The politics of coproduction during Latin America’s ‘Pink Tide’: Water, housing, and waste in comparative perspective

Coproduction brings together a diverse range of state and non-state actors to create and deliver public services. These processes, which occur across the Global South, have been widely studied. However, insufﬁcient critical attention has been paid to their politics. We address this gap in the literature by ana- lysing the politics of coproduction in Latin America during the ‘pink tide’ of the early twenty-ﬁrst century. Drawing on original qualitative research, this article explores the coproduction of three distinct public services—water, housing, and waste—in three countries where left-leaning presidents and governments were elected into ofﬁce—Ecuador, Bolivia, and Uruguay. We argue that coproduction is intrinsically polit- ical in these three cases; that is, the ‘political’ is internal to, and inherent in, coproduction. Our comparative analysis centres on two political dimensions—subject-making and collective autonomy—and shows that tensions around these two issues were central to coproduction in Ecuador, Bolivia, and Uruguay . The article demonstrates the importance of not treating politics as a mere contextual concern in coproduction analysis and shows that coproduction has the capacity to reshape political relationships and subjectivities. Taking politics into account is essential to understanding the dynamics and potential of coproduc- tion in the Global South. Our comparative analysis also provides new insights into Latin American politics, especially concerning ’pink tide’ governments and the provisioning of public services. (cid:1) 2022 The Authors. Published by Elsevier Ltd. ThisisanopenaccessarticleundertheCCBYlicense(http://


Introduction
Latin America experienced a seismic political shift in the opening decades of the new millennium as a wave of left-leaning politicians and parties swept across the region and reconfigured the political landscape. This so-called 'pink tide' involved a diverse range of political actors and projects, from reformist governments in Brazil and Chile to more radical regimes in Ecuador and Venezuela. Taking advantage of favourable global economic conditions, these governments set about reforming or rejecting neoliberal policies and practices, and positioning the state as a more central actor in the process of capitalist development. One important part of this process was a commitment to improving public services. Outcomes were mixed, both at local and national levels; but overall, the coverage and quality of services increased, following decades of decline under structural adjustment and neoliberal reforms. Public utilities and state agencies were important actors in this process. However, organized groups of citizens, including families, communities, social movements, and cooperatives, also made substantial contributions, and public services were often 'coproduced' between state and non-state actors (Joshi & Moore, 2004).
In this article, we consider the politics of these processes by focusing on the coproduction of three different public serviceswater, housing, and waste-in countries where left-leaning presidents were elected into office-Ecuador (Rafael Correa), Bolivia (Evo Morales) and Uruguay (Tabaré Vásquez and José Mujica)-in the early twenty-first century. Our comparative analysis and ethnographic methods make a unique contribution to the global coproduction scholarship, and coproduction research in Latin America, in particular. We show that coproduction re-shapes political relationships and subjectivities and how political struggles emerge around these issues. In employing an ethnographic approach we also shine a light on the lived experiences of https://doi.org/10.1016/j.worlddev.2022.105930 0305-750X/Ó 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). coproduction, providing new insights into the ways in which organized groups of citizens navigate and contest the process.
Our comparative analysis builds on the existing scholarship of coproduction politics in the Global South (see, for example, Mitlin, 2008Mitlin, , 2018McMillan et al., 2014;Goodwin, 2019). These studies show that coproduction takes place within existing political relations and structures and has the capacity to reconfigure them. Our aim in this article is to go a step further, to argue that coproduction is intrinsically political; that is, the 'political' is internal to, and inherent in, coproduction. Our understanding of the 'political' is not limited to formal political practices and institutions but is more broadly rooted in a wider set of unequal social relations characterized by antagonism (Mouffe, 2005(Mouffe, , 2018. Coproduction creates opportunities for individuals and collectives to reconfigure power relations, strengthen their rights, and develop alternative forms of citizenship (Mitlin, 2008(Mitlin, , 2018. However, it is also a site of struggle and antagonism, as conflicts emerge over the diverse discourses, practices, and interests involved in the process. Political struggles occur because of the different assumptions, logics, and objectives that various actors bring to coproduction and the confrontations, strong-arming, and compromises that inevitably follow. Integrating such diverse and differentiated actors into processes of public service delivery is unavoidably political. We demonstrate this by focusing on two important political dimensions of coproduction: subject-making and collective autonomy (Mitlin, 2008;Goodwin, 2019). We draw on the concept of 'translation' (Dinerstein, 2017(Dinerstein, , 2020 to show how these issues are mediated by the conversion of the demands of the organized citizens involved in coproduction into laws, policies, and constitutions. By analysing these issues, this article also contributes to debates on Latin American politics, especially concerning 'pink tide' governments and the provisioning of public services. The remainder of this article is structured as follows. The next section explains our conceptual approach to coproduction and subject-making and collective autonomy. We then discuss our data and methods. The subsequent sections analyse the three cases-Ecuador (water), Bolivia (housing), Uruguay (waste)-individually. Finally we consider them comparatively, identifying similarities and differences, and signposting future avenues of coproduction research.

