key role in the process because stakeholders did not want to generate conflicts during this period (as detailed below). Later, in October 2019, the country was paralyzed for 21 days because of the outcomes of the presidential elections. At the beginning of 2020, the public policy was rejected.

In addition to the Advocacy Group, different social and institutional actors were involved throughout the process (Table 1).

Tiraque has a long history of internal and external conflicts related to water access, management, and control. Thus, the making of this policy was practically carried out in a minefield of conflicts and power relations. It is in this context that the actors promoting the formulation of the public policy had to move cautiously to avoid unintentionally stepping on someone's toes. A misunderstood word or technical concept could have unleashed a new water conflict. For example, when reflecting about a water conflict between Tiraque and a neighbouring municipality (Punata), an interviewee mentioned that ‘it's like a time bomb ready to explode’ (R1).

Visible power is about observable decision-making regarding the construction of public policy. It is mainly in the hands of decision-makers whose interests are to be maintained, hence reducing participation. An objective of social organizations is to influence the construction of public policies in favour of the peasantry, granting them legitimate decision-making power in the country. Although their leaders make decisions in favour of the peasantry at the grass roots, power itself is managed in a hierarchical and vertical manner. In Tiraque, the main decision-makers are the Central Provincial, the Mayor, and the Municipal Council.

The Central Provincial is identified as the most powerful social organization in Tiraque. For example, the candidates running for Mayor of Tiraque are chosen from the Central Provincial. Hence, political decisions at this level of the organization might allow or block processes. If the Central Provincial asks the Municipal Council to approve a public policy, it is highly likely that it will be approved. As a municipal authority mentions: ‘Mostly social organizations exercise power, through the Municipal Council, to ensure that policies are accepted’ (R6).

The legal consultant, hired by the local NGO, developed a draft in Spanish of the public policy, which was made public at the social organizations' meetings between May and August 2019. At these meetings, the inclusion and exclusion of actors were based on their decision-making power, resulting in the presence of high-ranking leading figures from social organizations, NGO's staff, Municipal Council representatives, and municipal technicians. Although these meetings are supposed to be invited spaces where public and non-government organizations can freely participate, the invitation of specific actors turned them into closed spaces, characterized by four forms of exclusion: (1) rotating and supplanting leadership, (2) a culture of gendered decision-making, (3) avoiding conflict by not inviting, and (4) flawed transmission of information.

The drafting of the policy was hindered by the inconstancy in the attendance of leaders, due to a 2-yearly rotating leadership. As a result, some leaders did not fully understand the content of the draft and felt that they were being pressured to approve a policy that in their eyes was being imposed on them. This was amplified by leaders being often supplanted by other members of the organization without decision-making power. In this regard, an interviewee who was present throughout the process mentions: ‘I think that the challenge in our context is that all those who participate, [should have the power to] decide, right? But that really did not happen’ (R5).

Likewise, although female leaders were present in these meetings, their voices were not heard, hence not visible. Women participation was considered important as exemplified by the invitation and participation of the Bartolina Sisa Women's Peasant Unions to the meetings. However, although they presented their policy proposal, it was not even considered by the Central Provincial, because water is thought to be a masculine theme. This can be illustrated by the following comment: ‘It seems to me that women have had almost no participation in this regard, as Bartolinas have only received a very brief report that the NGO has given them’ (R2). Therefore, mere attendance did not guarantee effective and legitimate participation.

Interviews also revealed how communities and landholders with water sources in their properties, were strategically excluded from the process, to avoid conflict. This group is vulnerable because it is not part of the irrigation organizations. In this regard, a farmer mentioned that ‘many families were not part [of the process]. In the Jatunchinija community alone, there are 174 unaffiliated [families]’ (R8). However, organizers of these meetings explained that because these communities were unaffiliated, they could not be identified in time to be invited.

