Multilevel Governance of Social Policy: Social Services in Metropolises^*,**

Abstract

This article looks into the explanatory dimensions of social services governance in São Paulo and Belo Horizonte. It considers the vertical axis of the interconnections between levels of government and the horizontal axis of interactions between state and non-state actors. We argue that it is necessary to advance in integrated explanations on how these multilevel interactions restructure local governance patterns over time in order to fill a gap in the literature on multilevel governance. Cases are compared based on the triangulation of methods: literature review, document analysis, and in-depth interviews. We argue that four dimensions are essential to understand the effect of the intersection between axes: 01. municipal political organization; 02. mobilization of policy communities; 03. civil society actors’ profile and work; 04. types of state capacities. Our case analysis reveals, in São Paulo, a pattern that is less constrained by federal rulemaking and more focused on conflicts between civil society organizations (CSOs) and bureaucratic actors, and highly dependent on local political dynamics and ways of building regulatory capacities. In Belo Horizonte, in turn, we observed a greater influence of federal rulemaking and greater weight of policy communities, which are constantly mobilized to build capacities for policy provision, and not only for the regulation of services.

Multilevel governance; social services; social policies; SUAS; metropolises

What explains social services policy governance in metropolises? This is this article's guiding question, as it engages with the international literature on multilevel governance and with the Brazilian discussion on federative coordination of social policies. The concept of multilevel governance aims to understand the changes in the public policy-making process resulting from different types and paces of decentralization in terms of decision-making autonomy (vertical axis) and the incorporation of a growing set of different actors in these decisions (horizontal axis) beyond the state (HOOGHE and MARKS, 2003; TORTOLA, 2017).

Based on the assumption that it is important to analyze the subnational level even in policies regulated at the national level, we aim to specify how the vertical axis (federal regulation and the multiple interrelationships between levels of government) and the horizontal axis (the interactions between state actors and civil society organizations (CSOs) involved in the provision of basic social services) affect the municipal governance of this policy, whether by inducing or constraining it. To this end, we looked into the cases of São Paulo (SP) and Belo Horizonte (MG) as observation units. Our goal is to discuss how legacies and battles waged between state and non-state actors that unfold at the municipal level are modified through federal rules that structure the Brazilian Unified Social Services System (SUAS), as well as defined by local battles and choices.

In analytical terms, we aim to fill a gap identified in the international literature (TORTOLA, 2017): to discuss the specific effect of the intersection between the vertical and horizontal levels of multilevel governance, beyond the isolated characterization of each of these axes. There are still few studies that address developing countries using the multilevel governance framework, with some exceptions, such as Deshpande, Kailash, and Tillin (2017) on India, Kriisk (2019) on Estonia, and Meza (2021) on Mexico. In Brazil, this is also an incipient agenda, as seen in the works of Arretche (2016) and Bichir, Brettas, and Canato (2017). Additionally, we aim to go beyond the characterization presented by Arretche (2012) about the ways of federative coordination of social policies in Brazil, where policy decision-making is relatively concentrated in the federal government while municipalities are in charge of policymaking. This is because there is subnational decision-making autonomy even in nationally regulated policies, particularly in complex contexts and with greater concentrations of relative institutional capacities such as metropolises.

We adopt an analytical, non-prescriptive use of the concept of governance, according to the approaches presented by Marques (2013) and Cavalcante and Pires (2018): it is about understanding and analyzing patterns of interaction between state and non-state actors that occur in formal and informal arenas — mediated by public policy instruments and institutions —, and not about advocating for supposedly more appropriate arrangements to obtain certain desirable results. This concept implies not only going beyond the formal arrangement that organizes the policy. We argue that it is necessary to incorporate the time dimension into the analysis, to understand how historical legacies are constituted and battled over, and how they affect the correlations of forces between major relevant actors. In the specific case of social services, we argue that it is essential to look into how social services actors' margin of autonomy is affected ‘before and after’ this policy was federally regulated. A relevant part of the federal regulation derives from trailblazing municipal experiences — that is, it is not an autocratically defined decision at the federal level, but there are processes and mechanisms of dissemination in action, which underscores how important it is to look into the intersection between axes.

After the enactment of the 1988 Constitution and of the Organic Law on Social Services (LOAS), relatively generic guidelines with low inducing power guided this social policy, leaving great room for municipal experiences of implementation. After the creation of the Unified Social Services System (SUAS) —that is, after the enactment of the National Social Services Policy (PNAS), in 2004 —, federal rulemaking begins to constrain this local autonomy. As we will see and have already discussed in other writings (BICHIR, SIMONI JR., and PEREIRA, 2020), while these nationally defined benchmarks established since 2004 have different levels of power to induce local decision-making processes, they constitute a new layer of rules that are now considered in the interactions between local actors. We thus argue that it is necessary to look into how vertical and horizontal dynamics explain the social services policy-making pattern at the municipal level, constituting and modifying the power resources of state and non-state actors and key institutions in this field.

So why look into metropolises, considering that most municipalities in Brazil are small in size? The specific literature on social services policy shows that there are specific challenges facing the implementation of the SUAS in metropolises. Service provisioning legacies from civil society entities, generating unequal and fragmented capacities; size and diversity of vulnerable populations; heterogeneity of publics and local territories; fierce battles over budgets with other departments; dynamics of urban violence that affect the work of frontline bureaucrats, and others (BICHIR, BRETTAS, and CANATO, 2017; SÁTYRO and CUNHA, 2019; SILVA et al., 2012). While in the beginning of the implementation of the SUAS metropolises recorded worse indicators regarding service provisioning, recently there has been a greater convergence with cities of different sizes, at least in terms of structuring state capacities. Moreover, while in small municipalities, there is less variation in the models of provision of basic social protection, with a predominance of direct provision in public facilities; metropolises tend to have much more varied local governance arrangements. Metropolises are thus an analytically interesting environment to test more general explanatory hypotheses about the ways an increasingly wide diversity of actors involved in policy-making interact.

The case studies that provide the basis for this article were designed based on an initial hypothesis that proved to be simplistic: São Paulo and Belo Horizonte, trailblazing cases building dimensions that were later incorporated (whether partially or fully) into the national social services policy, were expected to have some comparative advantage in the period after the policy was regulated at the national level. Some of these experiences include the territorial and decentralized dimension of administrative organization, in both cases; the implementation, in São Paulo, of programs that conceived integrated services for vulnerable families¹, a pillar of what later became the Comprehensive Family Protection and Assistance Service (PAIF), a structuring strategy for basic social protection at the national level; the creation, in Belo Horizonte, of Family Support Centers (NAFs), public facilities that may be considered predecessors of the Social Services Reference Centers (CRAS). As our analysis advanced, however, it showed that, after the consolidation of the SUAS, a number of modifications and adaptations started to occur in these local contexts, with different (and sometimes negative) consequences for local policy provision patterns. These local patterns, as we will see, are affected by political dynamics and the history of policy structuring, as well as other explanatory factors outlined in the next section.

The time axis of our analysis begins with the first national and subnational regulations of this policy, spanning from the 1988 Constitution to 2020, to cover the regularity of the processes before the COVID-19 pandemic. The two case studies are based on a triangulation of methods and techniques: literature review, document analysis (legislation referring to the three levels, regulatory acts, rules and regulations), and around 25 in-depth interviews conducted with mid- and high-level public managers, actors connected to the legislative branch, actors who hold seats in municipal committees dedicated to this policy, former federal social services policy managers, and representatives of CSOs and social movements connected to the social services policy, in either or both cities.

The article is organized into four sections in addition to this introduction. In the first section, we present the analytical multilevel governance framework and its adaptation to the context of the Brazilian discussion. In the second section, we summarize the main milestones of the SUAS that exert, in different ways, inducing effects on subnational behaviors without, however, eliminating municipal decision-making autonomy to organize their social services policy management. In the third section, we present the main aspects of social services governance structuring in São Paulo and Belo Horizonte, aiming to demonstrate how local battles are transformed/mediated by federal macro regulations, and when local dimensions overlap the influence of central rulemaking. Finally, we synthesize our main comparative findings and address future research agendas.

Multilevel governance of social policy: lessons from the international discussion and specific Brazilian characteristics

The polysemic and multidisciplinary conversation about multilevel governance has developed since the 1990s in developed countries, notably in Europe and later in the US. In its interface with the discussion about federalism and decentralization (HOOGHE and MARKS, 2003), it is focused on the dynamics of relations between central and subnational levels of authority (vertical axis), as well as the patterns of relationship and interaction between state and non-state actors in the production of public action (horizontal axis) (KAZEPOV, 2005; KAZEPOV and BARBERIS, 2013), in a process permeated by techniques, tools, and instruments (LASCOUMES and LE GALÈS, 2007). This literature provides a more granular lens to consider the multiple axes of transformation of the state and the processes of public policy-making. While, on the one hand, the transformative and democratic potential of the local level is celebrated — such as in Sellers et al. (2020) —, other analyses are very conscientious when discussing the effects of decentralizing reforms and their potential effects (MINAS et al., 2018). There are also studies that aim to define specific indicators of local autonomy, without assuming that policy-making at this level would naturally be more appropriate or responsive to citizen demands (BRUCH and WHITE, 2018). Studies concerned with public policy-making processes, and not with the characterization of state reforms, tend to combine the lens of multilevel governance with models of public policy analysis, as advocated by Tortola (2017) and carried out by Meza (2021). By combining the multilevel governance lens and the punctuated equilibrium theory, Meza's work (2021) demonstrates, in the Mexican case, how decentralizing reforms managed to disrupt lasting balances at the municipal level and thus expanded the social policy agenda at this level of government (Meza, 2021).