Conceptualising coproduction politics in the Global South
Our use of coproduction follows the reading of the term as proposed by Joshi and Moore (2004: 40), which sees public services being produced through long-term interactions between state agencies and organized citizens, with all parties making significant contributions to the process. 1 We understand 'state agencies' to include state bureaucracies, public utilities and national and local governments, and 'organized citizens' to capture a diverse range of social collectives, including families, communities, associations, cooperatives, and social movements. We broaden the Joshi and Moore (2004) definition of coproduction to include international development agencies, non-governmental organizations and multilateral institutions as these diverse actors perform important roles in coproduction in the Global South (Goodwin, 2019). The roles and contributions of the various actors involved in coproduction are diverse and fluid, and the process therefore exhibits considerable variation, across both space and time. However, the active involvement of organized citizens in the delivery of public services is a constant feature of coproduction, and one of its defining characteristics.
While the Joshi and Moore (2004) reading of coproduction is a useful starting point, it is insufficient to capture the politics of coproduction as it largely sees coproduction as a technical process of public service delivery (Ahlers et al, 2014;Goodwin, 2019). We overcome this by drawing on insights from scholars who have explored the politics of coproduction in the Global South. Mitlin (2008), for example, shows how urban communities and social movements have used coproduction as a tool to exert greater influence over state policies and practices, build organizational capacity, and take collective control of infrastructure and services (see also Ahlers et al., 2014;McMillan et al., 2014;Mitlin, 2018;Mitlin & Bartlett, 2018;Chatterjee and Kundu, 2020). She sees coproduction as having instrumental value for these groups, as a mechanism to challenge power relations and increase political influence. Importantly, Mitlin (2008) argues that the organized citizens involved in coproduction generally seek to manage infrastructure and services over the long run rather than developing them for transfer to the public or private sector. Social collectives can therefore build collective autonomous control of infrastructure and services through coproduction. Mitlin acknowledges, however, that this is not always the case, and organized citizens involved in coproduction often lack the collective unity and external support to build collective autonomy, which results in low-income families and communities being marginalized or excluded from public service provisioning, a particularly important feature of our Uruguayan case study (see also Meagher, 2013;Watson, 2014;Rusca, 2015;Adams & Boateng, 2018;Rosaldo, 2022). Mitlin (2008) hints at the political subjectivities that emerge through these processes, arguing that coproduction creates space for organized citizens to develop relations and practices that strengthen collective consciousness and identities. However, by coming into closer contact with politicians, bureaucrats and planners, they also run the risk of becoming 'agents of the state' (Mitlin 2008: 357). Goodwin (2019) further explores this tendency of coproduction to promote engagement with and autonomy from the state, arguing that political tensions and opportunities stem from this process, which reshapes political relations and subjectivities.
In this article, we build on these insights to explore subjectmaking and collective autonomy in the coproduction of public services in Ecuador, Bolivia and Uruguay during the 'pink tide' of the early twenty-first century. Our comparative ethnographic approach enables us to offer a fine-grained analysis of these issues and provide important new insights into coproduction. We show that tensions in processes of subject-making and collective autonomy are present in all three cases, highlighting the intrinsically political nature of coproduction.
Our approach to subject-making takes inspiration from Foucault (1982Foucault ( , 2008 who sees this as a fundamental task of modern government. However, we place greater emphasis on political subjectivities based on collective relations and practices (see also Mitlin, 2008). The state is a key actor in the process of subjectmaking but citizens have the power to resist and form political subjectivities that challenge the rationality and authority of the state. Indeed, Rancière (2010) proposes that political subjects emerge through the very process of challenging and contesting the rationale by which political decisions are made. Development agencies, multilateral institutions and non-governmental organisations also perform important roles in this contested process. Hence, the terrain of subject-making extends beyond the state, which underscores our understanding of the 'political' as rooted in a wider set of unequal social relations and not limited to formal political practices and institutions (Mouffe, 2005(Mouffe, , 2018. New political subjectivities can emerge when citizens take greater collective autonomous control of infrastructure and services, disrupting taken for granted assumptions about who can act in political processes (Rancière, 2010). In this way, subjectmaking and collective autonomy are both central and intertwined aspects of coproduction politics. Political struggles emerge as social collectives attempt to protect or strengthen their collective autonomy while negotiating with the other actors involved in coproduction. Coproduction therefore demands a relational understanding of autonomy (Dinerstein, 2015). Collective autonomy implies social collectives have some control over aspects of service delivery, organisational norms and structures, and the use of labour and resources. But this is conditioned through their diverse interactions with state and non-state actors and the wider social, political and economic context. Hence, we understand collective autonomy as a relational process rather than a binary condition (i.e., autonomous or non-autonomous). The tension between engagement with and autonomy from the state is more acute for some collectives than others and relations between state agencies and organized citizens vary across space and time. Political tensions can emerge when the state reduces or increases its contributions to coproduction. Hence, the solution to tensions related to collective autonomy is not simply to reduce state support and control. What really matters is the form of state regulation and support and the space that is open for social collectives to influence decision-making and protect or expand their autonomy.
The concept of 'translation' (Dinerstein, 2017(Dinerstein, , 2020) further helps us understand political struggles connected to subjectmaking and collective autonomy in coproduction processes, especially when the state becomes a more dominant actor, which was the case during Latin America's 'pink tide'. Translation refers to the incorporation of the demands, proposals and practices of grassroots organizations and social movements into the institutional language of law and policy. Dinerstein (2017Dinerstein ( , 2020 uses the concept to analyse a wide range of struggles in Latin America. We restrict its use in this article to understanding the tensions that emerge when the demands and proposals of the social collectives involved in coproduction are incorporated into laws, policies and constitutions. The problem of translation, Dinerstein (2020) explains, is rooted in the intrinsic need for capitalist states to support capitalist relations, practices and processes, on the one hand, and ensure laws and policies protect these fundamental objectives, on the other. Elements of social collective demands and proposals that are not legible from the epistemological and ontological perspective of the capitalist state are excluded from the translation process or deradicalized and reconfigured to support capitalist relations and processes (see also Scott, 1998). Bureaucratization presents further challenges as the demands and initiatives of social collectives are integrated into hierarchical, standardized state institutions.
Translation is, among other things, a struggle over political subjectivity and collective autonomy (Dinerstein, 2020). Bringing about changes to laws, policies and constitutions offers organised citizens the prospect of strengthening their rights and interests but runs the risk of them losing autonomy and becoming integrated into state institutions. Furthermore, when negotiating with the state, members of social collectives are forced to articulate their demands and practices, which involves internal debates over their meaning. Hence, translation not only involves struggles with the state but also within and between the social collectives pushing for political change. Translation was a particularly important issue in Latin America during the 'pink tide', as left-leaning governments were expected to respond to the demands of the grassroots organizations and social movements that helped put them into office (Dinerstein, 2020). Coproduction was a key domain for these struggles, as our comparative analysis shows.

Data and methods -An ethnographic and comparative approach
Our analysis of the three selected cases-Ecuador (water), Bolivia (housing), Uruguay (waste)-draws on qualitative research, including ethnographic fieldwork, conducted independently in the three countries. Ethnographic fieldwork involves long-term participant-observation within a research location, be this a community, an organisation, or other spaces appropriate to the research topic. The researcher both observes daily routines and events, and participates in on-going activities within the field site in order to gain a deep and broad understanding of the research topic. It is an inductive and open-ended methodology (Howell, 2018). Participant-observation has multiple advantages in a study on coproduction. First, it ensures data is gathered over a significant period of time, enabling the researcher to follow a process as it unfolds; secondly, it does not rely on verbal data only-what people say-but can capture what people do, which allows for verification of received data and therefore rigour of evidence; thirdly, key questions are generated by the research participants, not the researcher, ensuring the analysis is firmly grounded in data, rather than in a pre-conceived hypothesis. The depth of ethnographic engagement varied, with research in Bolivia and Uruguay involving greater immersion in field sites than in Ecuador. In each case, however, extended periods were spent in the field, close attention was paid to everyday practices, extensive fieldnotes were written, and an inductive methodology was followed. The research conducted in each country is discussed in more detail below.
In Ecuador, multi-sited fieldwork was undertaken in several visits between 2015 and 2019, amounting to eight months in the country. This connected to preliminary research conducted in Ecuador between 2010 and 2011. A total of 58 semi-structured interviews and extended discussions were undertaken with representatives of water associations, social movements, state bureaucracies, political parties, local governments, and nongovernmental organisations. Most interviews took place in the highland region but interviews were also conducted in the coastal and subtropical lowlands to understand geographical variation in water services and struggles. The bulk of time in the field was spent in the highland provinces of Pichincha, Chimborazo, Cañar, and Azuay. Interviews were complemented with attendance at public and community meetings and street protests, informal conversations, observation, and field notes. Legal archives at the National Assembly in Quito and national and local newspaper archives at municipal libraries in Quito and Riobamba were also consulted to triangulate information derived through interviews, conversations and observations and gain a deeper understanding of water politics and services.
In Bolivia, interviews and participant observation were conducted over a nine-month period in 2018 with the beneficiaries of state social housing in a peri-urban area in the zona sur (the south zone) of the city of Cochabamba, where the Bolivian Housing Agency builds thousands of homes every year. Separate fieldwork took place over three 1-2-month periods in 2016, 2017 and 2019 in the rural municipality of Charazani in the department of La Paz, which received hundreds of social houses during the Morales government. Additionally, data was collected during three days shadowing housing agency employees as they carried out their work in Cochabamba, as well as in multiple semi-structured interviews with these employees and more senior staff within the agency. A material analysis of the agency's communication literature complemented the ethnographic focus.
In Uruguay, ethnographic fieldwork was conducted with Montevidean waste-pickers over several visits from 2010 to 2020. This involved research with both Uruguay's oldest waste-picker cooperative and the largest. The transition of the latter cooperative from one of informal workers at Montevideo's landfill to that of employed labourers at a recycling plant built as part of the national Ley de Envases (Packaging Law) was also followed. Qualitative research methods involved participant observation at various waste-picking sites and government offices, over 60 interviews with wastepickers, government officials, relevant NGO workers, and trade unionists, archival research in the library of the Junta Departamental de Montevideo (Montevidean Regional Council), and community art intervention.
We selected these cases for this comparative study for three main reasons. First, the provisioning of the public service in each case is consistent with the conceptualisation of coproduction outlined in the previous section, and therefore involves significant contributions from state agencies and organized citizens. Second, each case involves the coproduction of a different public service. Showing that similar political tensions were present in the coproduction of three distinct services-water, housing and waste -enabled us to make stronger claims about the intrinsically political nature of coproduction. 2 Third, left-leaning governments were elected into office in the three countries selected for the study. While the left-right political spectrum is ill-equipped to deal with the complexities of Latin American politics, these governments-Rafael Correa (Ecuador), Evo Morales (Bolivia) and Tabaré Vásquez and José Mujica (Uruguay) are widely considered to be part of the 'pink tide' of the early twenty-first century (see, for example, Ellner, 2019;Garcia Fernandez, 2021). Focusing our analysis on 'pink tide' governments enabled us to explore the politics of coproduction in countries that experienced similar, if distinct, processes of political change during the same historical period. This is a particularly interesting time to explore coproduction politics as the state became a more dominant actor in coproduction, while social collectives brought new hopes and ideas to the process.
Viewing coproduction through an anthropological lens, the data gathered on the three cases was analysed using content analysis and narrative analysis to systematically evaluate the field notes and interview transcripts. No software was used for this analysis, instead fieldnotes were organized and analysed by all authors manually. This was done individually then collectively to draw out common themes and identify similarities and differences across the data. Our comparative approach is rooted in anthropological evaluation of lateral and disjunctive (Lazar, 2012) comparison as a means to generate theoretical questions and insights, deepening scholarly conversations on particular regions or themes (Candea, 2018), rather than primarily deriving clear conclusions about cause and effect through comparing and contrasting cases. In comparing three diverse cases our aim is both to offer new empirical and theoretical insights into the politics of coproduction in the Global South and shine new light on the politics of public service provisioning in Latin America. The act of comparison focused on the logic and process of coproduction from the conceptual perspective outlined in the previous section. This comparative approach also enabled us to distil elements of coproduction that can be found across sites and services and make stronger claims about the politics of coproduction. Our claims are most relevant to the Latin American context but the political issues we discuss are reported elsewhere in the Global South , suggesting tensions related to political subjectivities and collective autonomy are common in coproduction processes across the world. The following sections explore these diverse processes in greater depth.

