Patterns of interaction between social movements and participatory institutions in the Brazilian policy subsystems1

In the post-constitution period, up to 2013, a consensus was established in Brazil about the idea and practice of social participation. The study agenda on the subject was no longer focused on social movements but, rather, on the analysis of the dynamics of civil society and formal arenas of participation that embrace representatives of civil society and state. The shift marked the institutionalist turn in the studies of social participation in Brazil. Large demonstrations against the ongoing austerity policies in Brazil, as in other countries, mark an inflection in this study agenda. Social participation within the participatory institutions has been criticised, and a new turn has emerged. By recognizing those changes, this paper comparatively describes and explains if, how and to what extent different patterns of participation are mixed in order to guarantee social rights in health, housing, and policies for women policy subsystems in Belo Horizonte, Minas Gerais.


Introduction
In the post-constitution period, up to 2013, a consensus was established in Brazil about the idea and practice of social participation. The study agenda on the subject was no longer focused on social movements but, rather, on the analysis of the dynamics of civil society and formal arenas of participation that embrace representatives of civil society and state. The main purposes were to expand social rights and to further Brazilian democracy (Abers and Von Büllow, 2011;Faria and Tatagiba, 2017).
The shift from analyses on social movements to analyses on participatory institutions (henceforth PIs) marked the institutionalist turn in the studies of social participation in Brazil.
This shift reflects the practice of social actors stimulated by the 1988 Federal Constitution, which required the creation of participatory institutions such as public policy councils and conferences within different public policy subsystems.
This institutionalist turn has given rise to different interpretations on the nature of the relationship between representatives of civil society and state within these and other participatory institutions such as Participatory Budgets, Executive Plans, and others. In this process of analytical convergence, public policies and participation have become related terms, notably regarding social areas .
In this context, scholars concerned with the expansion of political and social inclusion began to investigate a set of variables that would test the inclusiveness of these PIs. The decision-making process and capacity, the legitimacy of representation, the impact of decisions on public policy cycles, as well as contextual and political variables, were mobilized to explain those novelties (Lavalle et al., 2005;Dagnino & Tatagiba, 2007, Lüchmann, 2007, Cortes, 2009, Avritzer, 2010, Pires, 2011. In Brazil, as in other countries, large demonstrations against the ongoing austerity policies mark an inflection in this study agenda (della Porta, 2015). Social participation within the PIs has been criticized, and a new turn in the field of social participation has emerged.
From that point on, the emphasis on institutionalized participation has been challenged by the recognition of confrontations seeking to denounce the violation of social rights, injustices, as well as the limits of the State by means other than the formation of those institutionalized arenas.
By recognizing those changes, in both practice and analysis of social participation in Brazil, this paper describes and explains if, how and to what extent different patterns of participation mix in order to guarantee social rights in the city of Belo Horizonte, Minas Gerais 2 .
If it is true that there is a new turn in the field studies on social participation in Brazil (and elsewhere), that calls into question the established connections between participation and institutionalization (such as the creation of formal arenas as councils and conferences), this does not necessarily presuppose a break with those relations that have been consolidated for the last 30 years in the country. This analysis will be conducted in three subsystems of policies: health, housing, and policies for women 3 . Different methodological strategies, that involve documentary analysis of council resolutions and protest events (AEP), as well as in-depth interviews 4 with representatives of social movements and PIs, will be used.
After this introduction, we will follow three steps. In the first section, we will present a brief review of the literature on the subject, showing how a relational dynamic between social movements and PIs widens the study field on participation in different subsystems of policies.
In the second section, we will present the cases, data collection and treatment. In the third section, a brief historical contextualization of each policy subsystem will be presented, along with a comparative analysis of repertoires of action used in each one. Finally, in the fourth section, we will suggest possible explanatory variables for the similarities and differences found among them.