Flawed transmission of information was the fourth form of exclusion. Information was not transmitted to the grassroots because ‘leaders themselves did not understand the message properly’ (R2). Information was not passed to the sub-centrals nor to the communities. Reaching the 147 communities individually, considering the large number and the restriction to work during office hours only, was an impossible task for the local NGO. Therefore, transmission of information from higher to lower levels of governance was difficult.

Information transmission was also limited due to language barriers. Tiraque's population is mostly Quechua-speaking. Men commonly speak Spanish as well, while women mostly speak only Quechua. Dealing with quite technical aspects of the conservation of water sources and water recharge, and using Spanish as the main language, the lack of understandable information as input for discussion and debate resulted in confusion, doubts, and distrust about the impacts of the water policy on the management and control of the water supply and recharge areas. In this regard, the following testimony shows how some communities were excluded due to language gaps and legal technicalities:

This case shows that visible power is about who the decision-makers are. But it also shows that it is about who are present in the process but have little influence, such as women. Finally, it is also about those excluded for example because of language gaps. The following section presents how excluded or opposing actors contested the process and how interests influenced it.

Hidden power is concerned with who sets the agenda and how those affected manage to challenge it. Moreover, it focuses on hidden interests that different actors may have and on how decisions are being influenced outside the public eye. Although the construction of this water policy in Tiraque aimed to address a common need identified and prioritized by multiple local actors, its 3-year development process was mainly led by a local NGO. The NGO managed to put the construction of this policy on the agendas of social organizations and the municipality, through a linear and straightforward advocacy strategy. Therefore, by setting the agenda, the local NGO exercised hidden power. This is further supported by a municipal authority who mentioned: ‘the NGO presented the request [for the policy], from there we started’ (R6). The municipal authority related to the pressure felt as follows: ‘[municipal authorities] feel pressured, compelled’ (R6). Along the same lines, an academic mentioned that the ‘The agenda is established directly by the NGO. It has been generating the process. It is the driving force behind the public policy process. They are setting the agenda’ (R2).

To prepare the ground and to show the importance of the policy, a series of initial meetings, including theoretical and practical workshops, was organized by the local NGO. Each workshop lasted ‘one day with the participation of various local actors, irrigators, peasant organization, municipal authorities’ (R3). Due to its content and the used ‘farmer to farmer’ approach, the first workshop was well rated by the attendees. In this event, technical aspects of water dynamics were covered, and a local peasant from the area gave legitimacy to the conservation of water sources as he explained how their community changed their usos y costumbres to be able to protect water sources.

To ‘continue motivating local actors’ (R3), in another workshop an authority from a municipality in a lower region was invited to explain his region's experience with the Reciprocal Water Agreements for ‘water planting’ (siembra de agua), a commonly used term to refer to groundwater recharge fostered by human interventions for infiltration, retention, and regulation of runoff waters (i.e., conservation of recharge areas, reforestation, artificial lakes, etc.). As we will see later, this workshop had a major impact on the approval of the policy. Due to the cultural differences with higher located lands, regarding water rights, confusion was generated with the concept of ‘water planting’. From the point of view of the attendees, those who ‘plant water’ would have complete ownership of the water harvested. Because these events were open and included the participation of different actors, they were identified as invited spaces, where ‘social institutions took a leading role’ (R5). A university supported this process with a communication campaign that included a ‘micro-documentary and booklets designed to sensitize organizations’, which according to some researchers helped to ‘open the field’ for the policy construction (R2).