Ecuador: The coproduction of water services during the revolución ciudadana
Community water associations are the protagonists in the coproduction of water services in Ecuador (Goodwin, 2019(Goodwin, , 2021. This highly diverse group of organisations started to perform a progressively larger role in the delivery of water services in Ecuador in the 1960s and 1970s, especially in rural and periurban areas in the Andean region. The water that they deliver is used for various activities, including drinking, cleaning, washing, cooking, and irrigation. 3 Water associations vary in size, ranging from less than 100 to over 10,000 members. Most members come from low-income households engaged in small-scale agriculture and informal activities; however, some associations include middle-class members with formal employment or businesses as well as medium or large-scale farmers. The racial and ethnic composition of water associations varies across the highlands. Water associations require external support to construct water systems and maintain and develop them over the long run, which involves them interacting with various state and non-state actors to coproduce water services (as in Joshi & Moore, 2004; see also, Ahlers et al., 2014;Moretto et al., 2018;Pareja Pineda et al., 2022). The general pattern of coproduction in Ecuador, which was established in the 1960s and 1970s and remains largely intact today, involves the families and communities integrated into water associations contributing labour, finance and knowledge to the construction and maintenance of water systems, and state bureaucracies, local governments, international development agencies, and non-governmental organizations supplying materials, finance, and technical knowledge (Goodwin, 2019). Social movements, particularly local and national indigenous movements, contribute to coproduction by strengthening the mobilizing and negotiating capacity of water associations, even if not all water associations are affiliated to social movements and integrated into their organisational structures and decision-making processes. The composition of state and non-state actors involved in coproduction varies from case to case; however, water associations are constantly present and take responsibility for the daily delivery of water services.
In the Andean region, the collective labour practice known as the minga is a central component of coproduced water services (Goodwin, 2019(Goodwin, , 2021; see also Boelens & Doornbos, 2001;Armijos, 2013;Manosalvas, 2021). The minga, which is used both to construct and maintain water infrastructure, is based on principles of solidarity and reciprocity, and involves community water association members working collectively on an unwaged basis, although food, drink, and entertainment are often provided. Through these physical endeavours, collective identities and memories have been forged, which have generated a strong sense of collective ownership and control of infrastructure and services . Participation in community meetings and taking responsibility for the daily delivery of water services have added to this sense of collective control and contributed towards the formation of political subjectivities based on collective autonomous practices and relations. 3 Potable and irrigation water are usually managed by separate associations in Ecuador. In everyday practice, however, the line between the management of potable and irrigation water is often blurred as many rural and peri-urban households use potable water to irrigate small plots of land. The organisational separation of potable and irrigation water management is rejected by some highland communities, which manage water holistically through existing communal institutions. implemented a state-directed capitalist modernization project-the so-called revolución ciudadana (citizen's revolution). The revolución ciudadana involved rewriting the constitution, bolstering the state, investing in infrastructure, and expanding public services (see, for example, Becker, 2012;Martínez Novo, 2014). Oil revenues, tax reforms, and overseas loans underpinned the expansion of the state bureaucracy and the acceleration of public spending. The new constitution incorporated several innovative and potentially progressive dimensions, including declaring Ecuador a plurinational state and establishing buen vivir as a guiding principle for state policies and practices. However, the revolución ciudadana's progressive 'left hand' was accompanied by its repressive 'right hand' (Bourdieu, 1999; see also, Wacquant, 2009). This included centralising decision making within state agencies, increasing surveillance, and threatening, abusing, and jailing opponents of the president and government, including indigenous leaders, academics, and journalists. The revolución ciudadana thus demanded a certain type of political subject: a loyal citizen committed to the 'revolution'-i.e., state-directed capitalist modernization-who was willing to mobilize to defend but not critique it (Conaghan, 2015). A central pillar of the revolución ciudadana was the construction of a new water regime, which implied integrating community water associations into a state-centric, highly-regulated form of coproduction (Goodwin, 2019). Following decades of neglect, the prospect of greater state support and clearer legal recognition was welcomed by many water associations. However, the threat posed to autonomy by heightened state regulation and control was widely shared and most water associations and social movements contested the government's proposals. Exploring this struggle shines a light on the political tensions that can emerge when coproduction is scaled up and incorporated into constitutions, laws and policies. This, as Mitlin and Bartlett (2018) suggest, is an important but underexplored issue in coproduction research and analysis. The following sections focus on tensions related to subject-making and collective autonomy, using the translation concept to further our understanding of these issues in the context of coproduction.

Confronting translation through coproduction
The 2008 constitution provided the basis for the reconfiguration of water coproduction in Ecuador. It incorporates several historic water association and social movement demands, including recognising water as a human right, proscribing privatization, and entrusting the delivery of water services to public and community organizations. Reflecting the long history of coproduction and the demands of water associations, it also explicitly promotes statecommunity alliances to deliver and improve water services, thereby providing a constitutional basis for coproduced water services. The constitution declared that a new water law should be introduced within twelve months. However, the Correa government's initial legislative proposals were opposed by water associations and social movements, and the new law was not introduced until 2014, following a nationwide consultation, which was organized after protestors surrounded the National Assembly in May 2010 and demanded the rejection of the government's legislative proposal (Registro Oficial, 2014;Boelens et al., 2015;Goodwin 2019Goodwin , 2021 During this struggle, water associations and social movements confronted translation through coproduction in various ways. Here, we highlight three. First, the legal and bureaucratic framework established by the Correa government granted clear legal status to water associations and recognised their administrative and financial autonomy (Registro Oficial, 2014. Yet it also threatened to undermine their substantive autonomy by requiring them to conform to the demands of the law and related policies. Water associations were expected to fulfil numerous requirements to secure formal legal status, including adopting particular organisational forms, which required the blessing of state water bureaucracies. The legal code therefore denied the individual histories and identities of water associations and attempted to incorporate them into a single standardised bureaucratic framework. This threatened the vernacular structures and practices that water associations had developed through decades of organizing and attempted to reshape their autonomous political relations and subjectivities. The challenge to autonomy was widely felt. For instance, the former vice-president of Interjuntas-Chimborazo, a network of irrigation water associations in the central highland province of Chimborazo, believed the government wanted 'to be inside the organizations to control them', while another water activist said the government was acting 'like someone coming into your house and telling you how to live'. 5 The threat posed to community water association autonomy through the formalisation of coproduction in Ecuador reveals the tensions that can emerge when the demands of the organized citizens involved in coproduction for stronger legal recognition are adopted and translated into law (see also Llano-Arias, 2015;Al'Afghani et al., 2019).
Second, the new legal and bureaucratic framework incorporated one of the central demands of water associations and indigenous movements-the establishment of a Consejo Plurinacional del Agua (Plurinational Water Council)-but in a diluted and deradicalized form (Registro Oficial, 2014. The council envisioned by the indigenous movement CONAIE and the indigenous political party Pachakutik, set out in a proposal submitted to the National Assembly at the start of the legislative process in 2009, was to comprise representatives of the state, social movements, and water associations and become the national regulatory water authority. The proposal implied a more radical form of coproduction, which provided social actors involved in the process with genuine influence over decision making at multiple scales (see also Mitlin & Bartlett, 2018). However, once translated into law, the plurinational water council was relegated to a secondary role and all decisionmaking power was vested in state bureaucracies-SENAGUA and ARCA-which solely comprised state officials (see also Moretto et al., 2018). The Correa government's half-hearted commitment to the plurinational water council was reflected in its limited interest in creating it. Despite pressure from water associations and non-government organisations, the plurinational council has still not been established, over six years since the law was first introduced (FRH, 2020). It was through the translation of the plurinational water council into law that the 'utopian aspirations' of water associations and indigenous movements were perhaps most clearly frustrated (Dinerstein, 2015(Dinerstein, , 2017. The opportunity to establish a form of coproduction that provided genuine space for water associations and social movements to influence decisionmaking over the use and control of water at multiple scales had been lost. Third, one of the central historic water association and social movements demands-the prohibition of the privatization of water sources and services-was incorporated into the constitution (Harris & Roa-Garcia, 2013;Hoogesteger et al., 2017). Technically, this protected coproduced water services from being replaced by private operators and created space for the expansion of coproduction. Yet, as constitutional declarations were converted into law, the door was gradually opened to private sector involvement in water services. 6 Transnational private water companies lobbied hard to ensure the new water regime left space for them to generate profits through the delivery of water services. The new water law emphatically restated the prohibition of the privatization of water, but permitted private sector involvement in 'exceptional' circumstances (Registro Oficial, 2014. The circumstances detailed in the law and related legislation were sufficiently commodious to allow private firms to participate in the delivery of potable water and sanitation services. If, for example, community water associations were deemed by SENAGUA and ARCA to offer inadequate services, private operators could be allowed to deliver water services in their place. Many water associations feared that this legal mechanism would be used to transfer community water systems to the private sector or as a threat to contain political activism and discipline water associations (see also Boelens et al., 2015;Conaghan, 2015). This undermined trust in the Correa government and introduced new tensions into coproduction.