1 -Interaction between social movements and participatory institutions: selective review of the literature The relationship between social movements, civil society organizations and the State is a central subject in the social sciences. The interpretations of these relations are marked by controversies (käntrəˌvərsē) that involve both caution and/or opposition between the parties, and by the recognition of the existence of multiple possibilities of interaction (Habermas, 1986;Melucci, 1989;Tarrow, 1989;Avirtzer, 2000;Diani and Bison, 2003, Lavalle et al., 2004, Silva, 2007, Abers et al, 2014. If recognized, these relationships always demand qualifications. Authors such as Offe (2017), for example, condition this relation to the capacity of institutions to promote voice (inclusion) and to prevent the "exit" of those involved. The presence of these two criteria, inclusion and cooperation, define the possibilities for success of the interaction. Approaches centred in society have always viewed these relations with suspicion. Although authors recognize the importance of the institutionalization of social movement demands, they always call our attention for the risk of co-optation, deradicalization and demobilization of civil society actors (Habermas, 1988;Offe, 1989;Tarrow, 1994).
The analytical separation between these two poles, movements and State, as well as the valorization of contentious forms of action have been contested in Brazil and elsewhere (Abers and Tatagiba, 2014Lavalle et al., 2017, Goldstone, 2003and Felicetti, 2017). Different from such approaches, these authors will argue that social movements do not act only by contentious means, without cooperative ties with power holders in the production of policies (Lavalle et al, 2017). They also argue that the use of institutional channels does not necessarily lead to demobilization and co-optation of civil society organizations. This may even promote innovations and democratization in the public policies linked to them (Della Porta and Felicetti, 2017).
Although the movements are formed from the collective perception of social malaise (məˈlāz) (Melucci, 1996), operating in a conflictive and challenging way to make this sentiment public, they become sources of innovation whenever they can transfer their practices to political institutions. In doing so, they change their institutional boundaries. They are able to politicize institutional structures, helping establish equality criteria inside and outside political institutions (Della Porta and Felicetti, 2017;Avritzer, 2009). By transferring innovation, they often blur the traditional boundaries between mobilization and institutionalization. They form communities and coalitions that involve many actors inside and outside institutions (Diane and Besson, 2010).
These changes require a broadening of their usual patterns of action: dialogue and cooperation, not just contestation, become part of their repertoire. Depending on the assumed configurations, discursive coalitions can be conducive to institutional changes (Bohaman, 1996;Cortês, 2009). Otherwise, differing repertoires will be mobilized and the dispute between meanings and means to achieve such changes will be replaced (Hajer, 1999).
This permanent tension allowed the analysis, in specific contexts, of "multiple repertoires of interactions" (Abers et al., 2014), whose use will depend on a set of factors involving both political opportunities (Lavalle et al., 2017) as well as shared identities and bonds of solidarity that help to redefine such contexts (Wampler, 2015;Tatagiba et al, 2018).
Brazil, characterized until recently by the presence of a "participative citizenship regime", experienced different patterns of action and interaction: contestation, closeness, cooperation and negotiation (Abers et al., 2014). Social movements and their allies have used them to achieve their multiple goals, but they have not done so in the same way. It is exactly such a statement that requires explanation, thus justifying the investigation of the contexts that lead organized actors to use this or that type of action outside and/or inside the institutions, such as the public policy councils analysed here.
In addition to the contexts that determine the multiple interactions, we have the types of public policy subsystems as they involve different "state capacities", understood as administrative, financial, and "relational political" resources (Pires and Gomide, 2016). The capacity of the subsystems to articulate the inclusion of multiple actors in the processes of formulation, implementation and oversight of their public policies, ensuring more legitimacy, learning and innovation, will also depend on the characteristics and motivation of the actors, as well as on the conditions found by them to focus on the rules that define who and how they deliberate in the formal spaces for interaction with the State. As Wampler (2015) states, representatives of civil society organizations have acquired, in various Brazilian policy subsystems, more information and access to public representatives, either by direct means or through the created participatory institutions. This change is important since it demonopolizes and democratizes access to state resources, but, at the same time, it can also result in more control over the movements' activities and strategies and over their social representatives (Wampler, 2015;p.171).
This uncertainty, far from denying the relations, justifies their analysis. In this paper, we will map, in each of the three public policy subsystems, what municipal health, housing and women councils have done from 2003 to 2016 (3.2.1). We will also analyse the strategies used by the social actors in these subsystems to publicize their demands and to force state boundaries beyond the actions and achievements undertaken by the participatory institutions analysed (3.2.2). It seeks, therefore, to describe and explain comparatively social movements and civil society organizations' institutional and extra institutional patterns of actions to secure their demands. Before discussing the data, we will present the methodological strategies used to analyse the patterns of action in each policy subsystem.