Although the draft of the public policy was written, the Municipal Council did not grant its approval. This can be explained by several forms of hidden power such as the distancing of the leaders, questioning the legitimacy of the process, the deliberate obstruction of the process, information transmission flaws, and actors avoiding being involved. While initially the meetings and workshops were well-attended (up to 40 attendees), the attendance declined gradually. By the end of the process, barely seven people were present. In addition, some leaders had distanced themselves from the process. This context of a weakening leadership, a reduced willingness to participate, and an increasing pressure to conclude the processes urged the local NGO and social organization leaders to take it to the next level (Municipal Council). Although they were aware of the lack of empowerment, understanding or open dialogue, the legitimacy derived from the successful start of entire process and from the mere authority of the local NGO itself allowed for this move. Second, several actors questioned the legitimacy of the process and condemned the pressure exerted by the local NGO to carry out the process without the effective participation of key actors: ‘There has been participation, yes, but it seems to me that it has not been very legitimate’ (R7). Another informant links the weakening of the process with the lack of coordination, the unclear division of roles and a resulting lack of responsibility with the actors involved (R2).

Others actively and consciously obstructed the process in several ways. For instance, several leaders expressed that they could not participate in the decision-making since they first had to consult their communities. Such a process usually takes a month, and did not happen:

A third example of conscious obstruction is provided by the actions of the irrigators, the group most affected by water policies. The FRIAC is an important actor because although its members do not manage all the water sources of Tiraque, they do have the biggest ones. They were not against the idea of the water policy as such, but they were afraid of its content, as it might affect their control over and management of water as well as land sources. Therefore, they took a series of actions that ended in the rejection of the policy. For example, identifying water sources is a key factor in developing water policy. Therefore, a technical advisor attempted to make an inventory of water sources in support of the policy. The irrigation organizations ‘have not endorsed this information, they have not accepted that it is valid information’ (R2).

Three forms of hidden power triggered the fears of the FRIAC for the new policy, urging them to act against it. The first form relates to the specific context of the gatherings. There is the issue of language, combining Spanish with a technical-legal jargon, and the deliberate choice by the Central Provincial to organize the meetings in their own venues, hence giving the irrigators a secondary position in the room. The second form relates to the misinterpretation of the ‘water planting’ concept and ‘water harvesting’ rights explained before. An informant explained ‘that the concept has been misinterpreted because harvesting water is about collecting water: when it rains, you harvest it, store it and that's it and it's for your consumption and they can't claim rainwater’ (R2). Therefore, irrigators were afraid that by ‘planting water’ other people would gain access and control of water resources. Third, historical conflicts with the municipality of Punata created the fear that this policy of conservation of water sources and recharge areas would be beneficial for Punata as well, a perspective that was not appealing, as detailed later.

As a result, the FRIAC preferred not to enact the policy. Other irrigation organizations and the Central Provincial, decided to reject it as well for political interests (see below). As an alternative, the FRIAC successfully requested a pilot case to be developed to show them the real impacts of the policy, which is currently being implemented through a Declaration of Protected Area in a community that supports the policy. In this way, irrigators have successfully managed to obstruct the approval of this water policy and negotiated an alternative, setting a new agenda as a true representation of their powerwith and powerto. This contestation was organized at claimed spaces at the FRIAC's own meetings. Then, the FRIAC's representatives met with different social leaders in closed spaces to express their concern, which was finally taken to the Municipal Council were the policy has been evaluated.

Actors used different spaces and levels of power to reach their interests: neighbouring municipal water interests, political interest related to presidential elections, NGOs' institutional interests, and the Mayor's office interests. The institutional regulation of water in Tiraque is a common interest among surrounding municipalities that benefit from Tiraque's water. They have a long history of municipal conflicts around water access with Tiraque. A particular conflict arose when the municipality of Punata tried to expand a water catchment in Tiraque (Totora Khocha) only to benefit its own communities (Figure 2). After several years of negotiations, an agreement was reached giving Punata 60% of the rights and 40% to Tiraque (Saldías, 2009). Since then, irrigators from Tiraque are very suspicious and cautious regarding any water policy that could also benefit Punata. Currently, ‘Punata and Tiraque do not get along, although water belongs to everybody. They understand that it does, but it does not suit their interests’ (R4). The surrounding municipalities' interests in the water policy are part of the hidden power since they are shaping the process outside of the public eye.

