When Time Matters: Federal Changes, Temporality and Recentralization in Brazil ^*

Quando o Tempo Importa: Mudanças Federais, Temporalidade e Recentralização no Brasil

Quand le Temps Compte: Changements Fédéraux, Temporalité et Recentralisation au Brésil

Cuando el Tiempo Importa: Cambios Federales, Temporalidad y Recentralización en Brasil

ABSTRACT

Much of the scholarly work on Brazilian federalism has long focused on the ways in which decentralization has produced institutional paralysis under intergovernmental conflict. This article, by contrast, suggests that Brazil has been under a gradual transformation since its democratization, which led to a recentralization of power in the hands of the federal government. It presents a framework that explains how an initial decentralization process (1983-1994) turned into a centralization process (1995-2007) that ultimately increased the regulatory and coordination powers of the Brazilian federal government. In addition, the analysis shows that two main blocs of political parties competed to influence the sequence and speed of legislation related to the decentralization and centralization processes. Specifically, these parties used two competing temporal strategies – foot-dragging and pace-setting – to slow down or accelerate the approval of legislation regulating subnational autonomy. Once the centralizing bloc of parties enacted key centralizing legislation, they used the strategy of fence-building to institutionalize intergovernmental relations, thereby consolidating the coordination powers of the center. The main contribution of this article is to advance a temporal approach to federal changes, which explains how the Brazilian federal government centralized authority under a decentralizing Constitution.

decentralization; centralization; temporality; subnational autonomy; federalism; intergovernmental relations

RESUMO

Grande parte dos trabalhos acadêmicos sobre o federalismo brasileiro há muito se concentra nas maneiras através das quais a descentralização produziu certa paralisia institucional diante de conflitos intergovernamentais. Este artigo, ao contrário, sugere que o Brasil está passando por uma transformação gradual desde sua democratização, o que levou a uma recentralização do poder nas mãos do governo federal. Apresenta-se uma estrutura que explica como um processo inicial de descentralização (1983-1994) se transformou em um processo de centralização (1995-2007) que, em última instância, aumentou os poderes regulatórios e de coordenação do governo federal brasileiro. Além disso, a análise mostra que dois blocos principais de partidos políticos competiram para influenciar a sequência e a velocidade de implementação da legislação relacionada aos processos de descentralização e centralização. Especificamente, esses partidos usaram duas estratégias temporais concorrentes – segurar o fluxo ( foot-dragging ) e definir o ritmo ( pace-setting ) – para desacelerar ou acelerar a aprovação da legislação que regulamentaria a autonomia subnacional. Depois que o bloco centralizador de partidos promulgou legislação centralizadara, utilizou a estratégia de construção de cercas ( fence-building ), a fim de institucionalizar as relações intergovernamentais, consolidando assim os poderes de coordenação do centro. A principal contribuição deste artigo é avançar uma abordagem temporal das mudanças federais, explicando como o governo federal brasileiro centralizou a autoridade sob a égide de uma Constituição descentralizadora.

descentralização; centralização; temporalidade; autonomia subnacional; federalismo; relações intergovernamentais

RÉSUMÉ

Une grande partie des travaux universitaires sur le fédéralisme brésilien s’est longtemps concentré sur les façons dont la décentralisation a produit une paralysie institutionnelle dans le cadre de conflits intergouvernementaux. Cet article, en revanche, suggère que le Brésil a connu une transformation progressive depuis sa démocratisation, qui a conduit à une recentralisation du pouvoir entre les mains du gouvernement fédéral. Il présente un cadre qui explique comment un processus initial de décentralisation (1983-1994) s’est transformé en un processus de centralisation (1995-2007) qui a finalement accru les pouvoirs de réglementation et de coordination du gouvernement fédéral brésilien. En outre, l’analyse montre que deux blocs principaux de partis politiques se sont affrontés pour influencer la séquence et la vitesse de la législation liée aux processus de décentralisation et de centralisation. Plus précisément, ces partis ont utilisé deux stratégies temporelles concurrentes ( foot-dragging et pace-setting ) pour ralentir ou accélérer l’approbation de la législation réglementant l’autonomie infranationale. Une fois que le bloc centralisateur de partis a promulgué une legislation-clé, ils ont utilisé le fence-building pour institutionnaliser les relations intergouvernementales, consolidant ainsi les pouvoirs de coordination du Centre. La principale contribution de cet article est de proposer une approche temporelle des changements fédéraux, ce qui explique comment le gouvernement fédéral brésilien a centralisé l’autorité sous une Constitution décentralisatrice.

décentralisation; centralisation; temporalité; autonomie infranationale; fédéralisme; relations intergouvernementales

RESUMEN

Gran parte del trabajo académico sobre el federalismo brasileño se ha centrado durante mucho tiempo en las formas en las que la descentralización ha producido una parálisis institucional bajo el conflicto intergubernamental. Este artículo, por el contrario, sugiere que Brasil ha experimentado una transformación gradual desde su democratización, lo que llevó a una recentralización del poder en manos del gobierno federal. Se presenta un marco de análisis que explica cómo un proceso de descentralización inicial (1983-1994) se convirtió en un proceso de centralización (1995-2007) que finalmente aumentó los poderes de regulación y coordinación del gobierno federal brasileño. Además, el análisis muestra que dos grupos principales de partidos políticos compitieron para influir en la secuencia y velocidad de la legislación relacionada con los procesos de descentralización y centralización. Específicamente, estos partidos utilizaron dos estrategias temporales – generar demoras ( foot-dragging ) y marcar el ritmo ( pace-setting ) – que competían para ralentizar o acelerar la aprobación de la legislación que regula la autonomía subnacional. Una vez que el bloque centralizador de partidos promulgó legislación centralizadora clave, utilizó la estrategia de construcción de vallas ( fence-building ) para institucionalizar las relaciones intergubernamentales, consolidando así los poderes de coordinación del centro. La principal contribución de este artículo es avanzar en un enfoque temporal de los cambios federales, lo que explica cómo el gobierno federal brasileño centralizó la autoridad bajo una Constitución descentralizadora.

descentralización; centralización; temporalidad; autonomía subnacional; federalismo; relaciones intergubernamentales

INTRODUCTION

This article theorizes about federal dynamics in Brazil, during both the decentralization and the centralization processes from 1983 to 2007. Focusing on the Brazilian changing federal dynamics during these recent decades, it seeks to answer three questions: first, how have decentralization and centralization unfolded since the promulgation of the 1988 Constitution? Second, how did decentralization and centralization ultimately lead to recentralization , epitomized by the increase in the federal government’s regulatory and coordination powers? Third, how did political parties on both sides of the decentralization-centralization dichotomy leverage the temporal modalities (i.e., sequence and speed ) of legislative approval to increase their own powers?

Taking into consideration the intricate relationship between decentralization and centralization, this article analyzes both processes at the same time under the term recentralization , defined here as the outcome of a long process that encompasses both initial decentralization (1983-1994) and subsequent centralization (1995-2007)¹. More specifically, recentralization was the outcome of a counter-reaction against decentralization, which gave rise to a sequence of centralization laws. Recentralization does not mean the simple undoing of decentralization, but instead it implies that the center strengthened some of its powers – namely, regulatory and coordination powers.

Recentralization, as the outcome of decentralization- cum -centralization, reflects changes in the autonomy of subnational governments. The concept of autonomy here is exclusive powers of either state or municipal governments exercised without the interference of the federal government. As such, subnational autonomy belongs to the sphere of self-rule, under which subnational governments independently exercise authority in their own territory ( Hooghe et al., 2016: 23). According to this conceptualization, one can interpret an attempt by the federal government to increase its regulatory and coordination powers as limiting the scope of actions and policy decisions of subnational governments. Specifically, the federal government may use its regulatory power to establish new centrally-designed parameters or conditions with which subnational governments must abide. Then, the federal government requires coordination powers to both implement and monitor those new regulations. In short, decentralization reinforces subnational autonomy, while centralization increases the federal government’s regulatory and coordinating powers at the expense of subnational autonomy².

This research is in dialogue with various literatures of territorial politics in Brazil. However, it follows a different line of research in its interpretation that federal relations in Brazil have gradually transformed under the influence of political strategies in which temporality is an important strategic component. Henceforth, the argument for the increase in federal coordination and regulatory powers in Brazil is structured around the temporal approach , which this article develops. The core idea behind this approach is that the passage of time is the medium through which an outcome of a process occurs ( Grzymala-Busse, 2011 ).

It is critical to define temporality in order to understand the passage of time as an instrument of politics. Temporality is a dimension of time in which political processes unfold, and it encompasses different modalities of time: namely, the sequence and speed of de/centralization laws³. Politicians try to influence these temporal modalities to advance their own interests. Within this understanding of temporality, sequence and speed are resources that players can use to influence both the processes and their outcomes ( Howlett and Goetz, 2014 ).

A temporal approach to de/centralization interprets that temporality has the potential to influence both political interactions and the distribution of power. Those who used a temporal approach to federal politics in Brazil between 1983 and 2007 treated time as a medium through which they advanced territorial interests for the ultimate purpose of interfering in federal relations.