-Cases, data collection and treatment
This paper evaluates three Brazilian local, public policy subsystems: health, housing, and policy for women. Each comprises participatory innovations such as local councils responsible for debating, deciding and overseeing their policies. The municipal health, housing and rights for women councils are investigated based on their decision-making capacity, which includes the quantity and quality of internal resolutions produced by each. This paper claims each council's decision-making capacity depends on the institutional and financial resources held by each subsystem. This variable might explain the patterns of participation used in the subsystems. The councils are embedded in public policy subsystems that provide such resources.
In Belo Horizonte, the three councils have similar participatory paths, but they are immersed in policy subsystems with different administrative and financial resources. Thus, the paper assumes that democratic innovations linked to policies with greater technical, administrative, and financial resources (more institutionalized) are able to process more demands arising from social movements and civil society organizations. Therefore, their capacity of making decisions could be quantitatively larger and qualitatively better. By presenting more and better policy resolutions, they become spaces for deliberation. The other forms of engagement are processed by these deliberative spaces which "present the ability to transform civic and voluntary action in political and permanent forms of organization" (Avritzer, 2009).
To evaluate the relationship between more and less institutionalized patterns of participation, the paper uses the following methodological strategies. The first one is to classify each council based on Lavalle et. al (2016), who works with councils' decision-making (output), but not their effects on policies (outcome). According to them, this strategy "allows us to diagnose what the councils are actually doing and when they do; that is, what decisions they make and at what point in their life cycle" (Lavalle et al., 2016, p. 611).
The period of analysis is justified because it characterizes both the expansion of institutionalized arenas of participation, which were strongly supported by the federal government, as well as by the presence of many mobilizations, before and after 2013 (Galvão and Tatagiba, 2017).
The classification of the protests followed Galvão and Tatagiba's article (2017), which shows some characteristics that determine them. We use type of claim and repertoire of action as analytical categories. For the type of claim, we classify, as the authors, (1) salary demands; (2) demands for better working conditions (working day, safety, professional regulation etc); (3) identity demands (defence of minority group rights, culture, and identity); (4) social rights demands (better health, safety, care and mobility conditions, fare reduction etc.); (5) demands for development (infrastructure such as public sanitation, public transportation, hospitals, schools, housing, parks etc.), and (6)  The "health movement", formed in the late 1970s, was composed of academics, researchers, healthcare workers' unions, community leaders and professionals from the area. 13 Financial constraints did not allow us to have access to all events of protests published in the three newspapers from 2003 to 2016. The materials available to non-subscribers varied in each newspaper. In the newspaper "Hoje em Dia", they were available until 2011; in the newspaper "Estado de Minas" until 2012; and in the newspaper "O Tempo", until 2016.
14 Capital of the state of Minas Gerais, Belo Horizonte presents, according to IBGE (2011), a population of 2,3 mi inhabitants. It is the 3rd most populous city in Brazil, considering its metropolitan area. BH is mainly composed of adults, with a slight predominance of women (57%). Its main economic activity is commerce and services (59%) and its average per capita income was, in 2010, R$ 1,497.29. The city has a Human Development Index (HDI) of 0.810, considered "Very High" (IDHM between 0.8 and 1). 70.15% of the population aged 18 or over completed primary education and 54.16%, secondary education (SMSA, 2015).
They fought to build a national healthcare system that would offer universal and holistic access to healthcare for the Brazilian people.
The development of the national healthcare system was based on three principles: 1) constitutional guarantee of quality access to health services for everyone; 2) decentralization of public services to poor communities; and 3) guarantee of citizen participation and deliberation in health policies (Jacobi, 1989). To do so, the health system should be decentralized and organized in a hierarchical and decentralized way, with the participation of users, workers and health service providers (Côrtes, 2009).