We have a very specific position regarding this broader discussion on multilevel governance. A first analytical shift follows the path suggested by Arretche (2016): we aim to disaggregate the analysis by public policy area, as decentralization trajectories may vary significantly, and we do not seek to frame Brazil as a whole in a general and comparative typology. We thus move away from multilevel governance approaches that aim to classify national social policy reforms and changes in the role of the state into typologies. A second shift follows Tortola (2017): we use the lens of multilevel governance as an explanatory model for public policies, and not as a new theory about state transformations.

We therefore aim to combine the multilevel governance lens and some concepts of public policy analysis, in order to consider both the ‘vertical axis of the multiple interrelationships between levels of government’ — federal, state, and municipal, in the case of Brazilian federalism—and the ‘horizontal axis’ of interactions between state bureaucracies, civil society organizations (CSOs), social movements, legislators, and others. We argue that there is a need to further specify the effects of this intersection between the vertical and horizontal axes of governance, over time, in order to produce integrated explanations about how these multilevel interactions restructure local governance patterns. This is a gap in the international literature on multilevel governance, which tends to favor one axis or the other in empirical analysis, therefore underestimating the analytical potential of the term and restricting it to an ‘umbrella concept’ with a more descriptive use. We thus aim to specify when and how local processes and experimentations become repositories of practices and regulatory instruments that then gain scale and power of enforcement at the federal level. On the other hand, even when national parameters gain strong inducing power, this vertical influence is mediated by local processes and battles. In summary, while inducing efforts at the federal level displace the subnational game and change correlations of forces between actors, they do not explain in isolation what occurs at the subnational level.

Our arguments refer to a nationally regulated social policy — social services —, in which civil society organizations (CSOs) are key. CSOs are involved not only in the provision of social services and assistance, but sometimes gain prominence in formal and informal decision-making arenas, even helping to guide how they will be regulated (BICHIR, PEREIRA, and GOMES, 2021; BRETTAS, 2016). In this sense, theoretical references arising from the discussion on coproduction (in an analytical — not prescriptive — sense, as partially present in OSBORNE et al., 2016 and CONTEH and HARDING, 2021) and on mixed-provision of welfare (such as EVERS, 2005 and YANG et al., 2019) are also incorporated and ‘filtered’ from their prescriptive and normative tone through the excellent contributions of Gurza Lavalle et al. (2019) on the mutual constitution between the state and civil society. Particularly, we have made progress in specifying what ‘types of state capacities’ may eventually emerge from these interactive processes, separating the dimensions of ‘direct provision’ of services and the ‘regulation’ of this provision.

Because we consider, as Marques (2021), that cities in the ‘Global South’ can and should be compared with other cities in the world, as they do not have any intrinsic exceptionalities that impose exclusive analytical categories onto them, we also aim to contribute to the literature on the effects of the local level on national social welfare systems. Local levels of government draw renewed attention from studies on multilevel governance and from the specific discussion on social protection systems (such as the works of KAZEPOV and BARBERIS, 2013; SELLERS et al., 2020; SELLERS and LINDSTRÖM, 2007). Local institutions, however, are increasingly inserted in broader dynamics, as noted by Sellers et al. (2020). When selecting a social policy that can be characterized, in the terms of Kazepov and Barberis (2013), as a ‘local autonomy-centrally framed’ case, it is not enough for us to specify subnational levels’ space of decision-making autonomy, but rather to understand how they are induced and constrained by central macro regulation processes. Moreover, we aim to understand how they contribute to the historical consolidation of certain pillars of national policies, particularly when ‘successful’ local experiences undergo a process of vertical dissemination and scale up, either via migration of policy communities (Belo Horizonte’s case) or via occasional participation of certain policy entrepreneurs who have strong political and professional networks with top-tier decision-makers (São Paulo’s case).

We also aim to advance in relation to the contributions made by Arretche (2012). Since Bichir, Brettas, and Canato (2017), we have argued that the division between ‘policy decision-making’ and ‘policy-making’ underestimates the several decision-making chains, with broad and multiscale consequences, that occur at the implementation level — a dimension considered in studies on implementation (SAETREN, 2014; WINTER, 2006), but which are relatively ignored in the Brazilian discussion about federative coordination. Since that work, we have advanced in the specification of subnational policy-making regimes that operate even in the case of nationally regulated policies. In other words, taking the local level seriously even when looking into social policies regulated at the macro level implies demonstrating how and why this policy-making game does not begin or end with the federal government. Ideas that have been tried and tested locally can spread nationally, actors who gain legitimacy in the local game sometimes accumulate resources and instruments to influence the national game, policy communities can be taken to the top national decision-making level, depending on the dynamics of the political process over time. In this article, the analysis of the intersection between the vertical and horizontal axes is directed to the municipal level to consider how these multiple games organize battles that involve ways of translating social services into public policy. As we will see in the next sections, state and non-state actors battle over how governance arrangements will be made to translate different political projects and perspectives about the centrality of the state or CSOs in service provision, programs, actions, and benefits to the most vulnerable populations.

We thus argue that four key dimensions organize social services governance over time in these two cases:

1. Municipal political organization: in line with the assumptions of the theory of party government, the more progressive or unprogressive profile of municipal representatives affects the political weight social services will have on the agenda and in the municipal budget². In the cases we investigated, one metropolis has a mostly conservative profile (São Paulo) and one is a politically progressive metropolis (Belo Horizonte). Additionally, the profile of the municipal secretary appointed by the mayor for this department also affects how the policy will be conducted.

2. Organization of policy communities and municipal bureaucracy: the ways of structuring the municipal bureaucracy — not only through structured careers, but also with greater or lesser connection with the social services community (networks of specialized actors who share views and projects for the area, in more or less closed patterns of connection, in the terms of Côrtes, 2015) — are expected to affect department governance. As we will see, in São Paulo, there is a prevalence of political entrepreneurs who form more of an epistemic community for national battles³ than a policy community, and there is an absence of a career-based structured bureaucracy. Belo Horizonte, in turn, has career-based patterns, relatively regular civil service entrance examinations, and a locally and nationally active policy community.

3. Profile of civil society actors:the profiles of the CSOs involved in the provision of services and of the social movements that organize the demands for services, as well as the way they work, is fundamental in an analytically oriented perspective of governance. While, in São Paulo, there is a prevalence of a relatively diversified CSO ecology (in terms of organizational size, denomination, or degree of professionalization before or after the SUAS) involved in the provision of different types of services, Belo Horizonte also has social movements that are key in the constitution of spaces for participation and in the mobilization of social demands toward the government.

4. Types of subnational state capacities: in municipal-level social services policy-making, it is essential to differentiate the focus on the state's capacity to directly provide services (case of BH) or on the regulation of the provision of services hired through CSOs (case of SP). These two capacity dimensions are combined in different ways and lead to the creation of public policy instruments (in the sense defined by LASCOUMES and LE GALÈS, 2007) that affect the correlation of forces between relevant actors.

There are certainly interactions between the aforementioned dimensions, as it is not a matter of building an argument based on linear causality, but rather of identifying, in each of these cases, ‘how’ the combination of these elements helps to understand ‘explanatory configurations’ organized by the vertical and horizontal axes. The multiple possible combinations between these dimensions help to ensure variability of conditions and empirical substrate to analytically advance in a still under-explored field in the international discussion on multilevel governance: when, how, and why the interrelationships between federative dynamics and local governance battles affect the local game.

In this proposed model of analysis, two concepts are worth highlighting. The first is the public policy community, understood here as a limited and relatively stable number of members who share the same values and views about what the results of a sectoral policy should be. This group builds alliances between societal and governmental actors in order to influence decision-making processes, considering the most adequate alternatives for solving problems based on their values and views about the desirable results for the policy (CÔRTES, 2015, p.132); in this case, it is about building policies through different instruments and with a greater diversity of actors involved. Meanwhile, the concept of epistemic community is used to address actors specializing in an area of knowledge who have the legitimacy to operate in a public policy, implying more restricted networks and higher entry costs, in which expertise, know-how, and knowledge are the key dimensions (HAAS, 1992).

The second concept is state capacities, understood as the state’s resources, skills, and competencies in the establishment and implementation of its objectives (PIRES and GOMIDE, 2018). Specifically, we understand that state capacities can be coproduced in interactions between state and non-state actors, as highlighted by other authors (LAVALLE et al., 2019). Moreover, we move forward by specifying that two types of capacities are important to understand the investigated cases: provision capacity (direct provision of services) and regulatory capacity (creation of instruments to regulate the indirect network).