According to this approach, temporality in the form of sequence and speed is treated here as an independent variable because it had the potential to activate mechanisms that affected the recentralization trajectory. The notion that this trajectory depended on the sequence and speed of de/centralization implies that these temporal modalities influenced the distribution of power between national and subnational politicians. Sequence and speed impacted recentralization first through the unequal promotion of federal, state, and municipal interests and then by triggering causal mechanisms that favored subnational governments initially and the federal government later on.

I use the term temporal strategy here to refer to the crafted action plans of politicians and political parties who used sequence and speed to shape the de/centralization processes. In this article, these strategies were the basis of de/centralization because they created the key events – the approval of de/centralizing laws – whose effects on recentralization depended on their sequence and speed of approval. In the context of territorial conflicts over competing views of de/centralization, the instrumentalization of temporality allowed politicians to influence territorial politics strategically.

This article evaluates changes in the federal government’s role to both regulate and coordinate through the notion of territorial politics , defined here as the dynamics created by territorial players (for example, parties, government branches, bureaucrats) who used temporality and institutions to influence decisions that affected federal relations. These dynamics were based on territorially-bounded interests. In this article, the degree to which politicians consolidated their power base in a decentralized or centralized distribution of power determined these territorial interests to a great extent. The federal government dominated centralized territorial dynamics and acted as the main agenda-setter in territorial politics. On the other hand, subnational governments (state and municipal) dominated decentralized territorial dynamics, exercising so much autonomy that they sometimes interfered with the agenda of the federal government. Thus, in their respective attempts to influence territorial politics, different government levels pursued competing strategies to control territorial dynamics.

It is worth highlighting here that both state and municipal interests can be grouped under the label “subnational interests” for analytical purposes because, in Brazil, most of the laws pertaining to the de/centralization processes address both levels of government in joint pieces of legislation⁴. In addition, the aggregation of state and municipalities under a single subnational unit is analytically consistent with the territorial party politics adopted in this article, according to which the territorial agenda of parties are organized along the decentralization-centralization dichotomy.

Political parties are the main force that represent territorial interests. This article takes for granted that, within the context studied here, the political parties defined their territorial interests and competed in territorial dynamics based on the decentralization-centralization dichotomy⁵. Within the Brazilian fragmented party system, this analysis considers only four parties – those that have held a majority of the seats in the National Congress since democratization, namely the Brazilian Social Democratic Party (Partido da Social Democracia Brasileira – PSDB), the Workers’ Party (Partido do Trabalhadores – PT), the Party of Democratic Mobilization (Partido do Movimento Democrático Brasileiro – PMDB), and the Liberal Front Party (Partido da Frente Liberal – PFL). These parties are divided into two blocs: those that pursued centralization (PT and PSDB), and those that pursued decentralization (PMDB and PFL)⁶.

These parties fought to control the decentralization agenda because the decentralization-centralization dichotomy is very significant in Brazilian territorial politics. This article argues that both decentralization and centralization can be sought in the legislative sphere through the approval of laws and in the intergovernmental sphere through the creation of technical forums for policy elaboration and implementation. In sum, this article tries to explain how recentralization in Brazil took place through these parties’ temporal and institutional strategies.

This article seeks two main objectives. The first one is to identify and distinguish between temporal-based and institutional-based strategies that political parties in Brazil use. The second one is to identify causal mechanisms that explain how the interplay between temporal and institutional strategies led to recentralization.

As far as the first objective is concerned, this article considers three strategic responses of politicians to de/centralization: foot-dragging, pace-setting , and fence-building . Borzel (2002) , who recognized that the member states of the European Union (EU) adopted different strategic responses to regional integration policies, identified these strategies. The contexts in which norms and rules were approved and implemented differ between the EU and Brazil, so this research uses Borzel’s work only as a normative metaphor that provides a general framework for the strategies of political players from different government levels competing to influence decisions over a wide range of policies.

The three strategies applied to the Brazilian case mainly differ from Borzel’s conceptualization for the following reasons. First, this article, unlike Borzel’s, pays greater attention to the temporal dimension of political processes. Second, while the current analysis interprets political parties as the main agents behind the pursuit of some of these strategies, Borzel’s analysis understands that EU member states both design and implement the strategies. Finally, the types of institutions and territorial players in both analyses are distinct. Whereas here the institutions and players belong to a federal country operating under a constitutional framework, in Borzel’s article the main players and institutions are under the sphere of influence of multiple states and a supranational entity, the EU, all of which function without a clear legal framework.

In this article, the temporal strategies of foot-dragging and pace-setting involve tactics that try to manipulate both the sequence and the speed of law approval. In Brazil, politicians that represented subnational interests used foot-dragging to impose a decentralization sequence that started with the fiscal layer and was followed by the slow approval of laws that tried to change the existing state of decentralization. The rival of foot-dragging is pace-setting , which politicians that represented the interests of the federal government sought in two ways: by imposing fiscal centralization before administrative centralization and by speeding up the approval of centralization laws. That said, these two strategies were opposite to one another because politicians that implemented these strategies were fighting for different sequences and different speeds of law approval.

The institutional strategy of fence-building tries to create rules and regulations for intergovernmental dynamics⁷. In Brazil, this strategy was part of the tactics of both national and subnational politicians to create vertical intergovernmental institutions that would determine outside the legislative sphere the dynamics, either top-down or bottom-up , among different government levels.

With these strategies, one can understand territorial politics in Brazil as a conflict between different territorial interests under low institutionalization of intergovernmental relations, which favored the use of temporal-based strategies in the approval of de/centralization laws. Over time, however, temporal strategies became less effective at achieving de/centralization, since the intergovernmental sphere became more robust, with new institutions created to mediate vertical relations between the central and subnational governments.

As far as the second objective is concerned, causal mechanisms are critical to understand how de/centralization unfolded because they reveal how the different tactics under the above strategies triggered certain behaviors and institutional arrangements that caused specific outcomes. In other words, causal mechanisms are the behavioral and institutional connectors between components of the strategies and the trajectory of recentralization. The article identifies four causal mechanisms: fiscal profligacy behavior, fiscal prudence behavior, status quo bias, and institutional layering . Each mechanism is associated with a specific strategy, context, and effect on the trajectory of recentralization.

The de/centralization processes are analyzed through the 32 legislative proposals that shifted the balance of intergovernmental power in democratic Brazil between 1983 and 2007⁸. I argue that the 32 de/centralizing measures studied in this article are critical and sufficient to explain the different temporal strategies developed in the recentralization trajectory. These legislative measures were selected based on the following criteria⁹: Legislation approved before the 1988 Constitution qualified for this article only if it was the last decentralizing measure of the non-democratic regime in the fiscal and administrative layers of decentralization. Legislation approved after 1988 qualified if it was a constitutional amendment concerning a change in state and municipal authorities, or if it was a law that has since been deemed highly influential in the fiscal, health, and education fields.

Equally central to the argument that the Brazilian federal government increased its powers over intergovernmental relations is the institutionalization of coordination forums and the creation of a top-down intergovernmental interaction. This article argues that as the federal government gradually acquired regulatory powers over subnational governments, it increasingly coordinated its policies and actions through forums that displayed top-down coordination. The federal government institutionalized five primary forums – the National Educational Council (Conselho Nacional de Educação – CNE), the National Council for Fiscal Policy (Conselho Nacional de Política Fazendária – Confaz), the Committee of Federative Affairs (Comitê de Articulação Federativa – CAF), the Tripartite Inter-Managers Commission (Comissão Intergestores Tripartite – CIT), and the Bipartite Inter-Manager Commission (Comissão Intergestores Bipartite – CIB) – that have been in charge of coordinating several aspects of policymaking in the fiscal, education, and health fields.

This article is divided into seven sections, including this one. The next section explains the theoretical and methodological underpinnings of the temporal approach to recentralization in Brazil. It also dissects the causal mechanisms through which the winning strategies fostered decentralization initially and then centralization in later years. The section “Recentralization Strategies and Territorial Interests in Brazil”, identifies the strategies that both national and subnational players used. The section titled “Territorial Party Politics in Brazil”, empirically demonstrates that the parties consistently used different strategies to influence de/centralization over time. The section “ Foot-Dragging and Pace-Setting: Determining the Sequence and Speed o f Decentralization in Brazil” highlights the main temporal patterns in the de/centralization processes and explains why some strategies prevailed in different periods of recentralization. Subsequently, in the section titled “ Fence-Building: The Institutionalization of Intergovernmental Relations in Brazil”, the article shows how the strategy of building top-down intergovernmental forums gradually replaced temporal strategies and consolidated the coordination power of the federal government. The last section highlights the article’s main contributions.

THE TEMPORAL APPROACH TO RECENTRALIZATION IN BRAZIL

This section elaborates on the methodological foundations that underpin the temporal approach presented in this article. This section first explains how temporality is grounded in critical assumptions of historical institutionalism and then explains how the central elements of the temporal-centered approach – strategies, events, and mechanisms – can be traced throughout the recentralization trajectory in Brazil.