Negotiating translation through coproduction
Translation is a process, not a one-off event (Dinerstein, 2017(Dinerstein, , 2020. The translation of coproduction into law, which involved responding to historical demands of water associations and social movements, passed through various stages, starting with the writing of the new constitution and continuing during and after the legislative process, which lasted several years. Water associations and social movements influenced the writing of the new constitution in the Constituent Assembly and the legislative process in the National Assembly. The proposal developed by the indigenous movement, CONAIE, though not endorsed by all water associations (Hoogesteger, 2014), influenced legislative debates. 7 Meanwhile, mass mobilisations disrupted the government's plans and forced a national consultation on the new water law. The consultation, the first of its kind in Ecuador, was another important moment of translation, as the government seized control of the process and closed down spaces for genuine debate (Isch López & Zambrano, 2017). Consultation meetings were generally well attended across the country. 8 Water associations and social movements voiced their concerns and made concrete demands. However, with the parameters of the consultation narrowly defined by the Correa government before the process began, they were only able to make modest changes to the legislation. The government 'played with our hopes and dreams', one leading water activist lamented. 9 Although the consultation ultimately undermined support for Correa, it helped his governments contain and demobilize mass protests against the water law. Nonetheless, water associations and indigenous movements continued to resist the legislation after it was introduced. In some cases, water associations deepened their autonomous practices and relations and rejected the new water regime, including rewriting their internal statutes in opposition to the new law (Goodwin, 2021). Political subjects are thus not only shaped through laws and policies but in opposition to them. Water associations and social movements also negotiated with the state to influence the implementation of the law. Important concessions were won in this struggle, including the recognition of indigenous forms of water organisation, which eased tensions over the imposition of uniform organizational structures (Manosalvas, 2021). The Foro de los Recursos Hídricos, a national non-governmental organization, also played an important role in mediating between water associations and state agencies, showing how relations developed through coproduction can help social collectives navigate translation (FRH, 2020; see also Hoogesteger, 2014Hoogesteger, , 2016Dupuits, 2019).
Local political alliances also helped some water associations manage translation and deepen coproduction (see also Fieuw & Mitlin, 2018). For example, in the southern highland province of Cañar, CENAGRAP, a coalition of potable water associations, worked closely with the local government to expand and improve water services before, during and after the introduction of the new water law. 10 Ties between members of water associations and the indigenous political party Pachakutik, which run the local government, eased tensions over the new law and provided a basis to strengthen coproduction. Elsewhere, water associations aligned with Alianza País at the local level to gain access to public funds and strengthen coproduced water services. This shows that translation operates at multiple scales and tensions at the national level can be assuaged at the local level. In other cases, however, local politics complicated translation and undermined coproduction. For example, after the president of Interjuntas-Chimborazo was elected as a councillor for Alianza País , his participation in the network diminished and internal divisions widened (Hoogesteger, 2014: 176-181). This was one of several factors that contributed to the unravelling of the once-powerful network during Correa's presidency. The outcome was a surprise to the former vice-president of Interjuntas-Chimborazo who lamented: 'we thought we would strengthen our organization under Rafael Correa, but totally the opposite has happened'. 11 He was hopeful that the network would re-emerge but said it is difficult when 'the state is there at every moment.' 12

Bolivia: The coproduction of housing and vivir bien
In this section we examine how the coproduction of housing in Bolivia created a tension between national and local imaginings of development, and raises questions in relation to subject-making and collective autonomy. As has been shown in the Ecuadorian case, coproduction during the pink tide years involved the translation of the desires of multiple social actors, as well as the state, into policy and law. Throughout the Evo Morales governments (2006-2019), state programmes-including the coproduced social housing services examined here-were promoted through a discourse of living well (vivir bien, the Bolivian equivalent of Ecuador's buen vivir). Drawing on an idealised conception of life in the Andean community, the ayllu, vivir bien refers to the harmonious coexistence of humans, animals, nature and deities incarnated in the landscape. However, by Morales' second period in office, the meaning of vivir bien gradually became more anthropocentric, justifying the exploitation of non-renewable resources in order to enable human beings to live well (Schavelzon, 2015).
The main actors involved in the coproduction of social housing in Bolivia are the state housing agency and the recipients (as in Joshi & Moore, 2004). Beyond this, various community organisations that the recipients belong to, such as neighbourhood groups and local unions, also partake in the coproduction of the houses as well as the private companies that deliver the housing projects. In this section, the coproduction of social housing will be explored through three facets of the process: 1) The specific requirements of eligibility for social housing 2) The element of the production that the citizen brings to the production (contraparte) and 3) The house itself, including the materials, design and particular life values these engender and encourage (infrastructure). In exploring these three facets, an analysis is furthered of how coproduction creates political subjectivities and how it challenges and reshapes collective autonomies, building on the discussionof coproduced water services in Ecuador presented in the previous section. The current social housing programme, Vivienda Social y Solidaria was set up in 2006, 13 and is delivered by the state-housing agency, Agencia Estatal de Vivienda (AEVivienda), created in 2011. Social housing provided by the AEVivienda falls into one of six main categories, including the improvement of houses (Proyectos cualitativos, Mejorar casas), which involves the expansion and upgrading of existing houses (such as the addition of a kitchen or bedroom), and the construction of completely new houses (Construcción de Vivienda Nueva). These two categories make up the majority of social houses built in Bolivia; a third that we will discuss here is emergency housing, provided when people lose their homes due to unforeseen circumstances, such as flooding. 14 In addition to the provision of adequate and affordable housing, AEVivienda's stated mission includes 'constructing social equality and quality of life', and delivering a social housing programme that fosters 'communitarian living in harmony with Mother Earth in order to live well' (Agencia Estatal de Vivienda, n.d.). These missions tell us that social housing in Bolivia is about more than providing shelter: it is about creating particular ways of living and relating to one another. As in the Ecuadorian case of coproduced water services, even when recipients of social housing were allowed some input into the design of houses, their conceptions of living well were translated into the logics of a state development programme (Dinerstein, 2015(Dinerstein, , 2017 creating a gap that was politically charged. Most recipients of social housing in Bolivia selfidentify as indigenous (in the cases discussed, Aymara-and Quechua -speaking), and while social housing provision is welcomed by many recipients, it must be understood as part of a longer history of Bolivian state initiatives forcibly separating or tying indigenous groups to particular spaces and identities. 15