In Belo Horizonte, the development of the Unified Health System (Sistema Único de Saúde -SUS) began in the early 1990s, when Mayor Patrus Ananias (Worker's Party) came to power (1992)(1993)(1994)(1995)(1996) and reorganized the city's administrative structure into nine administrative regions. Each region created its Regional Municipal Secretariat that comprised a defined geographical population around a political and administrative space. A Sanitary District was created in each secretariat, following the same territorial limits. The District corresponds to the administrative and healthcare organization of the Municipal Health Secretariat and it is responsible for organizing and managing SUS in the city. All Districts follow the principles that guided the national healthcare reform movement: decentralization of services, universalization of access to healthcare, and citizen participation in policy decisions. For each BHU, there is a Local Health Committee (LHC) that is composed of service users, workers, managers and the surrounding community. These committees are the entrance into the health participatory system in Belo Horizonte, since "all those affected" can participate.
From the LHC, the other health councils, located in different territorial scales, such as district and municipal, are composed of elected and appointed representatives (delegates). This participatory system, created by Municipal Law 5,903/1991, represents a starting point for the formation of coalitions involving representatives of service users, workers, managers and health service providers in a network that pressures the government representatives for more resources and better healthcare conditions (Wampler, 2015, p.154).

-Housing Policy
The elaboration and implementation of the National Housing Policy followed principles and guidelines whose main goal is to guarantee to the population, especially those with low income, access to housing. To do so, the policy requires the integration of both the housing policy and national policy for urban development. To solve the precarious Brazilian housing situation, it is also necessary to articulate cooperatively the three levels of the government, federal, state, and municipal.
According to the study "Housing Deficit in Brazil", in 2010, the country had a deficit of 6,940,000 dwellings, of which the Brazilian Southeast concentrated 38% of the housing deficit (2,674,000 units). Responsible for more than half the deficit in the region, the state of São Paulo had a deficit of 1,495,000 housing units, followed by Minas Gerais (557,000), Rio de Janeiro (515,000) and Espírito Santo (106,000) (FJP /MCIDADES).
Belo Horizonte has presented a housing deficit since the city was founded in 1897.
According to the Social Housing Plan, "favelas" have always been part of Belo Horizonte's landscape. Workers' families as well as former residents who were not reallocated in the planning of the city used to live in these precarious houses, deprived of any infrastructure In the 1970s, having to struggle against this scenario characterized by the housing deficit at the local and national levels, the social movement of urban reform emerged. Without proper housing and basic infrastructure, individuals and families gathered to fight for better living conditions. Similar to the health reform movements, the movement of urban reform started in São Paulo comprising professionals such as architects, urban planners, lawyers, and progressive sectors of the Catholic Church. It was organized around the right to land, the return to democracy and the establishment of infrastructure that would benefit the low-income inhabitants (Wampler, 2015).

-On the councils' decision-making profiles
Graph 1 shows that the Municipal Health Council (HC-BH), from the three councils analysed, is the most active in terms of decision making. The Municipal Housing Council is the second most active one, which presented 34 resolutions in the same period. Out of those resolutions, as Graph 1 shows, 50% are related to General Policy Definitions, 41% are related to Self-regulation, and 9% to Policy Oversight.
Finally, the Municipal Council of Rights for Women produced a total of 16 resolutions in the same period. Out of these, 44% concern Self-regulation, 37% concern Internal Procedures, and 19% concern Policy Management. Therefore, C1 is composed of territorially universalized councils, with high federal induction and well-defined and consistent actions. These characteristics explain the councils' decision-making profile which is, according to Lavalle et al (2016), closely linked to their institutionalization process.  In our research, type C1 is represented by the Health Council (BH), because it produces a greater number of resolutions, which mostly concern health policies and not only its internal functions. As mentioned above, Health is one of the most institutionalised policies, presenting a highly successful process of financial and administrative decentralization, with Councils in almost every city.
The second type (C2) is composed of councils with medium territorial expansion and moderate federal induction. Their organization into respective policy areas is moderate, which affects their performance. Its decision-making profile is characterized by a lower number of resolutions that involve issues related to internal organization and housing policy. Council, we investigated the protests related to each policy analysed during the same period (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016) in Belo Horizonte. The protests were classified according to the types of demands and the types of repertoires of action 15 involved (Galvão and Tatgiba, 2017).