Comparative Historical Analysis

Temporality is important in the historical institutional tradition because it can make sense of both institutional changes and their outcomes¹⁰. From this perspective, historical institutionalism is a conceptual and methodological toolbox that emphasizes the role of temporality in influencing institutional changes and continuities ( Fioretos, Falleti and Sheingate, 2016: 4). However, considering the leading role of time in this analysis, this article is better situated under the umbrella of Comparative Historical Analysis (CHA), which has been explicitly used in work that explores territorial dynamics ( Falleti, 2005 ; Broscheck, Petersohn and Toubeau, 2018 ).

Methodologically, CHA tries to advance critical assumptions about outcomes that result from historical processes ( Thelen, 2003 ; Mahoney and Rueschemeyer, 2003 ). These outcomes are traceable to a chain of interconnected events that emerge out of negotiations between players (Falleti and Mahoney, 2016:212-215). Events are the essential components of processes because they reveal both the preferences and strategies of players and the causal mechanisms that guide the trajectory toward the observed outcome.

CHA argues that players try to advance their preferences with the ultimate goal that their group’s preferences will prevail and get institutionalized over other competing preferences ( Pierson, 2016 ). The abilities of different groups to make their strategies prevail over those of other groups shape the trajectory of this process.

In establishing how events lead to a specific outcome, it is important to differentiate between events that cause an outcome from those that are consequential to an outcome (Falleti and Mahoney, 2016:216-219)¹¹. That said, when a process analysis focuses on the sequence and speed of events, as in this article, these events are not necessarily causally related to the outcome but, instead, are consequential to the outcome.

To sum up, in the temporal approach to territorial politics, political strategies shape events that form a sequence and unfold at different speeds . These temporal modalities activate specific mechanisms with specific effects on territorial power relations.

This methodology enables to highlight three characteristics of the recentralization trajectory in Brazil: (i) it was consequential , since the events leading to the recentralization outcome indirectly and incrementally influenced this outcome through a chain of related events; (ii) it was multidirectional, since it changed directions – from decentralization to centralization – because early events prompted counter-reactions; and (iii) the competing strategies behind de/centralization were adaptive strategies, provided they were both weakly and strongly implemented throughout the process, depending on the political and intuitional context.

Strategies

In this article, the temporal approach focuses on past events – specifically, the approval of de/centralization laws. The analysis of past events identifies the strategies of players and associates them with the chain of laws. Any specific strategy had the potential to delay or speed up a law or give shape to a clear sequence of approved legislation.

Political parties are the main agencies in this analysis. Between 1983 and 2007, parties defined their territorial interests along the decentralization-centralization dichotomy and embraced both temporal and institutional strategies to influence the recentralization trajectory according to their territorial interests. From the large constellation of Brazilian political parties, this article will focus on PT, PSDB, PFL, and PMDB as representatives of the main territorial interests. PMDB and PFL, as the decentralizing bloc, used foot-dragging to maintain the status quo of the 1988 Constitution by delaying national legislation designed to regulate decentralized fiscal and administrative competencies. Meanwhile, PT and PSDB, as the centralizing bloc, used both pace-setting and fence-building to speed up legislation designed to regulate decentralized competencies and increase policy coordination.

From 1988 to 2007, PMDB, PFL, PSDB, and PT together were responsible (both directly and indirectly, through the National Executive Branch) for proposing 21 of the 26 post-1988 laws analyzed here. Thus, these four parties have been the leading political force behind territorial politics.

Two important issues guide this analysis of these parties’ strategies: how political parties built territorial interests and how they articulated them. In the party dynamics presented here, the parties did not necessarily converge toward a national-territorial agenda, nor did they coordinate an intra-party strategy using territorial interactions. Rather, the Brazilian party system and the party dynamics gave rise to competing strategies that represented dichotomous territorial interests¹²;

The temporal approach views each approved de/centralizing law as part of a temporal strategy . As such, the pace of law approval gives shape to the chain of events that ultimately renders one strategy as the winner. In Brazil, the first prevailing strategy was foot-dragging , which tried to consolidate the gains of decentralization until the mid-1990s. This strategy was subsequently replaced by a competing strategy, pace-setting , which became a winning strategy in the governments of both PT and PSDB. Fence-building , as an institutionalization strategy, offset the temporal strategies and the winners used it to consolidate the coordination power they acquired.

A critical building block of this article’s analysis is the two temporal modalities – sequences and speed – employed by the strategies of the federal and subnational governments within the legislative sphere. An analysis of sequence reveals that in Brazil, fiscal decentralization came before the administrative one. In general, the sequence of decentralization affects the distribution of power between national and subnational politicians ( Falleti, 2005 ). For instance, when subnational governments acquire fiscal autonomy before the administrative one, their new ability empowers them to spend with few constraints. Based on this rationale, the order of centralization also matters, that is, fiscal centralization before the administrative one allows the federal government to impose stricter fiscal discipline on subnational governments.

Speed refers to the accelerated or delayed approval of laws that affect the rate of political change ( Grzymal-Busse, 2011: 1282)¹³. Obstructing a legislative process in Brazil increases the possibility that a proposed law will be vetoed or altered. In other words, the longer it takes to approve a law, it is most likely that it will be altered or even dropped ( Ricci, 2003: 724; Hiroi, 2008 ). Also, the speed of approval of de/centralizing laws gives insight into the process of institutional change. Rapid institutional change reduces the opportunities for players who oppose the direction of change, provided that this type of change increases the revision costs ( Grzymal-Busse, 2011: 1283-1284). Thus, a speedy approval of decentralization often has little opposition, while a slow approval might allow for opposition to emerge and gather strength.

Unlike the temporal strategies advanced in the legislative sphere, the institutional strategy takes place in the intergovernmental vertical sphere. When the latter is institutionalized, politicians transfer more fiscal and administrative competencies to bureaucrats because the interactions in intergovernmental forums rely heavily on technical cooperation between different government levels ( Behnke and Mueller, 2017 ). Territorial players create institutions in the vertical intergovernmental sphere to impose either a bottom-up or a top-down interaction in this sphere.

Each strategy corresponds to a dimension (either temporal or institutional ), a cluster of players, and a specific set of territorial preferences within the context of Brazilian recentralization. Foot-dragging was embraced by politicians who represented subnational interests to influence the sequence of decentralization in the national legislative sphere – namely, to prioritize the approval of fiscal decentralizing laws before administrative ones. In addition, in order to maintain the decentralization gains, these politicians delayed the approval of any law that tried to regulate the decentralized competencies of subnational governments. Foot-dragging may be sub-classified as either strong or weak. Strong foot-dragging entails two successful pursuits: the approval of fiscal decentralization before the administrative one and a delay in the approval of centralization laws. Weak foot-dragging fails to accomplish one of these pursuits.

The pace-setting strategy emerged to thwart the effects of the foot-dragging strategy. As with foot-dragging, pace-setting may be subclassified as either strong or weak. Strong pace-setting entails two successful pursuits: to impose fiscal discipline by approving fiscal centralization before the administrative one and to enact laws that strengthen the coordination and regulation role of the federal government. Weak pace-setting fails to accomplish one of them.

Finally, the institutional-based strategy of fence-building is used to create institutions that regulate the dynamics of intergovernmental relations in areas such as health, education, and finances. Considering that intergovernmental dynamics are characterized by a power struggle between central and subnational governments, these dynamics may be dominated by either a top-down or bottom-up interaction. Politicians who follow a top-down fence-building strategy are trying to create intergovernmental forums that impose norms on subnational governments. By contrast, politicians who seek a bottom-up fence-building strategy are trying to crystalize the norms of intergovernmental forums in which subnational governments have the upper-hand. In Brazil, the observation that fence-building was deployed after the temporal strategies reveals that it offset the temporal strategies by consolidating the coordinating role of the federal government in the intergovernmental arena.

Table 1 identifies the components of these strategies and summarizes how a specific cluster of territorial players in Brazil used each component of the above temporal-based and institutional-based strategies.

Table 1
: Elements of Recentralization in Brazil

	Modalities	Types of modalities	Who benefits?	Strategies
TEMPORAL DIMENSION	Sequence	Fiscal layer ex ante administrative layer	Municipal and state governments	Foot-dragging
		Administrative layer ex ante fiscal layer	Federal government	Pace-setting
	Speed	Slow approval of laws	Municipal and state governments	Foot-dragging
		Fast approval of laws	Federal government	Pace-setting
INSTITUTIONAL DIMENSION	Institutionalization of ING	Bottom-up dynamics	Municipal and state governments	Fence-building
		Top-down dynamics	Federal government	Fence-building

Source: Own elaboration based on the “temporal approach” of this article.

Causal Mechanisms

Causal mechanisms are defined here as linkages that connect inputs (for example, strategies) to the outcomes of a process (for example, recentralization). Thus, causal mechanisms are “relational concepts” ( Fallleti and Lynch 2009: 1147) that interact with contexts and dynamics to produce an observable outcome. In Brazil, the recentralization trajectory was shaped by mechanistic causations that linked strategies to recentralization outcomes. I identify four causal mechanisms – fiscal profligacy behavior, fiscal prudence behavior, status quo bias , and institutional layering – that had specific influences on the recentralization process, as described in Table 2 .