Coproduction requirements
All participating households have to meet certain key criteria, including: having a clear need for a new or improved house, possessing legal paperwork, demonstrating ownership over the plot of land and any existing house, and the ability/willingness to set aside three months to work on the construction of the house. As access is based on need, most recipients are families with dependants and the single adult households that are awarded houses tend to fulfil the criteria of old age or disability. In addition to these eligibility criteria, the houses come with certain rules: a person can only receive one vivienda estatal in a lifetime; upon construction the named beneficiary has to live in the house (it cannot be let out to a third party); the contract drawn up between the beneficiary and the AEVivienda must include both husband and wife (even if the land itself is owned by only one party); and the house may only be sold many years after construction.
Resembling the demands coproduction placed on water associations in Ecuador, these requirements threaten exclusion. In an urban area where neighbours make collective applications (usually one application involves a minimum of twenty-five households), not all will be deemed needy enough, and some may not have the necessary paperwork. In rural communities, it is not a requirement for communities to apply en masse, but when applications are made collectively, they likewise tend to find that not all families will be awarded houses. The fact that the final decision concerning eligibility is taken outside of the community undermines the notion that a community understands its own needs best, as is implicit in vivir bien discourse, and interferes with any community's ability to regulate internal equality. Instead the needs and wants of a community are re-shaped as they are translated into actionable policy in accordance with wider government vision. In some Charazani communities, decisions were taken not to solicit houses at all in order to avoid envy between families should all not be awarded a house. This echoes what Boonyabancha and Kerr (2018) have shown in Thailand for government-run housing projects, that demand-driven coproduction is better at delivering diverse and fit for purpose housing than supply-driven ones, in other words, moving these decisions out of communities rarely improve outcomes.
The AEVivienda's criteria assumed that recipients were individuals/households rather than communities, and land ownership was private, not communal. As a result, rural communities where land was held collectively without private property deeds had to fulfil the requirements either by shifting actual ownership practices in their communities, or inventively 'performing' individual ownership. At times, community authorities produced 'mock' individual title deeds to ensure compliance. In Charazani, households had to apply for social housing in cooperation with community authorities, from whom they obtained certification that the land on which they were soliciting a house to be constructed was their own 'property'; to be awarded such certification they needed to have fulfilled the expected social obligations of their community, such as participating in the rotative authority and fiesta sponsorship system known as fiesta-cargo. In this way communal engagement and contribution became a condition of private property.
In urban areas where land is held privately, the problem was documentary proof of ownership. The property market in periurban Cochabamba is fast moving, with thousands of recent migrants from rural areas buying plots of land annually. Many buy from a 'loteador', a dealer who purchases large stretches of land from landowners, subdividing them to facilitate quick sales. Loteadores often work at the margins of legality and land changes hands without the proper documentation to go with it (see also Goldstein, 2012: 90-94). When new owners attempt to enter their plot in the land registry, pay property tax, or apply for social housing, they often discover that they are not in possession of the correct paperwork and cannot easily prove legal ownership of the land (Sheild Johansson, 2020). The coproduction process thus demands a level of formality that does not exist in peri-urban zones, resulting in the exclusion of many households and/or bureaucratic battles that define people's lives for many years. Here, state resources, and by extension basic rights of citizens, are made 13 It was predated by other state-led housing initiatives, such as Fondo Nacional de Vivienda (FONVI), but has been considerably more productive than its forerunners, building on average seven times more houses than previous projects in any given time period. 14 Ojeda et al. (2018: 539) argue that the benefits of coproduction in such cases is clear: members of civil society generally lack the tools and financial resources to mitigate or respond to disaster effectively, and government agencies do not possess the necessary local knowledge. 15 Such as the Disentailment Act of 1874, which dissolved indigenous communities and created private territories, the Law of Popular Participation (LPP), (1994), and the Law of the National Agrarian Reform Service (INRA) (1996). These government policies of the 1990s were inherently contradictory, with autonomy and survival being predicated on performing particular identities and conforming to imposed moral imaginaries about indigenous citizens (Andolina, 2003). As such, experiences of 'participation' and 'coproduction' have in recent history been linked with both the withdrawal of state resources and the circumscription of lives led. inaccessible by the reality of the 'informal' economy within which people live. 16 Finally, when the paperwork for the house is drawn up, the recipients must sign a contract. One element of this contract is a commitment to participate in workshops that aim to develop the construction skills of recipients and educate them on certain social issues. According to the AEVivienda, the training provided through the coproduction workshops improves the recipients' future employability as construction workers, and is therefore an added good. The workshops on social issues delivering the vivir bien element of AEVivienda's mission statement, cover topics such as domestic abuse, female education, hygiene, and 'proper' care for the surrounding environment. They explicitly aim at producing particular social relations and dispositions: not just a house, but also ways of inhabiting the house and the wider country, as a particular type of subject.
Taken together, the 'requirements' work to produce specific kinds of subjectivities-citizens who are aligned with the vivir bien ideology as well as practising a modern indigeneity, through living in new houses, formalising their lives through legal documents, and following certain hygiene expectations-besides undermining the autonomy of local collectives through intervening in norms around landholding.

Coproduction and the contraparte
It is important to note how and when communities are invited to co-produce and to specify what is meant by coproduction, as these factors greatly impact on how these projects develop and are experienced. In this instance, the contraparte is the element of the production that the recipient brings to the table and includes both upfront and future contributions (see also Goodwin, 2019). For both new houses and house improvements, the two main resources that the recipient brings to the coproduction process are land and labour. The AEVivienda provide expertise and materials, both delivered through private companies. Unlike instances of housing projects where the recipient community contributes to the production of the house at the design stage (see Ojeda et al., 2018), in urban Bolivia recipients had little influence at this stage and were only involved in construction.
Land, meanwhile, is the facet of the contraparte that most concerns people in urban areas. Coproduction is often touted as giving people a sense of (private) ownership over something that is public. In the urban Bolivian examples we have the opposite-here, privately owned land suddenly has a public element (house) attached to it. In the zona sur there was deep concern about the potential impact of putting a government house on one's land. In a context of a broader property market characterised by contested claims, adding another potential 'claimant' (the state) to the list was not done lightly. Once a plot had a vivienda estatal sitting on it, the owners were tied to that location, and the land was no longer flexible capital. In a sense, as one interlocutor said 'once you build a vivienda estatal, you no longer own your land, you have a home, but no land'. This element of the contraparte thus involves handing over certain rights to privately held land (including selling and sub-letting) and restricts private speculation on property, turning 'informal' speculators into 'formal' homeowners. Many people also anticipated that taxes/fees might be demanded by the government at a later stage, while others worried that future governments may want to be paid back for the full cost of the house, or were concerned about eviction-'Se va a expropiar' (they [the state] will take possession of it), was a common refrain. While the AEVivienda strongly denied the possibility of land being taxed in new ways once a social house was built on it, or of future payments, the building of a social house undeniably circumscribes how that land can be used.
In both rural and urban areas, the labour element of the contraparte usually involves committing at least one member of the household to working on the construction full-time for three months, resulting in significant loss of earnings. Many would draw on extended kin and neighbourhood networks to bring in extra labour, echoing the collective labour practices that underpin coproduced water services in Ecuador. Female-headed households where the labour could not be provided from within the home often paid a labourer to provide this part of the contraparte. Both forgoing wages and paying for an external labourer demanded capital, which if declared could deem the applicant ineligible. Providing the contraparte without undermining one's requirements thus required a certain amount of creative management of resources.
In rural Charazani, the emergency houses were ostensibly free, posing a different kind of a dilemma for inhabitants. Many people were suspicious of supposedly free gifts, which do not fit the Andean notion of reciprocal exchange and ayni (mutual aid), in which a gift given, or a service performed is always part of an extended and ever embedding network of interdependence, exchange, and care. While several beneficiaries referred to their houses as 'gifts from Evo' they also expected to later be asked to give something in return. As one recipient remarked with suspicion: 'it will produce taxes, it will produce like the plant produces' (Alderman, 2021).
Thus, while the contraparte, from the government's perspective, makes social houses affordable, allows recipients to experience a sense of ownership over them and provides an opportunity for the state to educate citizens in ways of living, from the recipients' perspective the contraparte continues the historical experience of the state circumscribing landownership, and collective work and decision-making practices for indigenous groups, thus reordering existing collective autonomies.