Evidently, political interests have influenced the process in Tiraque because during the final period of construction of public policy, there were many interests related to the presidential elections. During this period the priorities of the social organizations were influenced by political interests of Morales' Movimiento al Socialismo (MAS) campaign. On the one hand, social organization leaders were busy campaigning for MAS and did not spare time to lobby for new projects¹. On the other hand, leaders aimed to avoid any conflict that could jeopardize the political campaign such as losing the support of a highly influential sector such as the irrigators. The elections were followed by a period of great political unrest where social organizations did not want to trigger any conflict. Therefore, when the FRIAC opposed the water policy, social organizations and the Municipal Council rejected it to avoid any conflict.

Institutional interests also affected the process of construction of the policy. For example, the local NGO was interested in reaching the goals established by the international NGO and in its institutional planning. Usually, these interests generate pressure on the NGO's personnel since they must meet the established deadlines and outcomes of their funding organizations. However, in the process of policy construction they must also consider local dynamics and power relations, which can prolong the time needed to reach the expected goals. This clearly affected the local NGO promoting the water policy construction: ‘I believe that the NGO has also felt this pressure, because it has not been able to show many results of their project’ is mentioned by a country officer of the International NGO (R3). Hence, the interests of the international NGO influenced the interests of the local NGO and directed its course of action in the overall process. This is a representation of hidden power.

On their side, the Mayor's office is also interested in the policy because it could open doors to develop new sectors. For example, an interviewee mentioned that ‘this would lead to training in management of irrigation systems, opening new doors for production development’ (R1). This shows that although pursing the achievement of institutional interests may open possibilities to develop a policy, it may also generate unfavourable pressure in different actors.

Invisible power concerns social norms and values that shape what is accepted as normal by society (Brouwer et al., 2016). Therefore, invisible power, represented by usos y costumbres in Tiraque, shape the process of decision-making by defining spaces, protocols, leaders, and participation. Usos y costumbres were respected by the NGO and the municipality, by the Mayor and the Municipal Council. It is the usos y costumbres that can make decisions accepted as legitimate when they are made by authorities. Usos y costumbres were used to establish the NGO's strategy to set the agenda, the meetings at the Central Provincial, only leaders’ participation and top-down flow of information.

The NGO abided by the local usos y costumbres to define its strategy to include the construction of the public policy in the agendas of the municipality and social organizations. For example, the Central Provincial asked to have the meetings and workshops as part of their own meetings. Likewise, at the beginning of the process the NGO identified the need to engage strategically and separately with irrigator's organizations to avoid conflict. However, the Central Provincial asked to have general joint meetings, a request that was accepted by the NGO following the authority level:

Although according to the usos y costumbres, the information shared and generated at the Central Provincial meeting should flow down to the grassroots, this transmission of information was limited. A municipal authority mentioned that ‘it is not possible to reach the totality of the grassroots and there may be a dissatisfied minority, but they abide by what has been decided’ (R6). This shows that communities will usually abide to the decisions made even if they do not agree with them. Therefore, the NGO initially tried to show in a workshop how some usos y costumbres can be changed to protect water sources and water recharge areas. In fact, to implement this water policy, some social structures would have needed to be adapted. Although this case shows how actors challenged the water policy through their powerwith and powerto, it also shows how socio-political structures affect the process of decision-making and how certain groups will abide by the decisions made by their leaders. Moreover, as mentioned, some communities were afraid that their usos y costumbres were challenged by the new policy, opening space to the State to access and control their resources.

This case study illustrates that the PowerCube can be fully implemented for an integrated analysis of how power interrelates with actors' interests, knowledge, and socio-political structures in the (co)creation of a water policy. These interrelations define the course and result of the process and thus should be considered in future implementations of the PowerCube to enrich the analysis. Furthermore, through the implementation of this tool and Hunjan & Pettit (2011) guiding questions, the dynamics and characteristics of hidden and invisible power became observable, evidencing that the Power Cube is a promising tool for the analysis of power relations to unveil insights for the (co)creation of more inclusive, legitimate, and empowering natural resources and water policies.