The structuring of the vertical axis: the inducing power of the SUAS

After the enactment of the Brazilian Constitution in 1988, social services began a process of transition to the paradigm of rights, with the recognition of the broad concept of ‘social security’ (‘seguridade social’) based on three pillars —health, social services, and social security (previdência social). However, despite the regulatory advances represented by the 1988 Constitution and the LOAS, enacted in 1993, the building of state capacities and the specification of competencies and responsibilities starts in the Fernando Henrique Cardoso administration (1994-2002) and continues through to the Luiz Inácio Lula da Silva administration (2003-2010), with the definition of the pillars of the SUAS and national policies. In addition to the development of mechanisms for inducing and coordinating the policy (BICHIR and GUTIERRES, 2019; BICHIR, SIMONI JR., and PEREIRA, 2020). In the 2000s, starting in the Lula administration, significant advances were made in terms of the institutionalization of social services as a public policy, with the construction of the pillars of its systemic arrangement. In this institutional construction, the National Social Services Policy (PNAS), from 2004⁴, the SUAS Basic Operational Rule (NOB-SUAS), from 2005⁵, and the National Classification of Social Services (CNAS Resolution Nº 109, from 2009) are worth highlighting⁶. The so-called SUAS Law (Law Nº 12,435) from 2011, in turn, modernized the LOAS and consolidated the pillars of the SUAS.

The SUAS establishes two types of social protection for social services action: Basic Social Protection (BSP) and Special Social Protection (SSP). BSP aims to prevent social risks facing families and individuals in situations of vulnerability and social risk, having the Social Services Reference Centers (CRAS) as its main facilities. On the other hand, SSP aims to help specific publics who are already at social risk and have had their rights violated, such as homeless people or children who are victims of sexual exploitation or child labor. The Specialized Reference Centers for Social Services (CREAS) are the main facilities focused on implementing SSP. Furthermore, SSP is divided into two levels of complexity: medium complexity, which provides care for families and individuals whose family and community ties are not broken, but their rights have been violated; and high complexity, which serves people who are removed from family and/or community life, ensuring their full protection, including housing, food, hygiene, and protected work.

From the point of view of the vertical axis of multilevel governance, intergovernmental relations within the scope of the SUAS are currently based on the following divisions of competences: the federal government is responsible for the general regulation of the policy, for the transfer of federal resources to programs classified as such, the definition of the types of social services that may or may not be co-funded via fund-by-fund transfers (having federal, state, and municipal governments as coparticipants), macro strategies for monitoring and assessment, and the design of the strategy to train bureaucracies at the subnational level. While state governments are co-responsible for funding the policy, providing training and regionalizing services, especially medium and high complexity services, they play a very limited role in the direct provision of services, except in emergency situations and in cases of their own programs that are not necessarily coordinated with the SUAS. Municipal governments, in turn, also co-fund the policy and are responsible for the implementation of services at both the basic and special protection levels, depending on their level of abidance to the system (JACCOUD et al., 2017). There is also an arena of federative agreement so that subnational managers can take part in decision-making processes even when the agenda-setting power is concentrated in the federal government, as Costa and Palotti (2011) skillfully demonstrated in the case of the SUAS CIT (Tripartite Inter-Managers Committee), similar to the CIT of the Unified Health System (SUS), which includes managers from the three levels of government and the CIB (Bipartite Inter-Managers Committee), with municipal, state, and Federal District managers⁷.

From the point of view of the horizontal axis of multilevel governance, the SUAS is based on a social services and assistance network that combines direct and indirect provision of services, under the responsibility of civil society organizations (CSOs) that enter into agreements with the government for this purpose. These CSOs hold an important place regarding the policy, not only for offering different essential services, both in terms of BSP and SSP, but also for operating as important actors in the building of regulations that govern the SUAS, making up decision-making spaces such as the National Social Services Council. Despite the political battles around the meanings of public responsibility for the social services policy and the place of the third sector in this policy, it is possible to note that the regulation of CSOs involved in service provision has increased significantly since the 1988 Constitution, a dimension that is still not considered very often in the national literature (BICHIR, BRETTAS, and CANATO, 2017; BRETTAS, 2016). Moreover, we have an institutional architecture formed by popular participation councils at the three levels of government, having managers, users, and workers as their representatives, as well as forums and social movements that represent important segments of social services policy users. Agencies, actors, and institutions dedicated to related policies, or which address publics or similar topics, are also connected to the social services policy, such as councils on children and adolescents, councils on older adults, public prosecutor's offices, and institutions of the justice system and for the guarantee of rights.

Federal regulation also constrains municipal autonomy through a budgetary governance arrangement (and not only volume/availability of resources) and specification of benchmarks for the organization of human resources (NOB-RH) and ways of organizing services through the SUAS Census⁸ and the IGD⁹. It is therefore a ‘local autonomy centrally framed’ policy according to the multilevel governance typology by Kazepov and Barberis (2013).

Through the consolidation of this SUAS institutional arrangement, inducing and catalyzing effects of subnational dynamics are generated. While it is undeniable that there have been advances in terms of building state capacities — not only at the federal level, but with significant dissemination at the municipal level (BICHIR, SIMONI JR., and PEREIRA, 2020) —, we also know that it is not enough to have federal inducing efforts and regulation. The ways of translating and specifying these general dimensions according to agendas, legacies, and local battles make a difference and produce ‘subnational regimes of governance’. These aspects are explored in the next section.

The horizontal axis: the transformation of legacies and new conflicts in São Paulo and Belo Horizonte

While São Paulo and Belo Horizonte are different cases, they are analytically interesting to understand the scope of autonomy in the organization of social service provision, even as national regulatory frameworks defined by the SUAS are in force. In this section, we introduce some of the main structuring elements of local governance patterns according to the analysis axes introduced above and considering the period from the democratic transition to 2020. That is, we did not address the discussion on the restructuring of services in the context of the COVID-19 pandemic. Also, the detailed presentation of each case is not the purpose of this article.

As a starting point to consider the comparison between the cases, Table 01 below shows characterization data of the two service networks and budgetary data in a historical series. São Paulo’s facility network is more extensive and diversified, with a complex territorialization, and most of the services are provided through agreements established with CSOs. In Belo Horizonte, there is greater weight to its network of public facilities, even though its indirect network, which is more concentrated in special social protection services, is not insignificant.

Table 01
Demographic characteristics, social services and assistance network in São Paulo and Belo Horizonte

Characteristics	São Paulo	Belo Horizonte
Demographic characteristics
Total population	2010: 11,253,503 2019 (estimated) 12,252,023	2010: 2,375,151 2019 (estimated): 2,512,070
GDP per capita [2017]
Extremely poor % [2010]	0.92	0.41
Poor % [2010]	4.27	2.12
Characteristics of the social services and assistance network
Regional Division of the Services (Territorialization) (2022)	32 Social Services Oversight Offices, which have the same regional design as the subprefectures.	09 Regional Social Services Boards, which have the same regional design as regional administrations
Number of public facilities (basic, high complexity, and homeless population) (2022)	54 CRAS, 30 CREAS, and 06 Pop Centers	34 CRAS, 09 CREAS, 03 Pop Centers, 09 Regional Basic Protection Services[10]
Approximate % of services directly provided by the state (2022)	06% Direct Management	36% Direct Management
Approximate number of partnership-based services (2022)	94% Indirect Management Around 1200 partnership-based services	64% Indirect Management 122 partnerships.
Number of public servants in social services departments	926 (SIGPEC — Dec/2020)	770 (Permanent Education Plan/SUAS/2018)

• Source: Elaborated by the authors based on data from IBGE, UNDP, Municipal Social Services Plans from São Paulo and Belo Horizonte, Integrated People and Skills Management System/SP, Permanent Education Plan/BH.

In terms of municipal political organization, it is worth noting the contrast between the two metropolises. Regarding the ideological profile of their mayors, in one of the cities — São Paulo —, center-right mayors prevail, while in the other, BH, its mayors are more progressive leaning. São Paulo had the following mayors from the late 1980s to the recent period: Luiza Erundina (1989-1992, PT, Left); Paulo Maluf (1993-1996, PPR/PPB/PP, Right); Celso Pitta (1997-2000, PPB/PTB, Right); Marta Suplicy (2001-2004, PT, Left); José Serra (2005-2006, PSDB, Center-Right); Gilberto Kassab (2006-2012, then DEM, currently PSD, Center-Right); Fernando Haddad (2013-2016, PT, Left); João Doria (2017-2018, PSDB, Center-Right), and Bruno Covas (2018-2020, PSDB, Center-Right). In 31 years, left-wing administrations were in office for only 12 nonconsecutive years. In Belo Horizonte, in turn, there is a greater line of continuity of progressive mayors: Eduardo Azeredo (1990-1993, PSDB, Center-Right); Patrus Ananias (1993-1996, PT, Left); Célio de Castro (1997-2000 and 2001, PSB, Center-Left); Fernando Pimentel (2001-2004 and 2005-2008, PT, Left); Marcio Lacerda (2009-2012 and 2013-2016, PSB, Center-Left); Alexandre Kalil (2017-2020, PHS and then PSD, Center-Right). Therefore, there have been twenty-three years of continuity of left-wing parties in Belo Horizonte.