Coproduced infrastructure
The house itself is a third site of political tension. By entering into coproduction with the state, people agree to live in a specific house, which promotes a particular way of life. For instance, the number of bedrooms, type of toilet, and building material used are all dictated by the AEVivienda.
In rural Charazani houses were designed in local meetings in consultation with architects working for the construction companies and while they conformed to a 'social house mould', the population reported feeling included and listened to. 17 But there were also clear areas where local needs were misconstrued, demonstrating that in contrast with one of the claimed benefits of coproduction-that these type of mismatches can be avoided through the involvement of local knowledge-the translation (Dinerstein, 2017(Dinerstein, , 2020 of local suggestions and proposals into the logic of the AEVivienda meant that gaps in meaning emerged. Toilets and stables for alpacas were two sites of mismatch. According to recipients, neither was fit for purpose, yet both had been accepted as these were crucial elements of the AEVivienda's vision of rural houses and 16 See Meagher (2013) for criticism of coproduction as a process that places an undue burden on informal workers and businesses when accessing public services. See also McMillan et al. (2014). 17 Mitlin and Bartlett (2018) and Fieuw and Mitlin (2018) similarly argue that an important element of service provision through coproduction is the inclusion of community knowledge and preferences in the design process. Boonyabancha and Kerr (2018) argue that empowering communities to actively participate in housing projects, including the form the housing takes, leads to both more efficient use of state resources and variety rather than standardisation, though they also show in relation to their case study from Thailand, that attempting to achieve government set targets can lead to increased standardisation in a project-led approach over time. recipients felt there was little point requesting changes due to said vision and budget restrictions.
Building materials are another important issue to consider in rural areas. The social houses were red brick constructions, while rural Andean houses are traditionally made of adobe-a material thought to connect inhabitants with deities in the landscape around them. Through the materials that the mountain provides (mud and straw for walls, wood for roof timbers, and grass for the roof-covering), the house becomes 'a personified representation of the mountain' (Arnold et al., 1991: 6). In initial meetings with architects in Charazani, opinion was divided-most people favoured red-brick as a construction material, but a minority argued that the state should help them to build 'rustic houses' made of traditional materials. Alderman (2021) (see also Bold, 2016:171-173) contends that most people welcomed the change in materiality as a means to project themselves as modern, urban-oriented citizens, distancing themselves from their adobe houses and the association of mud and earth with inferiority, as projected by urban-dwellers (Canessa, 2012:193;Orlove, 1998:217). The materiality of social houses thus became a political vehicle for the inhabitants, but also created inter-community strife.
While rural areas were afforded some flexibility regarding the appearance of social houses, they had, on the whole, to conform to the adequate and modest aesthetics of a vivienda estatal. In urban areas the situation was stricter and any elaboration on the provided 'kit' house was forbidden. In fact, many participants of the program added rooms to their houses during construction, or features such as balconies and arches, which were met with dismay by AEVivienda employees. As architect Nellie, an AEVivienda procurement and project manager, exclaimed as she inspected the progress of a house with an illicit second floor and balcony: 'This does not look like a vivienda estatal any longer. The government can't be seen to be building these kinds of houses, or supporting people who can afford these kinds of houses, this is more than this family needs.' The recipients' ambitions and desires for a dream home did not conform with the AEVivienda's interpretation of equality and living well. While recipients made efforts to truly make their new homes their own, they usually had to submit to producing a house according to the government's vision of a good Bolivian home in which citizens could vivir bien (live well). This resonates with Fieuw and Mitlin's (2018) argument that a drive for professionalism in South Africa's housing programme stifles community action in coproduction processes (see also McMillan et al., 2014), and again highlights that coproduction often involves the translation of community visions of development into a form legible to the state.
In this Bolivian example, coproduction intervenes in practices of ownership, resource management, community relations and the very notion of what a house is and how it should be lived in. While the AEVivienda insist the state has no desire to lay claim to the private land upon which these state-funded houses are built, they nonetheless enforce a set of conditions that stipulate the kinds of life that can be lived on these plots, what may be built on the land, how it should be owned, and how people should conduct themselves-thus promoting a particular kind of citizen subjectivities and undermining local collective autonomies. Ostensibly drawing on highland indigenous notions of living well and a home being not just a place to live, but a space through which relationships with humans and non-humans are made and maintained, the viviendas estatales aim to do much more than provide an affordable roof over peoples' heads. While the houses are desired and appreciated by many, they also throw up new concerns about ownership, formality and inheritance, as well as triggering the age-old theme of state dictation of how indigenous groups should live in order to deserve access to resources, be it land or public goods.

Uruguay: The coproduction of waste services under the Frente Amplio
Unlike in Ecuador and Bolivia, pink tide government in Uruguay was not tied to a single figure. Two presidents dominated the Frente Amplio's stretch in government from 2004 to 2019: Tabaré Vásquez, who enjoyed two spells as President (2005-2010 and 2015-2020) and José 'Pepe' Mujica, who served from 2010 until 2015. With regard to coproduction and the provision of public services, the Frente Amplio sought to expand what it termed the 'social economy', establishing cooperatives as key partners of the state in the coproduction of certain services (as in Joshi & Moore, 2004). Uruguay has a long history of cooperativism, particularly in the housing sector, where the national housing cooperative federation (FUCVAM) was perhaps the most active social movement actor during the dictatorship (González, 2013) 18 and the Frente Amplio years also saw an expansion of cooperatives in the catering, care, cleaning, and, intermittently, waste and recycling sectors (Reyes & Guerra, 2019).

Coproducing waste and recycling services
Household waste collection and disposal in Uruguay is, as in most countries, a responsibility that falls to municipal governments. Household, commercial, and industrial waste generation is most pronounced in Uruguay's capital, Montevideo, which has a population of 1.4 million, 40% of Uruguay's total population. Greater Montevideo, which includes the neighbouring departamentos, has a population of close to 2 million, around 60% of the total. Waste services in Montevideo have been under the control of the Frente Amplio since 1989, while the government of Canelones, Uruguay's second most populous and neighbouring state, has been controlled by the Frente Amplio since 2005.
Waste policy and politics in Uruguay are inseparable from the presence of classifiers (clasificadores), the name given to wastepickers who have long carried out the lion's share of recycling in the country, before any municipal recycling schemes were implemented. The presence of waste-pickers is documented since the end of the 19th century, in the vivid accounts of flaneur and first mayor of Montevideo Daniel Muñoz (see Carrasco, 2006). A key part of the metal, paper, and then plastics industry, wastepickers have nevertheless been regularly repressed, and their activity was formally outlawed during the dictatorship, which established waste as the property of the local government. Waste-picker carts were burned, horses confiscated, and some workers rounded up, imprisoned and tortured.
On assuming the mayoralty, Tabaré Vásquez decriminalised the activity (O'Hare, 2017). Successive mayors of Montevideo have nevertheless had problematic relationships with clasificadores. Waste and rubbish in the city have long been one of the main issues that plagued local government, with pictures of waste piled high in the streets a regular feature of denunciatory newspaper reports from at least the 1960s onwards (ibid). Inadequate waste services troubled Montevideo's claims to 'infrastructural modernity' (O'Hare, 2022): as one newspaper headline asked, was the city the Switzerland (Suiza) or the trash-heap (sucia) of the Americas? (Época, 1966). Further, who held the blame for the accumulation of waste? Whilst corruption, poor management, and municipal strikes were singled out as contributing factors, so too were clasificadores, previously known as rummagers (hurgadores) for their apparent tendency to rummage through the city's bins in the 18 Pepe Mujica launched his own housing scheme for the country's poorest residents, Plan Juntos, to which he donated 90% of his own salary, sought further donations, and engaged volunteers from the country's Communist Party-dominated construction union. search for recyclables, leaving undesired materials spread out on the street. The question thus loomed: were waste-pickers coproducing a public service (recyclable collection and separation) or were they impeding the delivery of that service?