Visible power is an exercise by decision-makers, in this case the Central Provincial. Actors who exercise visible power usually make decisions in closed spaces and, most importantly, decide who can participate and who is excluded. In this case, some vulnerable groups were left out of the process and others, such as women groups, were part of the process, but their participation remained low. This indicates that attendance does not guarantee participation. Moreover, it shows that women's knowledge is still not considered as valuable. Similarly, Seemann (2016) found in one community of Tiraque that in the Water Users Registries, only 56 of 200 entries belonged to women even though women are involved in the daily chores of water management. By being left out of these registries, women have limited access to meetings. Therefore, women continue to be invisible in water-related decision-making arenas.

Along the same lines, access to timely and adequate information was a key for the exclusion of lower levels governance actors. As long as communities do not have access to timely and adequate information (i.e., due to language barriers), they cannot achieve ‘participation and autonomy’ (Chávez & Rojas, 2011, p. 174). If they are not informed, they will not be able to build their own speech and make their voices heard. A similar case was observed in the process of the Commission on the Legal Empowerment of the Poor from Peru. This process was criticized by excluding vulnerable groups such as ‘the poor, women, indigenous peoples, the landless and ethnic minorities’ (Seemann, 2016, p. 194).

Although certain actors exercise visible power, the FRAIC managed to challenge the process through their powerwith and powerto and stopped the approval of the process, setting a new agenda by asking for a pilot project. Indeed, by following the strategies and tactics of various actors, it is possible to see how the direction of a policy is changing (Brock et al., 2001). This was seen in studies carried out by Geng (2016) and by De la Cruz & Dessein (2021) in Peru, where groups of villagers managed to challenge the decision-making power through interactive networks of actors. Such ‘intertwining relations’ (De La Cruz & Dessein, 2021) demonstrate that policy-making can take the form of bricolage – where actors participate throughout the whole process as ‘doers’ – , or even of sabotage – where other actors oppose and sabotage the process as ‘stoppers’ (Paquet, 2001, p. 14).

Moreover, throughout the process there were closed, invited, and claimed spaces. However, spaces interact with different forms and levels of power, changing from one type of space to another (Gaventa, 2006). For example, participation was still limited at invited spaces because hidden forms of power defined what issues and whose voices were valid. When the FRIAC contested the public policy and managed to set a new agenda, their contestation was organized in claimed spaces (the FRIACS's meetings) and taken to higher levels of decision-making in closed spaces (meetings with social leaders). Similar observations were made in Asian cities and in Rwanda, where vulnerable people excluded from closed and invited spaces of participation mobilized to create claimed spaces for contestation (Patel et al., 2016; Nikuze et al., 2020). This evidence indicates that in policy-making dynamics, different spaces are used or created by actors at different levels to achieve their goals. Thus, levels of governance also interact with forms and spaces of power in different and varying ways.

Invisible power can be difficult to explore since it concerns social norms and values related to socio-political structures which require longer periods of time in the field to be able to identify and study them. Because of the socio-political unrest after the presidential elections of 2019, it was difficult to reach leaders of social organizations to achieve deeper interviews after the final rejection of the policy. Moreover, in Bolivia, usos y costumbres are important aspects in rural communities and are guaranteed by the Political Constitution of the State through Article 374 (Bolivia, 2009). A threat to local usos y costumbres can spark conflict. Saldías (2009) in a study on Tiraque and Punata found that irrigation associations of high-altitude communities negotiate water control based on their territory and their usos y costumbres. Therefore, to create changes through policy, either the policy must conform with local water usos y costumbres or there must be an adaptation or transformation of them. In this line, the rejection of the policy and the requirement to establish pilot projects can be seen in the long term as an opportunity because these interventions could provoke generational and lasting changes to invisible power, required for sustainable water management. Therefore, we recommend promoting small-scale interventions to provoke changes in usos y costumbres, and to generate a sense of ownership and earn the required legitimacy for the development of water resources policies.