Historical legacies and their transformation Post-1988 and Pre-SUAS

Turning a field of erratic interventions, guided by charity, benefaction, and poor relief (‘assistencialismo’), into continued actions guided by the right to a public policy was not an easy task in either of these two contexts. However, these two metropolises have been marked by quite different political choices and institutional designs in this initial moment that goes from the enactment of the 1988 Constitution to the early 2000s. In the absence of national frameworks with strong inducing power, this is a time with great municipal autonomy and broad room for innovation. In addition to the political dynamics itself, different choices in terms of consolidating or not social services bureaucracies and careers — which happened in Belo Horizonte and did not happen in São Paulo — made a lot of difference from the point of view of structuring actions.

In São Paulo, pioneering experiences in the translation of constitutional precepts began under Luiza Erundina (PT). Programs and services managed by the direct public administration were prioritized and a working group was created to establish minimum parameters for the indirect provision of services through CSOs (AMÂNCIO, 2008). However, these first efforts to institutionalize the policy barely had time to thrive, as the right-wing administrations that followed were unwilling to promote changes in the sector. The administrations headed by Maluf (PPR/PPB/PP, 1993-1996) and Pitta (PPB/PTB, 1997-2000) maintained the institutional fragmentation of actions, guided by ‘emergency social services’, and not by continuous actions and actions dedicated to preventing risks and vulnerabilities (YAZBEK, 2004). São Paulo was the last capital city to implement the LOAS, and some of the instruments and institutional spaces recommended in the organization of the policy were vetoed — such as the case of the Municipal Social Services Fund (FMAS) — or modified to reduce the room for civil society to conduct their work —such as changing the equal representation of the Municipal Social Services Council (COMAS)¹¹. However, after pressure from organized civil society and key social services actors, the Municipal Social Services Forum (FAS) began its activities in 1993, as an important space for coordinating different demands by non-state actors in this field (YAZBEK, 2004).

It is only during the second administration run by the Workers’ Party (PT) in the city — under Marta Suplicy — that an institutional structuring of social services as a policy begins, with the trailblazing development of pillars that later structured the SUAS. In addition to starting the effective operation of fundamental levels of the SUAS, such as the council and the municipal social services fund, the Suplicy administration chose to mobilize policy community actors at the department’s top decision-making level after the second year of Suplicy's term, appointing Aldaíza Sposati, a professor at PUC-SP, former city councilor, and important name in the sector’s policy community. Her profile allowed connections to be established with top decision-makers in Brasilia, as she was an entrepreneur of ideas from this area and an actor with political influence in the Social Services Sector of the Workers’ Party¹² (GUTIERREZ, 2015). As discussed in Bichir, Brettas, and Canato (2017), it is after this administration that important co-policy-making processes start, organized in complex multilevel governance patterns: from a vertical point of view, there is an alignment between national and local parameters, with the necessary adaptations; from a horizontal point of view, negotiations with civil society actors responsible for the provision of services are accentuated, having the COMAS as a major arena of battles and having, as a result, the definition of municipal parameters for the regulation of agreements between state and civil entities (MARIN, 2012). Some pioneering experiences tested in this administration were later incorporated as principles of the national social services policy. On the other hand, this municipal administration was characterized by the institutional fragmentation of income transfer programs and other benefits and services scattered among different departments.

In Belo Horizonte, there is a more favorable political and institutional starting point. From the point of view of the politics, the analyzed period begins with the Pimenta de Veiga/Eduardo Azeredo administration (PSDB)¹³, having a technocratic and emergency-driven perspective to social services, characterized by the management of agreements with social entities, without cost studies and continuous automatic transfers, and by the absence of a structured policy with programs, projects, human resources, and budget. We then have a coalition of left-wing parties — the Popular BH Front (1993-2003) —, which provided fertile ground for experimenting with social services as a public policy. Unlike São Paulo, Belo Horizonte was already characterized by the active organization and participation of civil society in the state structure even before the enactment of the Brazilian Constitution in 1988. There were also important institutional milestones for social services before the LOAS. For example, the Department for Social Development was created in 1989 from the merger of the departments of Community Action and Social Welfare, and the structure remained in place until its administrative reform in 2000.

The first administration ran by the Popular BH Front was headed by Patrus Ananias (PT, 1993-1996), who became known for consolidating a decentralized municipal model and for significantly expanding civil society participation, particularly in the implementation of the Participatory Budgeting. Several programs to serve specific publics were created or restructured, people’s participation was encouraged, several mechanisms for participation were improved (especially through Local Committees and regional and municipal councils), and there was expansion and professional training provided to the sector's personnel.

In the field of social services, the head of the then Department of Social Development was André Quintão¹⁴ (PT, 1993-1995), and its assistant secretary was the social worker Márcia Pinheiro, known for having participated in the LBA (Brazilian Aid Legion) and for being extremely active in the area — and later for heading the National Social Services Department (SNAS/MDS) and chairing the National Social Services Council in Brasília. Actions targeting specific publics, such as homeless children, were high on the agenda, such as the Miguilim Program, which also remained a top-priority agenda in the administrations of Célio de Castro, who was Patrus Ananias’ vice-mayor and took office as mayor afterwards. Moreover, the first civil service entrance examination for social workers was held in 1994, which drew in several professionals trained at PUC-MG who were also connected to other municipal administration experiences, PT activists, and several social movements — advocating for childcare centers, food and nutritional security, and others. There was, therefore, a concern to structure careers as an important element for the continuity of the area.

The basic social services institutionalized was developed during the Patrus Ananias and Célio de Castro administrations. In 1995, the Belo Horizonte Municipal Social Services Forum began to be organized, as an important space for the agglutination of the social services policy community that brought together workers, entities, and civil society actors. The Minas Gerais Forum included PT social workers, AMAS, Caritas, ASPRON, the Pastoral Homeless Commission, the Pastoral Human Rights Commission, and social workers. In 1995, the first Municipal Conference on Social Services was held, bringing together more than 2000 people. The project that provided the basis for building the BH LOAS in the following year was formulated during this conference. Some important pillars that eventually led to the SUAS were established in 1996, in the wake of the regulation of the LOAS in the municipality. Municipal Law Nº 7,099, of May 27, 1996, established the Municipal Social Services Fund (FMAS) and levels of social participation such as the Municipal Social Services Council (CMAS), the Regional Social Services Councils (CRAS), and the Local Social Services Councils (CLAS). Unlike other municipalities that only have a municipal council, Belo Horizonte has regional and local participation councils, which are only advisory and not deliberative in nature. These spaces were important participation mechanisms pre-SUAS.

By the end of Célio de Castro’s first term (1997-2000), specifically in 2000, a process of administrative political restructuring of the Belo Horizonte city government began, having intramunicipal decentralization, participation, information, and intersectorality as main structural pillars — which were connected with the agenda that was being built for social services in the municipality. One department, the Municipal Department for Social Policy Coordination (SCOMPS), was created to coordinate all social areas, including: Education, Social Services, Health Care, Sports, Culture, Supply, and Citizenship Rights. These areas were granted the status of attached departments allocated at the same hierarchical level and subordinate to the SCOMPS. The social services area thus became a subdepartment and was included as part of a broad agenda of interventions with intersectoral guidelines.

This intersectoral agenda was reinforced in the Fernando Pimentel (PT) administration, which also promoted relevant advances in the sector. In 2002, important structures for service provision were created: the Family Support Centers (NAFs), with a regionalized logic — one NAF for each of its nine administrative areas. This administration also promoted advances in the discussion about the regulation of the private network, the separation of childcare centers from social services to education; it removed the notary responsibility for granting certificates from the councils; a monitoring and assessment system and assessment indicators were created. All these trailblazing actions served as a model for the formulation of national guidelines, particularly the NAFs that served as a model for the creation of the CRAS at the national level in 2004. Another highlight is that, in Belo Horizonte, the legislature — both at the municipal and state level — was recognized early on as an important arena for the institutionalization of advances in social policies. While in São Paulo many important decisions were made through ordinances and decrees to speed things up, in Belo Horizonte the introduction of bills prevailed, aiming to guarantee, in theory, greater continuity and institutionality for the instruments regulating the policy and the role of actors.

The continuity of actions in Belo Horizonte in this pre-SUAS period can be explained by the political and budgetary priority given to the social services area, the continuity of left-wing mayors, and the maintenance of actors connected to the social services policy community at the head of the department. This community included social workers, psychologists, and sociologists with experiences in different municipal administrations within the broad sector of grassroots and democratic movements and organizations¹⁵ — connected to the movement of reconceptualizing social services and the social services sector of the PT. These actors were recognized by municipal administrations as having the experience and expertise to define the actions that should be taken in this area, and there was also, in some administrations, political alignment between these actors.

As our interviews pointed out, in this context where the policy was being built nationally, there is a process of learning and dissemination between municipalities of different sizes and regions via party networks — the ‘PT way of running government’ and the weight of the social services sector in the PT, as analyzed by Gutierres (2015). The BH experience in the regulation of precepts of the 1988 Constitution and the LOAS was built through the exchange with other PT municipal administrations, in a learning-by-doing process of learning and dissemination of ideas. The PT city governments learned together as they looked at their realities. The national spaces where the PT was being built were used to share and spread experiences. It is also interesting to note the processes of dissemination and learning between the cases of SP and BH: in 1997, André Quintão created, from a sketch made by Aldaíza Sposati for São Paulo, the Partnership Law Nº 7427/97 (regulated on May 16, 2000, by Decree Nº 10.241¹⁶), an important instrument that standardized the policy on agreements between civil society entities and the government. This period was also important for the strengthening and integration of the ‘PT Crew’ of social workers, an important part of the national policy community, who would later work in the Ministry of Social Development and Fight Against Hunger (MDS) — starting with Patrus Ananias himself, mayor of BH and pioneer in the implementation of the LOAS, as well as the first MDS minister.