(Un)Cooperative subjects?
When the Frente Amplio assumed national office in 2004, waste collection and disposal was carried out by a range of public and private, formal and informal, economic actors. The municipal fleet collected and disposed of household waste in most of the city, except the central district and old town, where delivery was privatised in 1995. A plethora of private actors were responsible for collecting and disposing of industrial and commercial waste, including several skip companies. All brought their waste to what was then Montevideo's only landfill, Felipe Cardoso, which is operated by the municipal government. Informal sector waste-pickers meanwhile were responsible for the collection and classification of recyclables, which they supplied to the national and international recycling industries.
The change that the government sought to bring about in this arena largely related to the activity of waste-pickers, whose labour they aimed to formalise and collectivise. The Ley de Envases (Packaging Law), a 2003 law that uses a voluntary levy on companies that produce and import packaging to finance recycling plants, formed the legal framework that would allow this to happen. Yet the implementation of the law was staggered across different departamentos, and the complexity of the Montevidean challenge, and the multitude of public and private bodies involved in its implementation, meant that recycling plants only became operational in Uruguay's capital city in 2014. The four plants only employ 128 workers, at most 10% of the city's waste-pickers, and they are only mandated to receive household waste from containers distributed throughout the central district, in supermarkets, and in large housing complexes. Multi-million-dollar trucks and accompanying containers were purchased by the private concessionary for the collection of recyclables and their delivery to the plants, paid for with government money through their public concession. 19 Waste collection and disposal is somewhat different from other infrastructures and coproduced services in that it involves the removal of a 'bad' (waste) rather than the provision of a 'good', as in the cases of water and housing. At least in the Global North and the formal Latin American city, water and electricity provision also generally involves minimal involvement from the public, whilst the removal of waste inevitably involves some public participation: the placing of items into a bin or bag and the transport of these into the street for collection. Recycling can involve a further labour of classification, where items might also be washed and cleaned before being placed into separate containers. Until the provision of Ley de Envases containers, however, the labour of classification was incumbent only upon waste-pickers, who embraced the title of 'classifiers' to highlight their valuable economic and environmental labour. With the implementation of the Ley de Envases, the coproduction of a recycling service was extended to households, which were asked to separate at home and take their recyclables to different points in the city, from where they would be transported to recycling plants to undergo further processes of classification by waste-pickers. To this extent, coproduction emerged as the delivery of a recycling service was expanded out from informal wastepickers, to household classifiers and the state managed collection of recyclables.
Such changes in waste and recycling policy involved the creation of new subjects. First, the ethical domestic recycler, or what Hawkins (2006) calls the 'anxious recycler', classifying materials at home. Such recyclers might already have existed, as in the case of those who separated out recyclables and other items that might interest clasificadores, hanging them in little bags from containers and trees on Montevideo's streets. At the very least, however, these activities became standardised: only stock recyclables like plastics, paper, metal, and cardboard were to be placed in the new containers. There were also several attempts to create new waste-picker subjects. The first of these was the emphasis placed on the 'cooperative waste-picker', the priority of Uruguay's Ministerio de Desarrollo Social (MIDES) and its program for clasificadores-Programa Uruguay Clasifica (PUC)-from 2006 to 2013, when clasificadores were urged to form cooperatives, whether these could be formalised or not. As occurred elsewhere in Latin America, the mistaken assumption was that waste-picking was a largely atomistic activity, even a dog-eat-dog world, whereas in fact clasificadores were often organised in family groups (Carenzo & Miguez, 2010). Yet the condition for waste-pickers to be able to continue working at the landfill, to continue coproducing a recycling service in alliance with the state, which brought them waste and recyclables, was the formation of a cooperative.
By the time of the construction of the Ley de Envases recycling plants in Montevideo, however, the focus had shifted onto the 'formalised waste-picker'. Clasificador cooperatives were deemed to have been a failure by the state and researchers (see, for example, Sarachu & Texeira, 2013), undermined by prevailing individualistic tendencies and kinship organisational structures in the clasificador population (not to mention a lack of financial and legal support from the state). The plants were not to be cooperatives but would instead be managed by NGOs with experience of working with 'vulnerable' people, with a transition to cooperatives envisaged further down the line. In this case waste-pickers would become fully formal workers, paying taxes and receiving social security. The push for formalisation went beyond the plants: businesses would now be penalised if they gave their recyclables to an informal waste-picker rather than contracting a formally registered waste management company or the municipality. Further, Montevideo's new Ley de Envases containers were specifically designed so that they could not be opened by street level waste-pickers, only by privately operated trucks. Whereas previously it hardly made sense to talk of waste-pickers as being informal (they were the only recycling 'form' in existence, see Fernández, 2012), state policies effectively created a cleavage in the population, even a process of 'dispossession-by-differentiation' (Kasmir & Carbonella, 2008), by counterposing cooperative to individual clasificadores, formal to informal. The contours of an acceptable coproduction of waste and recycling services were being traced, in line with hegemonic ideas of dignified work and modern waste disposal.
At one point it was even debated whether, in order to work in the recycling plants, waste-pickers would have to hand in their horses and carts. This never transpired, but when one wastepicker at a training session for formal work asked if there would be parking for horses and carts at the plants, he was told that this would not be appropriate. Yet, just as the state in Bolivia could only be seen to be constructing houses for the poor, and thus discouraged housing modifications and extensions that troubled a modest aesthetic, Uruguayan state officials were very keen to emphasise that the recycling plants were being constructed for needy subjects. Thus, when at the same session another clasificador asked about parking for his car, he was told that parking needs had been 'estimated based on the type of people entering the plant': clasificadores weren't expected to own cars. When this discussion was raised with a government official involved in the implementation of the plants, he explained that 'if we were seen building recycling 19 Interview, private concessionary coordinator, 10.10.14. plants for people with cars then we would be questioned by international organisations who would ask what this had to do with the eradication of poverty'. 20

Undermining Autonomy
Through the various iterations of its waste and recycling policy, one relation that the Frente Amplio sought to transform was that between the citizen or neighbour (vecino) and the waste-picker. Clasificadores were assumed to suffer stigma as they collected waste in the city, and the PUC sought to visibilize their activity, promoting the wearing of uniforms and distributing booklets informing them of their rights. One successful cooperative, Cooperativa Juan Cacharpa, established positive relations with the neighbours of a large housing cooperative in 2009: where they had previously been viewed with suspicion, they were now invited into homes, given Christmas gifts, and asked to help with removals (Zoppis, 2011). Yet with the failure of the cooperative collection model, this 'soft infrastructure' of human contact was replaced with the 'hard infrastructure' of municipal containers and waste-pickers were confined to recycling plants. Contact between plant-based recyclers and the general population became limited to a fleet of municipal motorised tricycles that at the time of writing had been given to only a small fraction of the clasificador population. The prohibition, by municipal decree, on businesses giving waste to informal waste-pickers meant that positive relations between business owners and clasificadores that might have existed for years were now compromised. For waste-picker representatives, such legal manoeuvres were unfortunately a more effective way of dispossessing waste-pickers than outright repression. 21 Through the marginalisation and criminalization of informal wastepickers, and the disbanding of wastepicker cooperatives, different forms of autonomy were undermined. The first was what Millar (2014) calls the 'relational autonomy' of boss-less, independent recycling work in and around the landfill and on the city streets. Despite hardship, clasificadores prized the ability to control their own working day, and avoid being ordered around by an employer or overseer, elements of autonomy that were conserved, if transformed, through the development of cooperatives (O'Hare, 2020). These were ways of working that were defended by the clasificador trade union, the Unión de Clasificadores de Residuos Urbanos Sólidos (UCRUS). Established in 2002 and influenced by but less successful than its equivalents in neighbouring Brazil and Argentina, the UCRUS is tasked with representing a wide body of waste-picking subjects. It has promoted and supported cooperatives and plants, whilst at the same time defending the right to work and circulate in the city with horse and cart and at the landfill, as clasificadores have historically operated. It boasts of being the only waste-picker trade union in the world to be affiliated to its national trade union congress (the Plenario Intersindical de Trabajadores-Convención Nacional de Trabajadores or PIT-CNT), yet despite the influence of the PIT-CNT over the Frente Amplio in what were increasingly corporatist governments (Silverman, 2011), the UCRUS failed to gain much traction, perhaps due to its anarchosyndicalist tendencies (O'Hare, 2017). The two most important UCRUS successes were the blockading of the landfill in 2002, which inaugurated the union and thanks to which landfill wastepickers could continue working, albeit as a cooperative, and the 'march of the carts' in 2008, which succeeded in discontinuing the confiscation of horse and carts and maintained their right to circulate in most parts of the city. Whilst these struggles were largely articulated through a 'right to work' discourse, the environmental role of waste-pickers was also stressed by their representatives, who were keen to emphasize that rather than being disruptive 'rummagers', clasificadores were involved in the co-production of an environmental service.
The Ecuadorean case discussed above raises the issue of the problematic 'translation' of social movement demands into state policy, and the questions that this raised for the autonomy of water associations. The UCRUS did put forward demands, some of which were ahead of their time, such as for a plastic bag charge which was eventually introduced in Uruguay in 2019, without any of the funds directed to recycling, as the UCRUS had proposed. Overall, the union found itself too weak to substantively influence government and laws such as the Ley de Envases, although it had slightly more success with the (as yet unimplemented) Ley de Gestión de Residuos (Waste Management Law) of 2019. They faced similar consequences of these laws as in the Ecuadorean case: restrictions on clasificador autonomy and attempts to fit diverse, vernacular forms of waste collection, transport and recycling into a single, formalised model. Only a certain form of waste and recycling coproduction could, it seems, be countenanced by the state: one in which the role of a restricted group of waste-pickers was reduced to formalised classification rather than collection and where there was little opportunity to add value to raw materials (through their processing, for instance). Although this did provide benefits for a small group of clasificadores, it did not, unlike the Chilean case of cooperative recycling coproduction, 'unleash wastepickers' potential' (Navarrete-Hernández & Navarrete-Hernández, 2018) and was accompanied by punitive measures for those who continued to labour informally, rowing back Tabaré Vásquez's landmark decriminalisation 30 years previous.