This case study shows that the attempts to implement a highly participatory process in the (co)creation of public policies eventually conformed with the traditional linear, silo thinking, and top-down schemes in which vulnerable groups continue to be excluded. To create a truly participatory process, a change must be achieved that allows actors to be given a voice in an effective participation that takes them from ‘access, to presence, to influence’ (Gaventa, 2006, p. 24). Therefore, more collaborative governance approaches should be developed that conform with Latin–American contexts.

The design of a water policy in Tiraque presents evidence that the (co)creation of such a policy is not a linear process because of the complex intertwining of power relations from the beginning until the end. Therefore, to foster more inclusive, equitable, empowering, and legitimate policies, it is crucial to analyse power relations throughout the whole process.

This paper shows that the PowerCube (Gaventa, 2005) is a very useful approach to unveil complex power dynamics and to have a close understanding of the different forms, levels, and spaces of power in natural resources and water issues. Moreover, it demonstrates that it can be expanded to study how power interrelates with actors' interests, knowledge, and socio-political structures. It is highly recommended to include these interrelations in future implementations of the PowerCube.

Our study shows that social organizations in the municipality of Tiraque exercise more visible power than the Mayor and the Municipal Council. Although the meetings organized for the construction of the public policy are expected to be invited spaces, they became closed spaces because only specific actors were invited. We identified four forms of exclusion: (1) rotating and supplanting leadership, (2) a culture of gendered decision-making, (3) avoiding conflict by not inviting, and (4) flawed transmission of information. Through these forms of exclusion vulnerable and marginalized groups were excluded. Women groups, for example, were present but their effective and meaningful participation was not guaranteed. Moreover, language barriers hindered the participation of different actors, specially from the grassroots.

The study shows that both actors in favour and against the public policy exercised hidden power by setting the agenda. On the one hand, the NGO managed to get the construction of the public policy on the agenda of social organizations and the municipality. To do so, the NGO abided by the local usos y costumbres regarding decision-making and developed a series of workshops and meetings at invited spaces for participation. On the other hand, the FRIAC was afraid of the possible impact of the policy and managed to have the policy rejected, asking for the development of pilot cases; therefore, setting a new agenda. Their contestation was organized in a claimed space. This shows that intertwined relations in processes of bricolage and sabotage shape the policy-making process.

Usos y costumbres were followed by the NGO and defined the protocol for the construction of the policy. However, usos y costumbres in Tiraque are also closely linked to invisible power because they define what is accepted as normal by society regarding decision-making. Even if some people or families do not agree with the decisions made, they usually will abide by them, seeing them as mandatory.

Different forms, levels, and spaces of power were used by actors to reach their interests. Many actors had political interests at the time, either for the presidential elections or to open new spaces for water management projects. The local NGO had institutional interests, forcing it to achieve the expected results set by their funding organization. Interests of other actors were also present outside of the public eye, such as the interest of the neighbouring municipality that benefits from the water from Tiraque.

Although the study focused on Bolivia for an in-depth case study, the findings are globally relevant for collaborative approaches in the (co)creation of natural resources and water policies. Hence, we suggest developing a power analysis before and during the (co)creation of a public policy to achieve true empowerment. Otherwise, vulnerable and marginalized groups might be excluded which might generate obstacles to achieve the expected result.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data cannot be made publicly available; readers should contact the corresponding author for details. The datasets generated during and/or analyzed during the current study are not publicly available due to the sensitivity of data and to keep the anonymity of the interviewees but are available from the corresponding author on reasonable request.

We express our gratitude to the different actors involved and to VLIR-UOS for funding this research.