Table 2
: Causal Mechanisms behind Brazil’s Recentralization

Causal Mechanism	Description	Influence on recentralization	Explanatory citations
Fiscal profligacy behavior	The exercise of subnational fiscal autonomy through high subnational expenditure and indebtedness	Reinforces the decentralization to subnational governments; favors the expenditure powers of the subnational governments vis-à-vis the fiscal regulatory powers of the federal government	Falleti, 2005 ; Rodden, 2006 ;
Fiscal prudence behavior	Reduction of costs of the high subnational expenditure by imposing limits and sanctions to subnational governments with high indebtedness	Imposition of limitations and sanctions on subnational expenditure and indebtedness reducing macroeconomic instability and strengthening the position of the federal government vis-à-vis subnational governments	Melo, 2005 ; Rodden, 2006
Status quo bias	Partisan polarization and institutional opportunity to delay parliamentary procedures to vote bills	Delays in the approval of laws that regulate constitutional provisions giving ample autonomy to subnational governments	Binder, Lawrence and Smith, 2002 ; Wawro and Schickler, 2013
Institutional layering	Creation of new institutions juxtaposing old institutions. Neglect and redirection of old institutions with the creation of new ones that reestablish new goals and dynamics	Institutionalize vertical intergovernmental relations through the imposition of either a top-down or bottom-up dynamic	Mahoney and Thelen, 2010

Source: Own elaboration based on the “temporal approach” of this article and on the scholarly literature.

The first mechanism, fiscal profligacy behavior , is the collective behavior of subnational governments to spend beyond their fiscal capacity. In Brazil, fiscal profligacy was driven by a primarily institutional incentive given that, until the mid-1990s, Brazil had no constitutional provisions, legislation, or binding norms that limited the spending of subnational governments or punished them for excessive spending. In other words, the overspending of subnational governments, which in turn relied on fiscal transfers from the federal government to cover their expenses, thus encouraging subnational indebtedness and ultimately forcing the federal government to bail out the subnational governments ( Rodden, 2006 ). Fiscal profligacy is a causal mechanism because it links the sequence of decentralization with the empowerment of subnational governments in the initial period of decentralization. In an analysis of the importance of the decentralization sequence , Falleti (2005) shows that fiscal decentralization, as the first decentralized layer in Brazil, benefited subnational governments by creating high expenditure autonomy, which substantially increased their bargaining power with the federal government to obtain bailouts and debt restructuring.

Fiscal prudence behavior is a mechanism that manifests in the collective action of presidents and members of parliament who try to limit high subnational expenditure and reduce indebtedness. Under conditions of macroeconomic instability – which in Brazil were partly attributable to subnational government overspending – national politicians try to create institutional changes that limit subnational expenditures and impose fiscal discipline. As such, fiscal prudence behavior represents an attitudinal change toward the federal government’s prior inability to limit subnational spending and its role as a lender of last resort to subnational governments. Fiscal prudence behavior operated as a causal mechanism in Brazil because it connected the beginning of a new sequence of laws that imposed restraints on the expenditure powers of subnational governments to the augmentation of the federal government’s influence in intergovernmental fiscal matters. In other words, after a change in the sequence of legislation approval this mechanism helped to increase the federal government’s leverage over territorial politics. This mechanism pushed back against the decentralization phase, in which the federal government showed a relative weakness as compared to the subnational governments in fiscal matters.

Status quo bias is a mechanism by which the parliamentarians who act as gatekeepers use legislative process instruments (for example, protracting legislative debates, proposing multiple amendments to the bill) on purpose to delay the approval of laws. As such, this mechanism links the strategy of foot-dragging with the effect of maintaining existing institutions, thereby preserving existing territorial power relations. This mechanism can occur when minority obstruction instruments (for example, filibuster, cloture) allow parliamentarians to introduce dilatory motions to block the approval of laws. The status quo bias is often used when there is a deep divide in partisan preferences ( Binder, Lawrence and Smith, 2002 ), or there is close alignment of party policy lines with strong policy preferences among the voters (Wawro and Schickler, 2013). However, parliamentarians can overcome the status quo bias by aligning interests and institutional opportunities to speed up the parliamentary approval procedure ( Thelen, 2003 ).

The Brazilian Constitution requires a supermajority to approve bills, so intra- and inter-branch conflicts allow minority groups to take advantage of the status quo bias mechanism. The approval process may be derailed by particular institutional configurations and legislative dynamics (for example, power asymmetry between the congressional chamber, the specific congressional chamber in which legislation is initiated, and the urgency of approval; Hiroi, 2008 ). The difficulty of maintaining government coalitions in a fragmented multiparty system favors obstructionism in legislation approval procedures. More specifically, between 1983 and 2007, the conflicts between the presidents with their own coalition partners and the opposition contributed to the long duration of approval in the Brazilian National Congress ( Hiroi and Renno, 2017 ).

Institutional layering is a mechanism that consists of creating new institutions that operate side by side with old institutions. Mahoney and Thelen (2010) suggest that gradual institutional change via layering occurs when strong veto players such as subnational governments block changes in old institutions, leading to the creation of new institutions alongside the old ones. Institutional layering is a causal mechanism because it connects the crafting of new institutions with an increase in the coordination powers of the federal government. In Brazil, the creation of intergovernmental forums represents a form of institutional layering in which a strong veto player – the federal government – imposed a top-down dynamic by creating non-binding norms, primarily in the fiscal and administrative layers, that slowly increased the coordination capacity of the Brazilian federal government. As the new intergovernmental forums gradually produced norms that subnational governments could implement at their own discretion, the dynamics shifted; institutions gradually neglected the old rules and adjusted to the new ones.

Figure 1 illustrates the timing of the mechanisms above mentioned in the Brazilian recentralization trajectory. It was initiated and sustained by fiscal prudence and then by status quo bias – both of which are considered self-reinforcing mechanisms insofar as they strengthened the decentralization process and advanced the interests of subnational politicians. Yet, the strategies of political parties had different strengths over time. For instance, the foot-dragging strategy associated with these mechanisms lost its strength, and as it waned, the self-reinforcing mechanisms diminished, creating a window of opportunity for negative feedback mechanisms to emerge (Falleti and Mahoney, 2015; Falleti, 2016). In this context, the mechanisms of fiscal prudence and institutional layering were invoked under the pace-setting and fence-building strategies that represented a reaction against decentralization and created the alternative centralization process.

Figure 1
: Timing of Causal Mechanisms along Brazil’s Recentralization

CONTRIBUTIONS TO THE DE/CENTRALIZATION LITERATURE

Table 3 summarizes the Brazilian literature on decentralization, which attributes decentralization to several key factors: the bottom-up nature of the drafting of the 1988 Constitution during the transition to democracy ( Abrucio, 1998 ; Souza, 1998 ), the vertical intergovernmental conflict ( Montero, 2001 ; Falleti, 2005 ), the fiscal and social policy autonomy acquired by subnational governments before and during democratization ( Arretche, 2004 ; Falleti, 2005 ; Rodden, 2006 ), and the federal institutional design that granted important powers to state governments ( Abrucio, 1998 ; Samuels and Abrucio, 2000 ; Samuels, 2003). Centralization, on the other hand, is attributed to the following causes: the Brazilian Chief Executive’s ability to increase political leverage and policy capacity through constitutional prerogatives ( Arretche, 2007 , 2009 , 2012 ; Melo, 2005 ), the policy inefficiencies under decentralization ( Almeida, 2005 ; Abrucio, 2005 ; Dickovich, 2007; Fenwick, 2009 ; Palotti and Machado, 2014 ), and the need for effective social policies and programs ( Rich and Gómez, 2012 ; Rodrigues-Silveira, 2011 ; Sugiyama, 2011 ).

Table 3
: Summary of the Brazilian Decentralization and Centralization Literature