In this context prior to the establishment of the SUAS, when municipalities had greater autonomy, a Belo Horizonte agenda stands out: the ambitious intersectoral proposal of BH Cidadania (BH Citizenship), which included the coordinated efforts of several departments in the social area (VEIGA and BRONZO, 2014). According to our interviewees, there was even some tension between municipal managers who wanted to leave ‘their mark’ in BH with the implementation of BH Cidadania facilities, and actors connected to the social services field who were already concerned with the consolidation of the area, not only at the municipal level, but also at the national level, through the creation of an ‘identity’ that this policy did not yet have. In addition, according to the interviews we conducted, BH Cidadania had advantages and disadvantages to it. While on the one hand the program facilitated the implementation and expansion of the NAFs and placed social services at the top of the municipal agenda along with other social policies (which was even reflected in terms of budget), on the other hand there was uncertainty about its role and institutional nature, as its facilities were managed by both the Social Services Subdepartment and the SCOMPS. And, the challenges came with the implementation of the SUAS, as discussed in the next section.

In summary, when looking into the dimensions of the political dynamics and organization of policy communities in the pre-SUAS period, we observe great contrasts. In Belo Horizonte, an integrated policy community was formed, playing a strong role in building the area. Social services in BH were based on a higher political priority given to this department, guided by the mobilization of actors from the policy community, the holding of civil service entrance examinations for the area, and prioritization of the direct provision of services. Meanwhile, in São Paulo, there was greater discontinuity and less institutionalization of social services as a public policy, and greater challenges in building an integrated policy community beyond some key actors at PUC-SP. The prevalence of right-wing and center-right governments negatively impacted the development of the area, contributing to connect social services to fragmented actions guided by the logic of poor relief. While there was no career structuring for the area, important advances occurred during the Erundina administration in terms of greater battles being waged over political space and regulation of indirect provision.

The Post-SUAS period: experiences, alignment, and misalignment

After the PT took federal office, in 2003, and the MDS was created, in 2004, there was an important window of opportunity for the consolidation of national parameters for the social services policy in Brazil (JACCOUD et al., 2017). Part of these municipal experiences then began to influence decision-making at the federal level, through different mechanisms. On the one hand, there was the work conducted by the São Paulo former municipal social services secretary, Aldaíza Sposati, a former PT city councilor, Social Service professor at PUC-SP, and leader of one of the epistemic communities within the social services community, as an entrepreneur of ideas. On the other, part of the BH public policy community moved around the federal government, after the creation of the MDS, in the Patrus Ananias administration. Several actors who worked in the social services department were invited to take on important positions in the MDS, assisting in the process of building the general parameters for structuring the SUAS. While, on the one hand, this shows how subnational experiences help to compose the decision-making repertoire of federal regulation — through formal and informal channels, and not only in arenas such as the CIT —, on the other hand, a space for autonomy remains in the subnational game. Varying according to their inducing power and the strategic calculations of local actors, the rules defined after 2004 (such as the PNAS) helped to reorganize the municipal game, changing correlations of forces, opening up opportunities for certain agendas and making it difficult to maintain others, such as the BH Cidadania intersectoral agenda. In this new institutional arrangement, state-level participation or lack thereof makes a lot of difference, and not only the relationship between the federal government and the municipal level.

In São Paulo, full abidance to the SUAS happened in 2005, amid the José Serra administration, with the PSDB (2005-2006), soon followed by Gilberto Kassab (2006-2012), with DEM (currently Brazil Union). This was a period of contradictory movements, which combined joint coordinating efforts with national SUAS guidelines and the search for leaving a local mark through the implementation of the municipality’s own social services programs, such as Ação-Família and Viver em Comunidade. There was room for local credit-claiming, associated with political-programmatic views, even as national parameters were in force. The Kassab administration (2006-2012) chose a municipal secretary connected to the social services policy community, Alda Marco Antônio, with the PMDB. She waged battles over ways of locally translating the SUAS in other directions, criticizing, for example, the centrality of the state in service provision, advocated by Aldaíza Sposati. The fact that she belonged to the policy community proved to be important for building rules and regulations, particularly on how partner CSOs worked, and the resources mobilized by the secretary were key: political capital, knowledge of SUAS rules, influence in the sector. According to the interviews we conducted, the actors in this field classify this administration as more ‘open to dialogue’, while Aldaíza's management is sometimes characterized by confrontation with the CSOs. This management also promoted the expansion of public social services facilities (notably, the CRAS), in line with national guidelines.

The Fernando Haddad administration (2013-2016), with the PT, maintained the trend of appointing secretaries with a technical-political profile from outside the social services field. Luciana Temer, who was not part of the social services community, but rather a lawyer regarded as part of the PMDB technical cadre, was appointed to the social services department because of the logic of forming a municipal coalition government and the political capital coming from her father, the then Brazilian vice president during the Dilma Rousseff administration (PT). There were advantages and disadvantages to distancing the government from the area: on the one hand, the secretary tried to propose innovations and agendas without much concern for the weight of her ‘legacy’ and the prohibitions and opportunities represented by the regulation of the area; on the other hand, the asymmetry of information between the secretary, the SMADS’ mid-level bureaucracy, and partner entities themselves unbalanced some of the battles waged, particularly those related to the regulation of CSO activities and their oversight.

The following administrations were characterized by great instability at the top of the social services department (LARA, 2020). João Doria (PSDB, 2017-2018) resigned after one year into his term to take over the state government, and his vice mayor Bruno Covas (PSDB, 2018-2020) took over as mayor. In less than two years, six different secretaries were entrusted with the management of this department. At the national level, there was a spending freeze with the enactment of the 95^th Constitutional Amendment, called the Public Expenditure Cap Bill; at the municipal level, cuts were also made to the social services budget. A major institutional change began around this time: the adaptation of the department to the guidelines of the Regulatory Framework for Civil Society Organizations (MROSC), which culminated in the enactment of Normative Instruction Nº 03/SMADS/2018 (São Paulo Municipal Government, 2018), the main document regulating and adapting the provisions of the MROSC to specific local social services characteristics. There were conflicts around the MROSC implementation process, permeated by difficulties in the operationalization of the new guidelines (LARA, 2020)¹⁷. On the other hand, the MROSC has established a new level for these socio-state interactions, the consequences of which for service provision still need to be investigated.

Despite the political instability of recent years and party alternations, two important pillars of transformation can be identified since the 1990s:

1. on the state side, there was an important process of building capacities for provision and especially regulations deriving not only from autonomous processes of structuring municipal bureaucratic agencies, but centrally from processes of interaction, negotiation, and battle waging with the CSOs responsible for offering services. These socio-state interactions do not occur horizontally with any or all CSOs, but with a select group of actors that can influence the formal and informal decision-making arenas, and sometimes voice the demands of other CSOs with less power resources, as discussed in other works (BICHIR, BRETTAS, and CANATO, 2017; BICHIR, PEREIRA, and GOMES, 2021).

2. considering the ‘organization and influence of the CSOs’ involved in this policy in São Paulo, there is an important process of professionalization and specialization — by segments of the population or by territories in the city. These transformations derive not only from changes in the field of third-sector organizations and general regulatory frameworks (as discussed in MENDONÇA et al., 2019), but also from specific national regulations for this policy sector. There is therefore little empirical basis to generalizing narratives based on the idea of ‘privatization’ or ‘philanthropization’ of the sector concurrently with the failure to hold the state responsible.

In short, the social services governance pattern stabilized over time in São Paulo is arguably strongly based on indirect service provision, defined in the interaction between state actors and CSOs. These interactions are guided by the SUAS national rules, but decided on mainly through local political choices and battles, in processes that alternate periods of greater state capacity to regulate indirect provision (especially in left-wing governments and when the department secretary is connected to the social services sector), and periods when CSOs enjoy greater autonomy, not only in service provisioning, but also in decision-making processes (in right-wing governments and/or those disconnected from the community). So political and programmatic battles over the meanings of ‘public responsibility’ for the right to social services found unique translations in the different municipal administrations, although we can draw lines of continuity after the SUAS was established.

In Belo Horizonte, in turn, the social services policy generally translates the precept of state responsibility included in the 1988 Constitution as a priority for direct service provisioning. Since the pre-SUAS period, there had been attempts to build public policies with more state characteristics in Belo Horizonte, within local budgetary restrictions. Governability, program expansion, and program improvement were ensured through conversations with the municipality’s social movements and organizations, as did the advance in regulating the partner network. This does not mean the absence of CSOs in Belo Horizonte, but a very particular characterization when compared to São Paulo: many of them have been connected to social movements since the 1990s — such as the movement advocating for childcare centers — and played an important role in structuring the social services policy.