Graphs 2 and 3 demonstrated the percentage of demands and types of repertoires of action involved in each policy.

Graph 2 -Types of demands about Health, Housing and Policy for Women protests in BH (2003-2016)
Source: Faria, 2018 Regarding the municipal health policy, a total of 165 protests were registered in the local newspapers from 2003 to 2016. As Graph 2 shows, out of those, 88.5% are protests related 15 According to Tilly (2006), repertories of action are a set of claims that have grouped, learned, but improvised character. That is, they are ways of acting, at the same time historically defined and constantly transformed by the creativity of the actors in specific situations (p. 35). and better salary (46.1%). The health workers' union is the main collective actor responsible for mobilizing these protests. Demands for social rights (6.7%); development (3%); politics (1.2%) and identity (0.6%) were also found, but with lower frequency.
Out of the 146 demands registered in the protests about housing policies, 94% are related to the development issue. Regarding this issue, there are claims concerned with construction and/or access to affordable housing and government measures to promote better conditions for population. Demands for social rights (5%), politics (3%) and identity (2%) also appeared with less frequency.
In policy for women, we gathered 46 protests whose great majority is related to social rights (82.6%). Demands collected about identities (6.5%) and political issues (4.3%), better working conditions and development (2.2%) were less frequent.  We identified 124 types of repertoires of action in the health protests. Strikes (33%) and demonstrations (33%) were the most frequent types, followed by sit-ins (26%). Other repertoires, such as marches, occupations, negotiations, and public classes were less frequent.

Graph 3 -Repertoires of action used in protest for Health, Housing and Policy for Women in BH (2003-2016)
With regard to the 146 types of repertoires of actions noted in the protests about housing policies, the most frequent were demonstrations (45.2%), followed by occupations (37.7%), and marches (11.6%). Public hearings (2.7%), negotiations (2.1%) and performances (0.7%) also appeared, although less frequently.
In policy for women protests, we identified 45 different repertoires of actions. Out of them, 40% are demonstrations, 40% are marches, 8.9% are cultural and artistic performances, 4.4% sit-ins, 4.4% occupations, and 2.2% public classes, as shown in Graph 3.

3.3-Analysing the cases
The data presented so far show that, in Belo Horizonte, social and political actors use different participatory strategies to achieve their demands. In do so, they form a "pluralistic regime of citizenship", which involves more or less institutionalised repertoires of actions.
Related to the public councils, a more institutionalised type of participation, our attention is drawn to the quantitative and qualitative differences in the outcomes of their decision-making processes. The Health Council presents a greater number of resolutions than the other two councils (Housing and Policy for Women, respectively) and the quality of the decisions is also very different.
The health resolutions are more concerned with the issue "definitions of the health policy", while the resolutions decided in the housing and policy for women councils are more closely concerned with their internal organizations. These data support the difference in the degree of institutionalization among the analysed democratic innovations already mentioned.
The same pattern is found when we analyse the demands and the repertoires used in the protests of each policy subsystem. The health policy is the subsystem that presents the greater number of mobilizations. Most of them involve demands concerned with the improvements of working and wage conditions. These demands, represented by the union movements, are generally not deliberated within the Health Council, although the workers and service users usually act as a coalition in the health subsystem (Cortes, 2009). According to Wampler (2015), the use of different repertoires in the health subsystem can be explained by the financial, administrative and political resources that it presents. The more resources the subsystem presents, the greater the dispute among the groups around them. Furthermore, the more pluralistic will be the repertoires of mobilized actions (p.202). In this case, the health workers' union (SINDSAÚDE-BH) is the main collective actor responsible for the mobilizations, whose most frequent actions are strikes and demonstrations. However, they do not act alone and are often supported in their actions by the leaders of civil society organizations (Wampler, 2015).
The housing subsystem presents the second largest number of mobilizations, whose demands are focused on development issues. Such demands are announced through demonstrations, as well as occupations and marches.
According to the actors who participate in this policy subsystem, the use of different The number of protests registered in the women's policy area confirms the low degree of its institutionalisation. The policy for women presented the smallest number of protests of the three policies analysed. Their demands are mostly concentrated on social rights claims, and the most common repertoires were demonstrations and marches.