	Authors	Causes	Main Actors	Main Institutions-Instruments
DECENTRALIZATION	Abrucio, 1998	Democratization	State governments	Governor’s influence over deputies at the national congress; Low checks and balances on the powers of the governors
	Arretche, 2004	Social policy autonomy of subnational governments (especially in the health sector)	Federal, state and municipal governments	Different capacity of subnational governments to exercise social policy autonomy
	Cheibub. Figueiredo and Limongi, 2009	Institutional powers of the executive branch of states	Federal and state governments, national congress	The powers of the governors and local political elites are limited by the centralized legislative procedures of the national congress and the agenda power of the national executive
	Falleti, 2005	Decentralization sequence: fiscal decentralization occurring before administrative decentralization	Federal and state governments, national congress	Temporal sequence of law approval
	Rodden, 2006	Inherently divided sovereignty of economic unions	Federal and subnational governments	Subnational fiscal deficits, federal bailouts, fiscal discipline, degree of subnational fiscal autonomy and dependence on the federal government
DECENTRALIZATION	Samuels and Abrucio, 2000	Subnational actor’s power over democratization	Subnational governments	Constitution drafting, fiscal autonomy of subnational governments, and nonconcurrent subnational elections
	Souza, 1998 , 2001	Democratization	National congress, central and subnational governments	Constitution drafting, separation between political and fiscal powers
	Montero, 2001	Vertical intergovernmental conflict	Federal and subnational governments	Pace of decentralization, macroeconomic crises, vertical coordination
CENTRALIZATION	Almeida, 2005	Effort to increase federal policy effectiveness	Federal government	Centralizing fiscal and social policies in a context of cooperative federalism
	Abrucio, 2005	Attempt of the federal government to increase policy coordination	Federal government under the Fernando Henrique Cardoso administration	Monitoring and evaluation of subnational governments, increase of regulatory powers of federal government, creation of intergovernmental coordination fora
	Arretche, 2007 , 2009 , 2012 , 2013	Legislation at the national congress, limited veto-powers of subnational governments	Members of parliament	National executive, legislative branches and other federal institutions
	Dickovich, 2007	Desire of the federal government to diminish the influence of governors	Federal and municipal governments	National executive and legislative branches
CENTRALIZATION	Fenwick, 2009	Desire of the federal government to diminish the influence of governors	Federal and municipal governments, policy actors	Social assistance policies, partnership federal-local governments
	Melo, 2005	Pursuit of macroeconomic stability	Presidents	National executive branch
	Palotti and Machado, 2014	Attempt to improve federal coordination in the educational, health and social policy areas	Federal government	Intergovernmental coordinating forums, incomplete federal pact
	Rich and Gómez, 2012	Policy efficiency in the health sector	Federal government	Expansion of regulatory capacity of the federal government
	Sugiyama, 2011	Social policy diffusion	Federal and municipal governments	Vertical and horizontal intergovernmental institutions

Source: Own elaboration based on the vertical intergovernmental analysis of this article.

In a departure from most of the scholarly work on Brazilian de/centralization, this article identifies the main cause of de/centralization as the strategic pursuit of territorial interests, which initially used temporal modalities of sequence and speed during the approval process for de/centralizing laws and then constructed intergovernmental institutions. In a longitudinal analysis of the processes of de/centralization, this article discloses the mechanisms that allowed the temporal and institutional strategies in this specific political context to shape a process that culminated in recentralization as the final outcome.

RECENTRALIZATION STRATEGIES AND TERRITORIAL INTERESTS IN BRAZIL

This section has three main purposes: (i) to provide an overview of the recentralization trajectory that evolved across three distinct periods of time; (ii) to dissect the linkages between different strategies, causal mechanisms, and their overall effects; and (iii) to explain the prevalence of the federal government’s strategies in the legislative and vertical intergovernmental spheres.

As far as the unfolding of the de/centralization processes is concerned, it is important to mention four general temporal patterns, identified in Figure 2 . First, the temporal-based strategies of pace-setting and foot-dragging were in conflict with each other, in such a way that only one of the two temporal strategies could dominate within each period of the de/centralization processes.

Figure 2
: Overview of De/centralization in Brazil

Second, the emergence of foot-dragging (1983–1994) and its eventual transition to pace-setting (1995-2007) all started with an attempt to impose a sequence of de/centralization. This highlights the importance of sequence as a required initial temporal modality to impose a dominant temporal strategy. The decentralization process started with a specific sequence , fiscal decentralization ex-ante the administrative one, before the 1988 Constitution. This sequence was reinforced by the fast approval of decentralizing laws under an autocratic regime, which led to the implementation of weak foot-dragging (number (1) in Figure 2 ).

Third, after the 1998 Constitution, political parties continued to implement either weak or strong versions of the temporal strategies. Starting in 1988, with the promulgation of the Constitution, the foot-dragging strategy was weakly implemented since the sequence of decentralization was altered to fiscal centralization before administrative centralization. Despite the weakness of its implementation, foot-dragging prevailed since the speed of law approval was slow (number (2) in Figure 2 ). The pace-setting strategy, on the other hand, was strongly implemented as politicians started centralization at the fiscal layer and achieved rapid approval of centralizing laws starting in 1995 (number (3) in Figure 2 ).

Finally, the fence-building strategy was pursued as early as 1991, but the institutions built to mediate policymaking between different government levels somehow became increasingly more robust with the incremental creation of top-down intergovernmental forums. It is worth noting that fence-building and pace-setting were implemented simultaneously, though fence-building started with the creation of two forums, one with top-down and another with bottom-up dynamics (number (4) in Figure 2 ). The fence-building strategy shifted to the strong implementation after 1995 (number (5) in Figure 2 ) with the successive creation of additional top-down forums.

Figure 2 provides a bird’s eye view of de/centralization in Brazil, depicting both the strong and weak implementations of the strategies at different points in time.

Periodization and the Unfolding of the De/centralization Processes

For analytical purposes, the Brazilian recentralization trajectory is divided into three periods. Period I (1974-1988) can be considered a critical juncture at which the main parties at the time, PMDB and PFL, joined efforts to create a constitutionally decentralized country. Period II (1989-1994) reinforced these early events, and conflict started to emerge due to diverging territorial interests between the main parties. Under this divergence of interests, PMDB and PFL embraced the foot-dragging strategy. Lastly, Period III (1995-2007) was dominated by centralization as the result of a chain of events that occurred in reaction to the earlier period; the strategies of pace-setting and eventually fence-building took shape and became prevalent. PSDB, and later PT, controlled the National Executive Branch and sought centralization.

Each recentralization period was associated with a change in the bargaining interactions among political players. Period I was marked by the tactical alignment of preferences among political players, which culminated in the creation of new macro institutions (for example, the 1988 National Constitution). There was a consensus around the basic institutional design of the new Constitution, which inherited many features from the 1946 Constitution ( Limongi, 2008 ). The cooperation between political parties was motivated by the democratization conjuncture, which compelled politicians to reach a pact for the possible consolidation of a democratic political regime. This resulted in alignment between PMDB and PFL in the territorial commission of the constitutional assembly ( Souza, 2001 ). During the constitutional drafting process, the opposition party to the outgoing regime, Brazilian Democratic Movement (Movimento Democrático Brasileiro – MDB), passed Constitutional Amendment 27/85, also known as the Airton Sandoval Amendment, a critical decentralization law that increased the amounts of revenue distributed both to the states and to the municipalities. This legislative measure marked the beginning of the sequence of fiscal decentralization ex-ante the administrative one, which was an essential mechanism within the foot-dragging strategy.

Period II began after the promulgation of the Federal Constitution, which enshrined several subnational governments’ rights. This period was marked by the broad implications of subnational expenditures for the Brazilian macroeconomic stability. This period also coincided with the three final years of José Sarney’s term in office (1985-1990), after which two presidents had short terms of approximately two years each. Throughout this period, the implementation of constitutional provisions caused several conflicts between the national and subnational governments, most notably on the fiscal front. Intergovernmental relations were paralyzed ( Samuels and Abrucio, 2000 ).

In the final years of the Sarney government, there was the proposition of two critical laws regarding territorial matters. PDT proposed the Basic Educational Law, which would create a decentralized education system, and PMDB proposed the Camata Law, which would limit public expenditures for both state and municipal governments¹⁴. The foot-dragging strategy substantially delayed both laws.

Under the first democratically elected president, Fernando Collor de Mello (1990-1992), legislative proposals from the National Executive Branch required intense negotiations with the National Congress, and this slowed down the approval rate. This happened with Law 8080, regarding health. Macroeconomic stability became the main priority under Itamar Franco’s term in office (1992-1994), which followed an unprecedented institutional crisis marked by the political impeachment of Collor de Mello. Although Franco did not belong to any political party, he relied on the support of a broad coalition that allowed his administration to adopt important measures to promote macroeconomic stability¹⁵. This, in turn, led to the adoption of the Social Emergency Fund (FSE), which enabled the federal government to increase its revenue¹⁶. This government marked the end of the foot-dragging strategy, a change that aligned with its attempt to end subnational practices that had led to a fiscal imbalance in the first place.

During Period III, with PSDB and PT occupying the national presidency, there was an attempt to carry out centralization through pace-setting and fence-building strategies, which aimed to regulate the subnational governments and coordinate policy implementation, respectively. With the promises of the macroeconomic stabilization plan underway, President Fernando Henrique Cardoso tried to sustain the pace of economic reforms that the previous government began and gathered the support of key political parties to carry out additional reforms¹⁷. The first Cardoso government (1995-1998) witnessed the end of much of the deadlock that surrounded some centralization laws (for example, the Basic Education Law – LDB) and the approval of several laws that affected subnational interests and had failed to pass previously (for example, the reduction of a state tax, Tax on the Movement of Goods and Services – ICMS, and the reduction of constitutional transfers to subnational governments).

From 1995 to 1998, the National Executive Branch proposed six constitutional amendments that were essential for the centralization of federal relations. These amendments were approved before the end of Cardoso’s first term in office. Part of the reason for this faster approval was that in the second half of Cardoso’s term, the party coalition that supported the government controlled 77% of the seats in the House of Representatives, the highest parliamentary support of any government from 1998 to 2007. During Cardoso’s second term in office (1998-2002), although the government coalition had weaker control in the House of Representatives, approximately 40% of the proposed laws (out of the 32 laws identified in this article) were approved. The National Executive Branch increased its efforts to create transparency in public financial accounts. These efforts culminated in the 2000 Fiscal Responsibility Law (LRF), which was one of the main signs that the power of the state governors in Brazil had waned. By the end of Cardoso’s second term in office, the federal government had acquired important regulatory powers.