The implementation of the SUAS opened up the possibility of expanding Belo Horizonte’s own network and building local social protection more systemically. It is observed, however, that the political dynamics of the municipality had changed, both because other political forces joined the city administration and as an effect of the migration of an important part of the policy community to the federal administration. In this sense, Fernando Pimentel’s second term in office (PT, 2005-2008) is marked by contradictory movements. On the one hand, the municipality sought to align itself with SUAS guidelines (the NAFs started to comply with CRAS requirements); on the other, the implementation of BH Cidadania becomes stronger as a municipal program to combat poverty, outside social services institutions. There were political conflicts over the priority of BH Cidadania as a brand for social services and the SUAS national guidelines.

Political challenges also characterized the following administration, under Marcio Lacerda (PSB, 2009-2016), which did not have the full support of the political party that ran the city of Belo Horizonte from 1993 to 2008, the PT. Thus, after the election of Lacerda, there was a breakdown of the group that previously ran the department: while many management-level public servants who had been hired through entrance examinations continued to head important programs, other managers took over the conduction of the local social services policy. Lacerda appointed names outside the community to head the Regional Social Services Boards, important spaces that ensure management decentralization and proximity to local territories. Still, there were important advances during this administration, both induced by the federal government and by local decisions, such as, expansion of CRAS, structuring of Special Social Protection and public service entrance examination held in 2012, after a request from the Public Prosecutor's Office, in which a significant number of professionals were hired. In his second term, the city administration sought to improve its compliance with national guidelines, but without the centrality or protagonism that characterized Belo Horizonte. In 2015, the SUAS Law was created, updating the 1996 LOAS-BH based on SUAS guidelines.

Alexandre Kalil (2017-2021, PHS, then PSD) marks the resumption of ‘pioneering management’ at the head of social services. Several actions were carried out aiming at realigning social services to SUAS parameters. They included the reformulation of the Social Services Boards (DRAS) for greater presence in local territories (all DRAS directors under this administration had a long background in social services); appointment of CRAS and CREAS facility coordinators through public personnel selection processes; creation of Regional Basic Protection teams to cover spaces that were not served by the CRAS, and others. An administrative reform was carried out and the entire institutional structure of the social services department was redesigned; spaces for participation, such as the CLAS and regional committees, were promoted, a public service entrance examination was held for the election of DRAS managers, in which all of them came from the social services community, and others.

While, on the one hand, structuring public service entrance examinations and careers for social services employees helped to consolidate bureaucratic capacities and ensure continuity in the provision of services, on the other, it is interesting to note the incorporation of new actors with different views (politically and programmatically) on how this policy should be structured in Belo Horizonte. The interviews we conducted revealed battles waged between ‘different generations’ of social services in recent administrations. On the one hand, young managers who passed the 2012 public service entrance examination — and coordinated efforts through the Municipal Workers’ Forum (FMT) — sometimes take SUAS as a ‘given’ and seek to discuss the municipality’s specific characteristics, reflecting on whether or not to have federal guidelines adjusted for the local reality. On the other hand, a historical generation of SUAS pioneers returned to Belo Horizonte after the impeachment of President Dilma Rousseff in 2016 and began to advocate for the importance of locally following these federal guidelines that they fought so hard to build, with the necessary adaptations.

In summary, in Belo Horizonte there is a greater weight of political activists working for state protagonism in social services policy provision, as well as the effects of long continuity of left-wing administrations and the mobilization of members of the social services policy community among top decision-makers. CSOs are part of the policy in a more clearly defined area of work (particularly in Special Social Protection) and in balance with direct service provision, and most post-SUAS administrations seek to comply with federal guidelines, even if to the detriment of their own brands, like BH Cidadania.

Table 02 shows the summary of these governance patterns considering the four main dimensions of analysis.

Table 02
Main analytical dimensions that explain the differences in local governance in São Paulo and Belo Horizonte

Dimensions	São Paulo	Belo Horizonte
Municipal policy organization	Metropolis with a mostly conservative profile and most secretaries with a technical-political profile coming from outside the social services Community	Politically progressive metropolis with a majority of secretaries connected to the social services community
Organization of policy communities and municipal bureaucracy	Predominance of policy entrepreneurs who form more of an epistemic community for national battle waging than a local policy community, and absence of a career-structured bureaucracy	Constitution of a public policy community working locally and nationally. Creation of career standards, holding relatively regular public service entrance examinations, and a policy community
Profile of civil society actors	Partner CSOs who enter into agreements with the state to provide different types of services are the key social actors; even when they coordinate efforts with broader social movements, these actors play a greater role and are more professionalized. They operate in formal arenas including councils and also through networks of informal relationships	Broader social movements are key to the constitution of spaces for participation and mobilization of social demands toward the government. CSOs play a more restricted role in the provision of specific services
Types of subnational state capacities	Focus on the regulation of indirect provisioning and structuring of services beyond national regulations	Greater focus on direct provisioning and structuring of services in line with national regulation

• Source: Elaborated by the authors, based on interviews and document analysis.

Final considerations

This article addressed how the vertical and horizontal axes of multilevel governance connect to each other and explain, over time, the patterns of social services policy organization in two metropolises, São Paulo and Belo Horizonte.

From the point of view of the relationships between different levels of government—the vertical axis of multilevel governance —, it makes a difference to consider what is the inducing power of federal regulations, what is the space of subnational decision-making autonomy even in the presence of general parameters, and to understand which actors participate in these decision-making processes. From the enactment of the Brazilian Constitution in 1988 to 1993, the first subnational experiences for building social services as a public policy took place in an environment that was still relatively rarefied from the point of view of clear guidelines for the area. After the enactment of the LOAS, in 1993, these pioneering experiences spread, but still had no systemic perspective, yet had great room for CSOs to operate, even at the level of national policy design (one example is the Solidarity Community Program during the FHC administration). Also, during the FHC administration, arenas of federative agreement were created for this policy —CIT and CIBs —, similar to the health policy. The 4^th National Conference on Social Services took place in 2003, marking a great social mobilization for the systemic building of social services, with the participation of several social movements and the Social Services Sector of the PT (GUTIERRES, 2015). Since 2004, with the establishment of the PNAS, and even more strongly so since the enactment of the NOB-SUAS (2005), the systemic building of the SUAS began with the development of strong inducing mechanisms, as already mapped in the literature (JACCOUD, 2020; JACCOUD, BICHIR, and MESQUITA, 2017).

However, many of the decisions that outline the social services policy in Brasilia derive from the participation — both formal and informal — of actors with great local protagonism. Coordinated efforts between local innovative experiences, testing of ideas on policy design, and federal rulemaking can be explained both by the centrality of certain entrepreneurs from the field itself — which is the case of Aldaíza Sposati in SP — and by the migration effects of local policy networks, as is notably the case of an important part of the BH social services policy community —the ‘Patrus crew’ that started to hold key positions at the SNAS¹⁸. On the other hand, the game in Brasilia has its own dynamics and challenges, and even the social services community from Minas Gerais that migrated to Brasilia failed to follow closely Belo Horizonte’s local dynamics. This actually leads to tensions and new accommodation needs after returning to the local level, with generational tensions and tensions between political groups waging battles over the meanings of the institutionalization of social services as a public policy beyond that recommended by federal rules (part of the battle refers precisely to the ‘sufficiency’ of or ‘adjustment’ of the rules that come from Brasilia to the reality of Belo Horizonte).

When we look in more detail at the horizontal axis over time, we observe that legacy effects matter from the point of view of capacity building, of how civil society actors work and what their profile is, and of the organization and influence of policy communities and bureaucracies. In São Paulo, benefaction and philanthropy are central (MESTRINER, 2008), although part of the CSOs is also combined with the movement advocating for the building of social services as a public policy. In this sense, the continued protagonism of CSOs in service provision does not allow us to establish simplistic diagnoses that assume that it is merely about a permanence of the past in the present or an attitude contrary to the institutionalization of the area. There is a complex ecology of civil society actors, with different resources and influence on government decisions, helping to build production capacities for this policy.

It is also worth noting the pervasive influence of PUC-SP in the conformation of ideas and programmatic views on social services, but its low influence on local political dynamics; a policy community such as that observed in Belo Horizonte is not established, nor are career standards or entrance examinations structured to allow specialized actors to be absorbed into the area. The local political dynamics, dominated by politicians who have a conservative profile, are also an obstacle to the continuity of actions. However, it is interesting to note that even in the left-wing municipal governments running the city before the policy was built at the national level (Luiza Erundina and Marta Suplicy), there are limits within the left itself to the work of these social workers: Aldaíza’s individual protagonism as an entrepreneur in the field does not translate into the establishment of a solid policy community mobilized by occasional left-wing administrations (even under Suplicy, Aldaíza only became secretary in the second year into her term, and in the Haddad administration, there is a political choice to ‘use’ the social services department to form a support base in the city legislature, by appointing an outsider to run the area.