Conclusion: The politics of coproduction in comparative perspective
The coproduction of public services in Latin America has provided opportunities for rural, urban and peri-urban communities, social movements, and organised labour to influence the design, implementation and management of infrastructures that satisfy basic rights and needs, such as access to water, housing, and waste collection. While service improvement often results, coproduction is just as much about the transformation of relationships between citizens and the state. In these processes, states are given an opportunity to shape particular models of citizenship and sovereignty. However, state plans and visions are rarely passively accepted by the organized citizens involved in coproduction and political struggles emerge as the interests, practices and objectives of the various actors involved in the process collide.
Building on this, we have argued that politics is internal to coproduction rather than a mere contextual concern or strategy for organized citizens to renegotiate with the state. We have shown this by exploring the coproduction of three distinct public services-water, housing, waste-in three countries where leftleaning governments were elected into office in the early twenty-first century-Ecuador, Bolivia, and Uruguay.
Our analysis has centred on two interrelated issues that are common to the three cases. First, all the coproduction projects sought to create and shape a particular type of subject that broadly aligned with the interests and ideologies of the respective pink tide governments. Second, the struggle over subjectivity and citizenship was also a struggle over collective autonomy, where water associations, social movements, indigenous communities, waste pickers and labour unions sought to preserve certain forms of autonomy whilst also engaging the state. These issues provide new critical perspectives on the lived experiences and politics of coproduction. 20 Interview with official, Department of Solid Waste and Chemical Substances, National Environment Division (DINAMA), 28.4.14. 21 Interview with Patricia Gutiérrez, UCRUS Secretary, 6.12.14.
The subject-making capacity of coproduction has been noted in the existing literature (see, for example, Mitlin, 2008;Goodwin, 2019). However, we have delved deeper into this contested process and illustrated how it takes distinct forms in different contexts. In Bolivia, coproduced housing sought to encourage private land and house ownership; and while it enforced collective applications for urban coproduced housing, it did so only on the premise that some families would be deemed ineligible. In Uruguay, waste and recycling policy saw attempts to create an environmentally-conscious citizen-recycler who would classify waste materials domestically, but also efforts to transform the subjectivity of clasificadores, informal waste-pickers who had traditionally carried out the lion's share of recycling in the country. In Ecuador, laws and policies were introduced and other tactics and discourses employed to attempt to mould obedient political subjects dedicated to the socalled revolución ciudadana. The types of political subjects that 'pink tide' governments attempted to shape through coproduction thus varied in each case. One common feature, however, was the effort to forge 'modern' subjects, whether it be waste pickers expected to leave behind the horse and cart, indigenous families required to inhabit state-designed houses, or water associations expected to embrace new technologies and organisational forms to manage water systems.
Struggles over the efforts of 'pink tide' governments to impose their particular modernising visions of capitalist development on citizens involved in coproduction were closely connected to the issue of collective autonomy. The existing literature shows that tensions between autonomy from and engagement with the state are common to coproduction (see, for example, Mitlin ,2008;McMillan et al., 2014;Goodwin, 2019). Within our research sample, these tensions came most to the fore in Ecuador. Legislation introduced by the Correa government recognised the financial and administrative autonomy of water associations. Yet the process of legalisation provoked challenges and changes to the associations, which in order to be rendered legible by the state needed to conform to particular rules and adapt prescribed organisational structures. This threatened their substantive autonomy by undermining their capacity to maintain or create their own structures, practices, and relations. Intriguingly, state interference in water associations in Ecuador was described by one activist as being 'like someone coming into your house and telling you how to live'. This was literally what occurred with the coproduced housing in Bolivia, which placed restrictions on the houses that were built or repaired through the AEVivienda program. Whilst the central prohibition concerned the sale of the house, and the land that it sat upon, rather than what occurred within, it was clear that state involvement also shaped how residents inhabited and modified the houses they had helped to construct, with balconies and additional rooms discouraged, for instance. More broadly, the final decision over who would benefit from new housing or repairs was largely taken outside of communities, some of which decided not to apply to the scheme to preserve their autonomous decisionmaking. Something that Uruguayan waste-pickers and their union sought to preserve where possible were the forms of autonomy available in 'boss-less work', whether that took place in cooperatives, in landfills, or on horse and cart. The autonomy of wastepicker cooperatives was also undermined by the transition to NGO-managed recycling plants. Uruguay's largest waste-picker cooperative, operating on municipal land, was disbanded and evicted by the local government, which instead offered the workers places in one of the new plants. Although pay and conditions improved, autonomous control over space and time was what waste-pickers missed most. The cooperative had been a social hub rather than simply a workspace. The recycling plant, by contrast, was a hygienic, policed space, complete with a security guard, who initially sought to ensure that workers did not make off with anything from the conveyer belt. In each case, opportunities for the organized citizens involved in coproduction to influence the laws and policies that impacted them were limited, echoing problems reported elsewhere in the coproduction literature (see, for example, Mitlin & Bartlett, 2018;Ojeda et al., 2018;Moretto et al., 2018;Li et al., 2019;Rosaldo, 2022). The problems of translation that we have discussed in each case illustrate this (see also, Dinerstein, 2017Dinerstein, , 2020. To conclude, the political issues that we have discussed in this article create opportunities as well as problems. Our critical analysis has focused more on the latter than the former. However, this should not be read as a uniform critique of left approaches to coproduction. Quite the opposite, in fact. While 'pink tide' governments instituted important advances in the sphere of coproduction (and elsewhere), they often employed a top-down, state-centric form of authority, which limited improvements, stoked political tensions, and eroded popular support. It is thus important for left governments to take coproduction seriously and learn from these experiences to build more effective models of collaboration across different services, which as well as strengthening participatory democracy, can help to reduce inequality and exclusion and create spaces for new forms of politics. A future research agenda should explore the on-going impact and legacy of the 'pink tide', how coproduction policies and experiences shifted as it ebbed, and how the contours of coproduction are again being redrawn as Latin America welcomes a new generation of left-leaning leaders in countries such as Chile. This research agenda should not only consider how the politics of coproduction compare across Latin America but also elsewhere in the Global South. Through this global comparative research agenda, a fuller understanding of the significance and potential of coproduction should emerge.

Declaration of Competing Interest
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.