Under the first term of President Luiz Inácio Lula da Silva (2003-2007), a governing coalition formed from a wide array of parties with different ideological backgrounds. The fence-building strategy had already begun to increase the coordinating role of the federal government, and Lula’s government sought additional instruments to foster intergovernmental coordination for a more effective implementation of policy measures. The Committee of Federative Articulation (CAF), created for this purpose, was designed to improve the relations between central and municipal governments. The initiative was consistent with a fence-building strategy since it aimed to institutionalize intergovernmental relations.

The Legislative Sphere

What was needed to implement these pace-setting and foot-dragging strategies in the legislative sphere? With an understanding of the effects of temporality on the recentralization process, political players used the aforementioned tactics to carry out temporal strategies , taking into account several legislative dynamics that were decisive for both the sequence and speed of law approval.

To recapitulate, the choice of which layer – fiscal or administrative – to decentralize first is critical to the outcome given that the sequence of de/centralization activates different mechanisms. The decentralization of the fiscal layer before the administrative one activates fiscal profligacy behavior , while the opposite sequence , fiscal centralization ex ante the administrative one, triggers fiscal prudence behavior . In Brazil, the fiscal profligacy mechanism was activated in the specific context of the transition to democracy that started in the late 1970s and provided a window of opportunity for political bargaining between the different political elites. During this transition, fiscal benefits were given as a concession by the party politically aligned with the authoritarian outgoing regime, the National Renewal Alliance (Aliança Renovadora Nacional – ARENA), to the opposition party, the MDB, in order to keep a favorable position in the future democratic regime. More specifically, the concession took the form of the approval of laws that would meet the long-lasting plea of subnational governments to increase the revenues as well as the expenditure autonomy of states and municipalities ( Medeiros, 1986 ).

When territorial players centralize the fiscal layer before the administrative one, they try to increase the ability of the federal government to limit subnational expenditure by curtailing their discretion to spend and by imposing subnational expenditure obligations before they grant fiscal autonomy. In addition, this sequence forces subnational governments (that remain dependent on national fiscal transfers) to comply with the federal government’s established expenditure targets and criteria. As such, this sequence has the potential to trigger the fiscal prudence behavior of subnational politicians, allowing the federal government to impose limits on the expenditure of subnational politicians. In Brazil, the fiscal prudence mechanism started as part of an attempt to enact socio-economic reforms ( Melo, 2005 ).

In the Brazilian National Congress, legislative approval rests on several determinants that are useful for our understanding of how temporal strategies delayed or sped up the approval of legislation. Obstructionism triggers the status quo bias mechanism, thereby resisting changes in the territorial politics that had been established by the decentralized arrangements in the 1988 Brazilian Constitution. One can identify conditions under which politicians in Brazil slowed down the process of law approval between 1988 and 1995, thereby decreasing the possibility that the status quo would shift, until new legislation initiated a centralization process in 1995. Hiroi (2008) identifies determinants that allow politicians to obstruct laws in Brazil: the lack of a coalition majority in Congress, the possibility of extensively amending laws, and a long duration of legislative approval, which increases the chances that the proposed law will be withdrawn. These determinants turned easier for the foot-dragging strategy to advance.

However, the status quo bias mechanism might be rendered less effective if the federal government decides to fast-forward legislation. Hiroi (2008) identifies another set of determinants that make it possible to speed up the approval of laws that are critical for the pace-setting strategy: an economic crisis, presidential activism that promotes legislative approval, initiation of the legislative procedures in the House of Representatives, and the inclusion of provisional changes within the bills under consideration.

Figure 3 shows how certain conditions (for example, political regime change, the emergence of elite bargaining to reach a new federal pact) allowed the first sequence of decentralization to empower subnational players by activating the fiscal profligacy behavior mechanism. Also, this figure illustrates how fiscal profligacy arose out of a change in the conditions (for example, policy paradigm shift, success of the macroeconomic stabilization plan) and culminated in the federal government gaining the ability to impose fiscal balance on subnational public accounts. In terms of the relationship between speed and the status quo bias mechanism, Figure 3 displays the obstruction of legislation under specific conditions (for example, absence of a coalition majority in both congressional chambers), which evoked the continued empowerment of subnational governments through the status quo bias . Yet, by speeding up the approval of legislation under certain conditions (for example, economic crisis, presidential activism), players that defended centralization deactivated the status quo bias that had favored subnational governments.

Figure 3
: Conditions and Effects of Temporal-Based Causal Mechanisms in Brazil’s Recentralization

The Vertical Intergovernmental Sphere

How was fence-building pursued in the vertical intergovernmental sphere? The federal government tried to consolidate its coordination powers over vertical intergovernmental relations by institutionalizing this sphere. Institutionalization is best understood as the creation of institutions with a specific mandate that regularly and formally brings together different levels of government for the primary purpose of policymaking. The development of intergovernmental forums in Brazil shows that, in the long run, the Brazilian National Executive Branch offset the temporal-based strategy used after the transition to democracy and brought negotiations concerning the implementation of centrally orchestrated policies to the intergovernmental arena.

In the years after the transition to democracy in Brazil, intergovernmental relations in the country were poorly institutionalized ( Selcher, 1989 ). This is partly attributable to the failure of the 1988 Constitution to create a coordinating framework for interactions between different government levels. Until the mid-1990s, vertical intergovernmental federal relations in Brazil were mainly informal or took place within forums established prior to the democratic regime. Under decentralization, the decision-making process became more dispersed as policy decisions were driven to multiple directions (for example, bottom-up , horizontal, and top-down ) ( Sugiyama, 2011 ). This dispersion of policy authority gave rise to vertical intergovernmental conflict leading, for instance, to the judicialization of territorial politics ( Do Vale, 2013 ). In this context, the institutionalization of vertical intergovernmental relations in Brazil should be viewed as an attempt to align policymaking with the effective implantation of policies.

The interactions between different government levels within these forums reflect the competition between the levels. While they were created with common, cooperative motives, territorial players on each level try to maintain their autonomy and influence other government levels. Within these forums, there are two types of interactions: top-down and bottom-up ( Behnke and Mueller, 2017 ), depending on whether the agenda is set by the National Executive Branch or the Executive Branch of subnational governments, respectively.

In the mid-1990s, the Brazilian federal government progressively became the effective agenda-setter for most active coordination within the first intergovernmental forums created ( Palotti and Machado, 2014 ). However, the coordination initiatives of the federal government have not always succeeded due to the lack of management capacity and resources ( Grin and Abrucio, 2017 ). Despite the hurdles associated with coordination in Brazil, this article interprets the growing institutionalization of the intergovernmental sphere as the federal government’s attempt to consolidate its coordination powers.

TERRITORIAL PARTY POLITICS IN BRAZIL

This section has three purposes: first, to explain differences in the territorial preferences of the decentralizing and centralizing party blocs; second, to explain how party nationalization affected each bloc’s ability to advance its territorial interests in the legislative sphere; and third, to provide empirical support for the argument that parties with a large congressional coalition and a chief executive committed to the party agenda were better able to enforce their strategies to approve centralizing legislation.

In Brazil, there is a lack of institutional instruments within the party system in favor a convergence of interests between parties that hold different territorial interests, as Riker (1964) envisioned, such that diverse party interests converge within one federal system. Indeed, Brazilian territorial party politics and the party system are notorious for their inability to accommodate territorial interests. Cooperation among parties in pre- and post-election coalitions is harder to achieve under Brazil’s fragmented multiparty system, in which state-level elections affect national legislative elections, and state-level party branches exert power over the entire political party ( Borges, 2019 )¹⁸.

Party nationalization has played a critical role in the division of decentralizing and centralizing party blocs in Brazil because parties acquire different territorial interests as they nationalize.¹⁹ The differences between both blocs have become even more pronounced due to disparate motivations behind political competition (programmatic versus clientelistic politics) and the ability of the president’s party to build and sustain a congressional coalition.

This article holds that features of the Brazilian party system help to explain why the centralizing bloc was able to follow the pace-setting strategy more forcibly than the decentralizing bloc was able to implement their foot-dragging strategy. As this article shows below, the centralizing bloc of parties, which were more nationalized and organized around a programmatic agenda, benefited from strong presidents who led a more cohesive multi-party coalition in Congress. These elements contributed to the centralizing bloc’s successful advancement of its territorial strategies.

Brazilian Party System and Territorial Politics

Although Brazil has shown higher levels of party nationalization in recent years ( Jones and Mainwaring, 2003 ; Borges, 2015), parties have nationalized differently over time. PT and PSDB, among others, increased their vertical nationalization, while PMDB became less vertically nationalized because it was trying to be more competitive in state-level elections (Borges, 2015:677). The different types of nationalization have made it difficult for the parties to converge around a national agenda.