In Belo Horizonte, social movements (and not CSOs) play the major role and there is a pro-participation view of civil society: the relevance of the pro-childcare mobilization and the political visibility of the issue of homeless children since the first Patrus Ananias administration were essential as initial pillars for the building of the social services policy. Moreover, the first experiences of PT administrations after the enactment of the Brazilian Constitution and after the LOAS, as well as the conversations between the public administration and PUC-MG (through formal and informal relationships), helped to consolidate a very active public policy community, including the coordination between social policy subsystems (notably social services and food security). There are, therefore, other political and institutional dynamics and a coalition of left-wing parties and managers who valued the social services agenda as a public policy.

After the implementation of the SUAS, other factors must be considered to understand the dynamics of this policy in these two cities. In general terms, we can argue that the governance patterns for this policy are generally guided by the rules defined nationally by the SUAS, but they are centrally defined by local battles and interactions, conditioned by: 01. the legacy of the sector — history of provision capacities in civil society or in the government in conversations with social movements; greater social participation or greater decision-making isolation; 02. political choices for policy organization — particularly the profile of top decision-makers (with members coming from this policy community or outsiders who are more interested in forming the support base of the mayor in the city legislature) —; particularly, the profile of the secretary and their programmatic ideas greatly affect the correlation of forces with CSOs; 03. how different instruments (regulatory, financial, and provision structuring) are decided on and operated over time in the relationship between state actors at various levels (federal/state/municipal) and branches (executive/legislative and judicial), in the interaction and battle with non-state actors (CSOs, social movements, citizens/users). Not all of these aspects could be discussed in detail in this article, but they are important as dimensions to be considered in future agendas.

As a future research agenda, we highlight the importance of consolidating a more integrated analytical model, based on the results of qualitative and quantitative research, as well as new case studies aimed at testing the external validity of the analytical elements outlined here. This way, we can move towards explanatory mechanisms that combine formulation and implementation in multilevel social policy governance processes.

References

• AMÂNCIO, Júlia Moretto (2008), Parcerias entre Estado e sociedade civil: significados e desafios na gestão de políticas públicas. O caso da assistência social em São Paulo. Master's dissertation Instituto de Filosofia e Ciências Humanas. Universidade Estadual de Campinas.
• ARRETCHE, Marta Teresa da Silva (2016), Federalism, social policy, and reductions in territorial inequality in contemporary Brazil. In: New order and progress: development and democracy in Brazil. Edited by SCHNEIDER, Ben Ross. Oxford: Oxford University Press. pp. 162-186.
• ARRETCHE, Marta Teresa da Silva (2012), Democracia, federalismo e centralização no Brasil. Rio de Janeiro: Editora FGV. 232 pp..
• BICHIR, Renata Mirandola; BRETTAS, Gabriela Horesh; CANATO, Pamella (2017), Multi-level governance in federal contexts: the Social Assistance Policy in the city of São Paulo. Brazilian Political Science Review Vol. 11, Nº 02, pp. 01-28.
• BICHIR, Renata Mirandola and GUTIERRES, Kellen Alves (2019), Sistema Único de Assistência Social: ideias, capacidades e institucionalidades. In: As políticas da política: desigualdades e inclusão nos governos do PSDB e do PT. Edited by ARRETCHE, Marta Teresa da Silva; MARQUES, Eduardo Cesar Leão, and FARIA, Carlos Aurélio Pimenta de. São Paulo: Unesp. pp. 218-237.
• BICHIR, Renata Mirandola; PEREIRA, Guilherme Nunes; GOMES, Maria Laura (2021), Interações socioestatais e construção de capacidades nas políticas públicas: o caso da assistência social na cidade de São Paulo. Novos Estudos CEBRAP Vol. 40, Nº 01, pp. 57-79.
• BICHIR, Renata Mirandola; SIMONI JR., Sérgio; PEREIRA, Guilherme Nunes (2020), Sistemas nacionais de políticas públicas e seus efeitos na implementação: o caso do Sistema Único de Assistência Social (Suas). Revista Brasileira de Ciências Sociais Vol. 35, Nº 102, pp. 01-23.
• BRETTAS, Gabriela Horesh (2016), O papel das organizações da sociedade civil na política pública de assistência social no Brasil: dilemas e tensões na provisão de serviços. Master's dissertation Escola de Artes, Ciências e Humanidades. Programa de Pós-Graduação em Gestão de Políticas Públicas. Universidade de São Paulo.
• BRUCH, Sarah K. and WHITE, KaLeigh K. (2018), Politics, State discretion, and retrenchment in safety net provision: evidence from the U.S. in the Post-Welfare Reform Era. Cambridge Journal of Regions, Economy, and Society Vol. 11, Nº 03, pp. 459-483.
• CAVALCANTE, Pedro Luiz Costa and PIRES, Roberto Rocha Coelho (2018), Governança pública: das prescrições formais à construção de uma perspectiva estratégica para a ação governamental. Boletim de Análise Político-Institucional Nº 19, pp. 09-15.
• CONTEH, Charles and HARDING, Brittany (2021), Boundary-spanning in public value co-creation through the lens of multilevel governance. Public Management Review Vol. 25, Nº 01, pp. 104-128.
• CÔRTES, Soraya Vargas (2015), Policy community defensora de direitos e a transformação do Conselho Nacional de Assistência Social. Sociologias Vol. 17, Nº 38, pp. 122-154.
• COSTA, Bruno Lazzarotti Diniz and PALOTTI, Pedro Lucas de Moura (2011). Relações intergovernamentais e descentralização: uma análise da implementação do SUAS em Minas Gerais. Revista de Sociologia e Política Vol. 19, Nº 39, pp. 211-235.
• DESHPANDE, Rajeshwari; KAILASH, KK, and TILLIN, Louise (2017), States as laboratories: the politics of social welfare polices in India. India Review. Vol. 16, Nº 01, pp. 85-105.
• EVERS, Adalbert (2005), Mixed welfare systems and hybrid organizations: changes in the governance and provision of social services. International Journal of Public Administration Vol. 28, Nº 10, pp. 737-748.
• GUTIERRES, Kellen Alves (2015), Projetos políticos, trajetórias e estratégias: a política de assistência social entre o partido e o Estado. Doctoral thesis Instituto de Filosofia e Ciências Humanas. Universidade Estadual de Campinas.
• HAAS, Peter M. (1992), Introduction: epistemic communities and international policy coordination. International Organization Vol. 45, Nº 01, pp. 01-35.
• HOOGHE, Liesbet and MARKS, Gary (2003), Unraveling the central state, but how? Types of multi-level governance. American Political Science Review Vol. 97, Nº 02, pp. 233-243.
• JACCOUD, Luciana B. (ed) (2020), Coordenação e relações intergovernamentais nas políticas sociais brasileiras Brasília: IPEA. 479 pp..
• JACCOUD, Lucina B.; BICHIR, Renata Mirandola, and MESQUITA, Ana Cleusa (2017), O SUAS na proteção social brasileira: transformações recentes e perspectivas. Novos Estudos CEBRAP Vol. 108, pp. 37-53.
• KAZEPOV, Yuri and BARBERIS, Eduardo (2013), Social assistance governance in Europe: towards a multilevel perspective. In: Minimum income protection in flux Edited by MARX, Ive and NELSON, Kenneth. Palgrave Macmillan. pp. 217-248.
• KAZEPOV, Yuri (ed) (2005), Cities of Europe: changing contexts, local arrangements, and the challenge to social cohesion. In: Cities of Europe: changing contexts, local arrangements, and the challenge to urban cohesion Hoboken: Wiley-Blackwell. pp. 03-42.
• KRIISK, Kersti (2019), Distribution of local Social Services and territorial justice: the case of Estonia. Journal of Social Policy Vol. 48. Nº 02, pp. 329-350.
• LARA, Maria Fernanda Aguilar (2020), A implementação do Marco Regulatório das Organizações da Sociedade Civil (MROSC) na assistência social do município de São Paulo: uma análise dos discursos de profissionais do SUAS. Master's dissertation Instituto de Psicologia. Universidade of São Paulo.
• LASCOUMES, Pierre and LE GALES, Patrick (2007), Understanding public policy through its instruments: from the nature of instruments to the Sociology of Public Policy Instrumentation. Governance Vol. 20, Nº 01, pp. 01-21
• LAVALLE, Adrián Gurza; CARLOS, Euzeneia; DOWBOR, Monica, and SZWAKO, José (eds) (2019), Movimentos sociais e institucionalização: políticas sociais, raça e gênero no Brasil pós-transição. Rio de Janeiro: EDUERJ. 411 pp..
• MARIN, Thais Rodrigues (2012), Entre a caridade e os direitos sociais: a poli´tica da poli´tica de assiste^ncia social no munici´pio de Sa~o Paulo (1989-2012). Master's dissertation Faculdade de Filosofia, Letras e Ciências Humanas. Universidade de São Paulo.
• MARQUES, Eduardo Cesar Leão (2021), The politics of incremental progressivism: governments, governances and urban policy changes in São Paulo. New Jersey: Wiley. 304 pp..
• MARQUES, Eduardo Cesar Leão (2013), Government, political actors and governance in urban policies in Brazil and São Paulo: concepts for a future research agenda. Brazilian Political Science Review Vol. 07, Nº 03, pp. 08-35.
• MENDONÇA, Patrícia Maria Emerenciano de; MEDEIROS; Anny Karine de, and ARAÚJO, Edgilson Tavares de (2019), Models for government-nonprofits partnerships: a comparative analysis of policies for AIDS, social assistance and culture in Brazil. Journal of Public Administration Vol. 53, Nº 05, pp. 802-820.
• MESTRINER, Maria Luiza (2008), O Estado entre a filantropia e a assistência social. São Paulo: Cortez. 320 pp..
• MEZA, Oliver (2021), Punctuated equilibrium in multilevel contexts: how federal and state level forces feedback into shaping the local policy agenda in Mexico. Policy Studies Journal Vol. 50, Nº 03, pp. 575-594.
• MINAS, Renate; VIBEKE Jakobsen; KAUPPINEN, Timo M.; KORPI, Tomas; and LORENTZEN, Thomas (2018), The governance of poverty: welfare reform, activation policies, and social welfare benefits and caseloads in Nordic countries. Journal of European Social Policy Vol. 28, Nº 05, pp. 487-500.
• OSBORNE, Stephen P.; RADNOR, Zoe, and STROKOSCH, Kirsty (2016), Co-production and the co-creation of value in public services: a suitable case for treatment?. Public Management Review Vol. 18, Nº 05, pp. 639-653.
• PIRES, Roberto Rocha Coelho and GOMIDE, Alexandre de Ávila (2018), Governança e capacidades estatais a partir da abordagem dos arranjos e instrumentos de políticas públicas. Boletim de Análise Político-Institucional Nº 19, pp. 25-32.
• SAETREN, Harald (2014), Implementing the third-generation research paradigm in policy implementation research: an empirical assessment. Public Policy and Administration Vol. 29, Nº 02, pp. 84-105.
• SAMUELS, David J. and ZUCCO, Cesar (2018), Partisans, antipartisans, and nonpartisans: voting behavior in Brazil. Cambridge: Cambridge University Press. 196 pp..
• SÁTYRO, Natália Guimarães Duarte and CUNHA, Eleonora Martins Schettini (2019), Jogando luzes sobre os determinantes burocráticos e de gestão da política de assistência social. Opinião Pública Vol. 25, Nº 02, pp. 401-432.
• SELLERS, Jefferey M. and LINDSTRÖM, Anders (2007), Decentralization, local government, and the welfare state. Governance Vol. 20, Nº 04, pp. 609-632.
• SELLERS, Jefferey M.; LIDSTRÖM, Anders, and BAE, Yooil (eds) (2020), Introduction: taking local institutions seriously. In: Multilevel democracy Cambridge: Cambridge University Press. pp. 01-20.
• SILVA, Maria Ozanira da Silva e; ARAÚJO, Cleonice Correia and LIMA, Valéria Ferreira Santos de Almada (2012), Implantação, implementação e condições de funcionamento do Suas nos municípios. In: O Sistema Único de Assistência Social no Brasil: uma realidade em movimento. Edited by COUTO, Berenice Rojas; YAZBEK, Maria Carmelita; SILVA, Maria Ozanira Silva and DEGENSZAJN, Raquel Raichelis.São Paulo: Cortez Editora. pp. 88-113.
• TORTOLA, Pier Domenico (2017), Clarifying multilevel governance. European Journal of Political Research Vol. 56, Nº 02, pp. 234-250.
• YANG, Jae-Jin; KIM, Hey Sung; CHOI, Seong Eun; RYU, Lanhee, and CHOI, Young Jun (2019), What makes hybrid insourcing successful: a new public - private partnership model for social welfare services. Asian Social Work and Policy Review Vol. 14, Nº 03, pp.0 1-11.
• YAZBEK, Maria Carmelita (2004), Assistência social na cidade de São Paulo: a (difícil) construção do direito. Observatório dos direitos do cidadão: acompanhamento e análise das políticas públicas da cidade de São Paulo. São Paulo: Instituto Pólis/PUCSP. 71 pp..
• VEIGA, Laura da and BRONZO, Carla (2014), Estratégias intersetoriais de gestão municipal de serviços de proteção social: a experiência de Belo Horizonte. Revista de Administração Pública Vol. 48, Nº 03, pp. 595-620.
• WINTER, SøRen C. (2006), Implementation. In: Handbook of public policy. Edited by PETERS, B. Guy and PIERRE, Jon. London: Sage. pp. 151-166.