Since women participate in almost every municipal council in Brazil (Faria, 2015;Wampler, 2015), there is no "women's absence" in the participatory policy subsystems in Belo Horizonte. Although, there is a problem related specifically to the process of institutionalisation of the policy system since the council presents low levels of federal induction, of territorial diffusion, and of organization within their policy areas. Even with the institutional induction promoted by the Secretariat of Policy for Women during the 2003-2016 period, the women's policy was not able to be institutionally organized.

Conclusions remarks
Brazil, until 2016, could be characterized by a regime where there was a plurality of arenas through which social movements and civil society organizations could publicize their material and cultural claims, and organize their actions and access to the political system. Some of these arenas, such as policy councils, composed by the representatives of civil society organizations, state and market, expanded considerably over the last thirty years. As a consequence, access to the state increased considerably. This expansion, whose climax was during the Worker's Party government, made scholars and activists believe in an institutionalist turn in Brazil's field studies on social participation.
This perception was interrupted by the new social actors and practices that invaded the Brazilian public scene since June 2013. They have forced us to review our analyses about that supposed institutionalist turn. The analyses of different repertoires of actions, mobilized by the actors to achieve their demands, was the main purpose of this paper that described and explained them as three policy subsystems in Belo Horizonte (MG).
The hypothesis supported that formal arenas (municipal policy councils), immersed in more institutionalized policy subsystems (those that present greater financial, technical and administrative resources) absorb more demands of civil society organizations. By presenting resolutions that are quantitatively greater and qualitatively better, they established arenas for deliberations that process more conflicts, thus preventing other repertoires of action to be mobilized. The degree of institutionalization of policy subsystem was our explanation for the difference in participatory patterns in the three policy subsystems analysed. Institutionalization of the demands through successful councils would prevent the other types of actor mobilization.
Based on the data, we showed that, although the health council produced the greatest number of resolutions, all of which consistent with their functions, this policy also presented the greatest number of protests. Though these protests are organized around issues that are not usually debated in the municipal health councils (work and salary conditions), health workers are frequently supported by service users of health system inside and outside the councils, forming a political coalition in this subsystem (Cortês, 2009).
The data show us that the health subsystem, the most politically and financially institutionalized Brazilian subsystem, uses, notwithstanding, different participatory patterns to push its demands and guarantee healthcare in Belo Horizonte. This finding contradicts our initial hypothesis. It shows us that mobilization and institutionalization work together in this policy.
Housing policies added more empirical evidence that two patterns of action are combined. The social actors in this subsystem practice institutional and extra-institutional patterns of action. According to the participants, to participate in Housing Council is not a problem, since it is "a space to be occupied". The problems are the executive and legislative powers and their limits of action. In this case, we have a policy subsystem that presents an expressive number of protests, whose demands are related to the most discussed themes in the Housing Council. The actors use different repertoires of action to achieve their ends, but demonstration and occupation of urban lands are their most frequent choice.
The third subsystem analysed, the policy for women, is the least institutionalised subsystem, whose council is less structured. It produced fewer resolutions, almost all related to the Council's internal organization. Additionally, the policy for women comparatively presented much fewer protests. According to the proposed hypothesis, given the low degree of institutionalization in that policy, without its own budget, technical and administrative resources, greater activism was expected, especially from 2003 to 2016, when women's policy gained prominence in the national political agenda and was highly stimulated at subnational levels (Faria, 2015).
The data showed why women's organizations criticized the formal participation in the Municipal Council of Rights for Women, but not why they do not engage more in the contentious repertoires. It is possible to say that the lack of institutional organization of the women policy in Belo Horizonte also compromises women's activism. This absence can be explained by both, the institutional resistance to women's demands and/or their transversality.
Other policies may contemplate their issues, such as women's healthcare programs. Even so, the data show us that there is no policy for women's subsystem in Belo Horizonte and their contentious repertoires are less expressive when compared to those mobilized by the two other policies. Considering the historical trajectory of the movements in each policy, the degree of institutionalization of each subsystem can be an important explanatory variable for these variations. As showed by the health subsystem, the more institutional resources they conquered, the more conditions for mobilization they present, even in a time of supposed institutional turn.