These two-party blocs have important differences both in their internal organization and in their competitiveness²⁰. PT’s internal organization is considered the most centralized of the parties analyzed here, while PMDB is the least centralized ( Ribeiro, 2013 , 2014 ). Each party’s internal organization has consequences for its approach to elections. PMDB and PFL often are more competitive in subnational elections, while PSDB and PT dominate presidential elections – every president from 1995 to 2016 came from one of the two parties (Melo, 2015).

The parties also differ in their motivation (clientelistic versus programmatic) to hold political posts, and once in office, they have different abilities to build and sustain a congressional coalition in support of their initiatives²¹. Some legislators base their careers around local clientelistic networks. Traditionally, PFL and PMDB (among other parties) were known for establishing clientelistic ties, so they were viewed as more “office seeking” ( Borges, 2011 ) and lacking “programmatic coherence” ( Melo and Câmara, 2012: 102). By contrast, PT and PSDB (among others) were known for programmatic motivations, which enabled a more cohesive approach to their programmatic agenda²². The tendencies of each party bloc to embrace either clientelistic or programmatic politics affected their implementation of territorial strategies. In general, office-seeking parties are more inclined to sacrifice their territorial agenda to obtain particularistic benefits, while programmatic parties might be more committed to their territorial agenda. Thus, during the recentralization trajectory, the centralizing bloc was more willing to defend its centralization agenda, while the decentralizing bloc was more inclined to negotiate its commitment to the decentralization agenda.

Within the Brazilian multiparty system, coalition politics can enhance the president’s policy-making ability ( Figueiredo and Limongi, 2000 ; Figueiredo, 2007)²³, and the success of this strategy depends to some extent on the president’s leadership capacity regarding cabinet formation. Presidents who distribute cabinet positions more evenly across the parties that compose the government coalition have been more successful at building a cohesive coalition to support the government’s lawmaking initiatives ( Amorim Neto, 2002 ), taking for granted that parties can impose party discipline in the National Congress ( Neiva, 2011 ). For instance, the post-1988 presidents from PSDB and PT had robust lawmaking capacities because they united their coalition through cabinet distribution, which in turn created cohesive vote support for the Executive’s legislative initiatives ( Amorim Neto, 2019 ). Thus, presidents from the centralizing bloc were able to both build and sustain a coalition behind their territorial strategy.

In sum, two important attributes of party politics in Brazil enabled the centralizing parties to oppose the decentralization process by initiating a centralization process. First, PT and PSDB had a clear programmatic agenda that revolved around their pursuit of macroeconomic stability and the establishment of social redistributive programs ( Melo, 2005 ), which required the centralization of both power and authority. Second, coalition politics after 1995 forced more office-seeking parties with lower programmatic coherence (i.e., PMDB and PFL) to sacrifice their decentralization agenda so that they could benefit from the distribution of cabinet positions in the PSDB and PT governments.

The Prevalence of Centralizing Parties

An overview of the 26 de/centralization laws proposed after the promulgation of the 1988 Constitution corroborates the centralizing bloc’s success at using the pace-setting strategy to initiate a sequence of centralizing laws. Most of the centralizing laws were presented when PSDB or PT controlled the National Executive Branch – that is, since 1995, when the pace-setting strategy started to prevail over foot-dragging . From 1995 to 2004, it took an average of 928 days to approve a law, a marked contrast to the 1668-day average under the PMDB term in office (1988-1991) ( Figure 4 ).

Figure 4
: Speed of De/centralization Measures by Proponent

Fewer laws were proposed and approved when a centralizing party did not hold the presidency (i.e., between 1988 and 1995) than when one did (i.e., between 1995 and 2007). Specifically, between 1988 and 2007, PMDB and PFL collectively proposed four laws, while PT, PSDB, and associated presidents proposed 17 laws.

In addition, of the laws proposed by the centralizing bloc, a majority (13 of the 17) were approved within the same presidential term, while all laws proposed under the PMDB president were approved in the subsequent presidential term. Generally speaking, the centralizing bloc’s ability to achieve a faster rate of approval indicates that congressional support was stronger when the president came from centralizing parties. On average, the coalitions formed under the PSDB’s and PT’s national governments held more seats in the House of Representatives (an average of 61.4% from 1995-2007) than the coalitions formed by other governments (an average of 49.8% from 1988-1995) ( Table 4 ).

Table 4
: Decentralization Laws and Presidential Coalitions (1983-2007)

Government Period (party of the president)	Presidential coalition†	Party of the president	Parties in Coalition	% of seats in the lower chamber		Decentralizing law ordered by date of proposal (proponent)	Decentralizing law ordered by date of approval
				Party of the president	Coalition
1983-1988	Autocratic governments	---	---	---	---	Passos Porto Amdt (PDS), Calmon Amdt (PDS), Airton Sandoval Amdt (MDB)	Passos Porto Amdt, Calmon Amdt, Airton Sandoval Amdt
1988-1990 (PMDB)	Sarney 2	PMDB	PMDB-PFL	41,4	63,03	LDB (PSDB)* Camata Law (PMDB), Law 8080 (Executive)	---
1990-1992 (PRN)	Collor 1	PRN	PRN-PFL	5,1	33,94	---	---
	Collor 2	PRN	PRN-PFL-PDS	6,1	34,59	Law 8142 (Executive)	Law 8142
	Collor 3	PRN	PRN-PFL-PDS	8	34,59	Amdt 15/96 (PFL)	Camata Law, Law 8080
	Collor 4	PRN	PRN-PFL-PDS-PTB-PL	6,2	43,54	---	---
1992-1944 (No party)	Franco 1	No party	PFL-PTB-PMDB-PSDB-PSB	---	60	---	---
	Franco 2	No party	PFL-PTB-PMDB-PSDB-PP	---	59,64	---	---
	Franco 3	No party	PFL-PMDB-PSB-PP	---	55,27	---	---
1995-1998 (PSDB)	Cardoso I 1	PSDB	PSDB-PFL-PMDB-PTB	12,5	56,14	Amdt 10/96 (Executive), Amdt 5/95 (Executive), Amdt 20/98 (Executive), Amdt 19/98 (Executive), Amdt 14/96 (Executive), Amdt 41/03 (PCB)*, Amdt 29/00 (PSDB), Amdt 34/01 (PCB)	Amdt 5/95, Amdt 10/96
	Cardoso I 2	PSDB	PSDB-PFL-PMDB-PTB-PPB	16,6	77,19	Kandir Law (PSDB), Amdt 30/00 (PSDB), Amdt 53/06 (PL)*, Amdt 25/00 (PPB)	LDB, Amdt 14/96, Amdt 15/96, Kandir Law, Amdt 19/98, Amdt 20/98
1998-2002 (PSDB)	Cardoso II 1	PSDB	PSDB-PFL-PMDB-PPB	18,3	73,88	LRF (Executive), CFL (Executive), Amdt 27/00 (Executive), Amdt 31/00 (PFL)	LRF, CFL, Amdt 25/00, Amdt 27/00, Amdt 29/00, Amdt 30/00, Amdt 31/00, Amdt 34/01
	Cardoso II 2	PSDB	PSDB-PMDB-PPB	18,2	45,22	Amdt 39/02 (PDT)*	Amdt 39/02
2003-2007 (PT)	Lula I 1	PT	PT-PL-PCB-PSB-PTB-PDT-PPS-PV	18	42,88	Amdt 51/06 (PT), Amdt41/03 (Executive), Amdt 42/03 (Executive)	Amdt 41/03, Amdt 42/03
	Lula I 2	PT	PT-PL-PCB-PSB-PTB-PPS-PV-PMDB	17,7	62,38	Amdt 47/05 (PT), Amdt 58/07 (Executive)	---
	Lula I 3	PT	PT-PL-PCB-PSB-PTB-PV-PMDB	17,7	57,7	---	Amdt 47/05
	Lula I 4	PT	PT-PL-PCB-PSB-PTB-PMDB	17,7	58,28	---	---
	Lula I 5	PT	PT-PL-PCB-PSB-PTB-PP-PMDB	17,7	69,59	---	Amdt 53/06, Amdt 51/06, Amdt 55/07

Figueiredo
• † In Brazil it is possible to observe the formation of different coalitions within a presidential term.
• * Laws that were proponed by parties outside the government coalition

Between 1988 and 1990, although PMDB alone held both the presidency and 41% of the total seats in the House of Representatives, the party proposed few decentralization laws. Similarly, although Itamar Franco (1992-1995) had support from a majority of the Congress, he did not propose or approve any decentralizing law during his term.

Brazilian presidents often formed a multiparty territorial coalition with parties from both the decentralizing and centralizing blocs, thereby transcending territorial interests. Indeed, both PMDB and PFL were part of 16 out of 17 national government coalitions formed between 1988 and 2007 and thus were involved in most of the decisions approved in the recentralization trajectory ( Table 4 ). Nonetheless, if we consider only the executive-initiated legislative proposal, when the president belonged to PMDB (1985-1990), laws were approved more quickly than those laws proposed by the presidents that belonged to PSDB and PT ( Figure 4 ). This happened because parties in the centralizing bloc proposed more constitutional amendments, which required greater parliamentary support than other legislative measures, and because party fragmentation within the National Congress increased over time, making it harder to build support for the legislative measures.