• Data replication: https://doi.org/10.7910/DVN/SC8ZLX
• *
  This article is one of the products of the research project ‘Multilevel governance in nationally structured social policies: the case of social services’, coordinated by Renata Bichir at the Center for Metropolitan Studies (CEM), a Center for Research, Innovation, and Dissemination (CEPID) funded by the São Paulo Research Foundation (FAPESP). We thank FAPESP for funding our research. The authors thank the other researchers in the group for their comments, for the suggestions received at the 13 ABCP Meeting, and the anonymous reviewers with the BPSR for their suggestions.
• **
  This article was presented at the 13th Meeting of the Brazilian Political Science Association (ABCP), which took place from September 19 to September 23, 2022, in an online format, and was named as the best paper presented in the subfield of “State and Public Policies”.
• 1
  Aldaíza Sposati, municipal secretary during the Marta Suplicy administration in São Paulo, was actually a consultant for the National Social Services Department (SNAS) of the former Ministry of Social Development and Fight Against Hunger (MDS) to develop the concept of ‘socio-family matriciality’ and define technical guidelines for the work conducted in the CRAS centered around this concept.
• 2
  In this article, we follow the classification of Brazilian parties adopted by Samuels and Zucco (2018).
• 3
  Understood here in the terms of Haas (1992, p. 03): “a network of professionals with recognized expertise and competence in a particular domain and an authoritative claim to policy-relevant knowledge within that domain or issue area”.
• 4
  The PNAS established parameters for the implementation of the SUAS, defined the types of security under the responsibility of social services (reception, income, coexistence, autonomy, circumstantial risks), and also specified the organization of public facilities in all municipalities as a gateway to social services, separated according to the level of complexity of the service, establishing basic protection and special protection.
• 5
  The NOB-SUAS institutionalized the principles of the SUAS, reinforcing the universality of politics and the responsibility of the state. Moreover, it established hierarchies and standards for social services and assistance and defined more clearly the roles of each sphere of government.
• 6
  Classifications establish the standards for basic and special protection services, provided directly or indirectly, so that they can be entitled to federal co-funding. This means that while there may be occasionally specific social services programs in states and municipalities, if they are not framed within national parameters, they cannot receive federal funds via FNAS (JACCOUD et al., 2017).
• 7
  The political arrangement of the SUAS also includes the National Council of Municipal Social Services Managers (CONGEMAS) and the National Forum of State Social Services Secretaries (FONSEAS) as spaces for coordination and representation of municipalities and states and the federal district, respectively.
• 8
  The SUAS Census is the main process for monitoring and assessing social services collecting data through an electronic form, which is filled by the State and Municipal Social Services Departments and Councils.
• 9
  Decentralized Management Index. It is the main indicator that aims to assess the quality of the decentralized management of services, programs, projects, and social assistance benefits within municipalities, states, and the Federal District. It is through this index that the federal government assesses whether there will be a transfer of funds to subnational entities.
• 10
  Pop Centers are facilities that offer services to homeless populations. Regional Basic Protection Services are facilities that are not classified at the national level and cover areas that are not served by CRAS facilities.
• 11
  Created by Law Nº 12,324, regulated by Decree Nº 38,877, of December, 1999.
• 12
  The Social Services Sector of the Workers' Party was an important space for building the SUAS and the PNAS, as well as for the dissemination of ideas at municipal levels (GUTIERREZ, 2015).
• 13
  The elected mayor Pimenta da Veiga ran city hall from 1989 to 1990, when he stepped down. His vice mayor, Eduardo Azeredo, then took over and remained in office until 1992.
• 14
  André Quintão was Municipal Secretary of Social Development and city councilor of Belo Horizonte, as well as State Secretary for Labor and Social Development (2015/2016).
• 15
  The expression ‘PT way of running government’ was coined from the experiences of local policies coming from the first city governments run by the Workers' Party.
• 16
  Available at ˂https://leismunicipais.com.br/a/mg/b/belo-horizonte/decreto/2000/1025/10241/ decreto-n-10241-2000-disciplina-os-procedimentos-para-estabelecimento-avaliacao-e-prestacao-de-contas-de-convenios-entre-o-municipio-e-entidades-sem-fins-lucrativos-para-a-promocao-de-acoes-no-ambito-da-politica-de-assistencia-socia.
• 17
  Some of these issues included: constant changes made to the rules regulating the MROSC, short time for implementation, lack of training to perform new jobs required by the MROSC for the position of partnership manager, and others.
• 18
  Working at the DGSUAS, at the Social Services and Assistance Monitoring, playing a key role in the very definition of the SUAS Census, CapacitaSUAS, and other rules (effects not only for top decision-makers, but also for mid-level bureaucracy).

• Funding information: São Paulo Research Foundation (FAPESP); Process Nº 2013/07616-7.

Data availability

Data replication: https://doi.org/10.7910/DVN/SC8ZLX

Publication Dates

• Publication in this collection
  05 July 2024
• Date of issue
  June 2024

History

• Received
  07 Nov 2022
• Accepted
  03 Sept 2023