FOOT-DRAGGING AND PACE-SETTING: DETERMINING THE SEQUENCE AND SPEED OF DECENTRALIZATION IN BRAZIL

How did foot-dragging become a dominant strategy in the early stages of the decentralization process, and how was it slowly replaced with the pace-setting strategy after 1995? This article addresses these questions by identifying two temporal patterns that contributed to the increase of the center’s regulatory power: (i) the replacement of the initial sequence , fiscal decentralization ex ante the administrative one, with a new sequence , fiscal centralization ex ante the administrative one; and (ii) the approval of fiscal laws, which were the most disputed laws in the recentralization trajectory, at a faster rate than administrative laws. These patterns are critical to understand the triumph of pace-setting over foot-dragging and the counter-reactive centralization sequence .

Sequence

One sequence emerged as dominant in each of the three periods of the recentralization trajectory. Period I was shaped by fiscal decentralization ex ante the administrative one, which triggered the fiscal profligacy mechanism ( Figure 4 ). In 1983, the Brazilian National Congress approved the Passos Porto Amendment to increase subnational revenues. Two years later, in 1985, the enactment of the Airton Sandoval Amendment enlarged the tax bases of states and municipalities while increasing their revenue through federal government transfers.

As constitutional provisions were implemented in Period II, the federal government initiated a new sequence of fiscal centralization ex ante the administrative one by prioritizing that approval of laws that would increase the federal government’s regulatory power. This sequence began with the Camata Law of 1995. Finally, Period III witnessed the reinforcement of this new sequence and the rise of fiscal prudence behavior.

The entire trajectory of the de/centralization processes in Brazil involved 19 alternations between fiscal and administrative layers ( Figure 5 ). It is interesting to note that although the sequence of fiscal decentralization ex ante the administrative one lasted for only a short time (until the beginning of Period II), it had an enduring effect in that subnational over-expenditure was widely practiced in Brazil until it was curbed by a series of legislations in the late 1990s.

Figure 5
: Sequence of De/centralization in Brazil (ordered by date of approval)*

Speed

In Period II, speed was used to delay the approval of laws, but in Period III, the centralizing parties managed to propose and approve legislation at a greater rate than the decentralizing parties, as Figure 6 demonstrates.

Figure 6
: Speed of Approval of De/centralization Laws in Brazil (ordered by date of proposal before the Lower House)

Across the full recentralization trajectory, the average approval speed of de/centralization laws was actually fastest in Period I – an average of 167 days for the three decentralization laws proposed in this period – because institutional constraints relaxed under the regime change. This rate slowed down considerably in Period II (an average of 1,318 days per law), but it increased again in Period III as the federal government pushed for faster approval (an average of 410 days), even though the measures were clearly counter-reactions to previous decentralization measures.

The speed of approval in Period III indicates that the federal government and the centralizing bloc collectively gained leverage in federal relations. This leverage is clearly reflected in the speed of approval of fiscal laws as compared to the approval of the administrative ones, which can be appreciated by stratifying the speed of approval by the type of law: from Period II to Period III, the average approval speed of fiscal laws decreased from an average of 961 days in Period II to 611 days in Period III, while the average approval speed of administrative laws decreased from 1300 days to 1184 days.

FENCE-BUILDING: THE INSTITUTIONALIZATION OF INTERGOVERNMENTAL RELATIONS IN BRAZIL

This section shows that vertical intergovernmental relations in Brazil have become more institutionalized under a top-down interaction such that territorial issues increasingly played out in the vertical intergovernmental sphere. These issues enabled the center to consolidate its growing influence over territorial dynamics.

The Institutionalization of Intergovernmental Relations in Brazil

Brazilian territorial dynamics initially developed under a Federal Constitution that promoted power-sharing among different government levels yet provided few coordinating instruments to manage federal relations ( Grin and Abrucio, 2017 ) and insufficient incentives for collaboration ( Arretche, 2004 ). This article argues that temporal-based strategies became less effective as intergovernmental relations were increasingly institutionalized through the creation of more robust intergovernmental structures.

Under the fence-building strategy, institutionalization strengthens the role of technocrats. Hence, in Brazil, cooperative intergovernmental relations became more technical and policy-oriented over time ( Palloti and Machado, 2014 ). The roles of politicians in intergovernmental structures may become more limited as the institutionalization of this sphere provides opportunities for bureaucrats to increase their own influence ( Trench, 2006 ; Behnke and Mueller, 2017). Even though this sphere tends to be less political than the legislative one, the different government levels are still competing with one another to influence intergovernmental interactions.

In Brazil, the centralizing bloc sought to institutionalize the intergovernmental sphere to increase coordination under formal institutions that would limit subnational autonomy. Specifically, it was possible to achieve institutionalization through the creation of instruments that coordinated policies by imposing a top-down interaction. Most of the intergovernmental institutions were formalized under the PSDB and PT presidencies (for example, CAF, CIB, and CIT) or were inherited from the autocratic regime before 1988 and maintained with some adjustments (for example, Confaz, CNE)²⁴.

The forums analyzed here were formal intergovernmental institutions created under presidential decrees or ministerial directives, and they had official, regular meetings to bring together representatives of at least two government levels. As such, this article excluded from the analysis any forum that convened on an ad hoc basis and/or held meetings not organized under a statute. The first intergovernmental forums in democratic Brazil were specific to the health and education sectors, designed to foster policy coordination to improve the heterogeneous quality of service delivery across the country. Eventually, there was the creation of non-sector-specific forums such as the CAF with an open federal agenda.

An analysis of the five intergovernmental forums mentioned above reveals that the intergovernmental sphere has become increasingly robust through institutionalization. Forums reported an average of about 33 meetings a year between 1990 and 1995, which grew to 45 between 1995 and 2000. From 2003 onward, the formal intergovernmental forums averaged 58 meetings a year.

With the advancement of centralization, federal relations became more robust ( Figure 7 ). In Period I, decentralization occurred in an intergovernmental sphere that had few institutional constraints because there were few established intergovernmental forums in Brazil. This scenario began to change in 1991, when the federal government started to improve its relative intergovernmental bargaining position.

Figure 7
: Evolution of Intergovernmental Forums in Brazil (number of official meetings per year)

In Period II, there was a conflict between national and subnational interests, and intergovernmental forums were slowly established as part of a clear attempt to coordinate intergovernmental relations in Brazil ( Abrucio, 2005 ). CIT and CIB, created in 1991 and 1993, respectively, were introduced through ministerial directives to increase coordination in the health sector, while CNE was conceived as a result of the 1995 Basic Law on Education.

In Period III, there was coordination in the fiscal sector, which had been an area of great conflict since the transition to democracy. The growing institutionalization of the intergovernmental sphere in the fiscal sectors started in 1997 with the creation of Confaz, which was put into place so that state and municipal governments could meet to align the norms and rules of the national tax system. In 2003, the Brazilian presidency created CAF to optimize the coordination role of the federal government across a diverse set of issues. In contrast to other forums, CAF is not sector-specific. It has a broad mandate and can coordinate policies in many sectors. Given that it was conceived to strengthen center-municipal relations, a notable characteristic of CAF is the exclusion of state governors from the meetings.

Top-Down versus Bottom-Up Interactions

Each forum is dominated by either a top-down or bottom-up interaction (Behnke and Mueller, 2017). While a top-down interaction signals the power of the National Executive Branch to act according to centrally established guidelines, a bottom-up interaction indicates the attempts of subnational executives to influence federal government policies. The article holds that in the long run, the Brazilian federal government established top-down interactions as the norm, and this was motivated by the desire to consolidate the federal government’s coordination powers.

The directionality of an interaction is defined by four criteria: (i) who establishes the forum, (ii) how the policy coordination proposals are originated, (iii) how decisions are made, and (iv) whether decisions made in the forums are binding. In a top-down interaction , the forum is created and coordinated by the center, the federal government proposes binding decisions, and there is no possibility of veto by subnational entities.

The manner in which a forum is created reveals much about its dynamics. If a forum is created formally through a presidential decree, then it is likely its rules and norms were established on the federal government’s own terms, whereas a forum established through a legislative measure might have negotiated the rules and norms through ample debate. In addition, the specific instrument used for the forum’s creation indicates which institutions have the potential to influence vertical intergovernmental relations.

The presidency of each forum reflects which institutions have the greatest potential to influence decision-making within it. If the federal government controls a forum’s presidency, then most of its activities are likely coordinated under a top-down interaction . If, instead, a subnational government controls the presidency, then coordination likely follows a bottom-up interaction .

Finally, the type of intergovernmental interaction is evident in the outcomes of the deliberations – i.e., whether the outcomes are recommendations or agreements – and whether decisions are enforced through binding or nonbinding instruments. For instance, a binding agreement reinforces the interactions within the forum, while a non-binding recommendation comes with little enforceability.

These criteria were used to assess the five permanent and legally established forums in Brazil. Overall, this article finds that vertical intergovernmental relations favored a top-down interaction . Four of the five forums (excluding CIB) are dominated by a top-down interaction (as seen in Table 5 ), supporting the prediction that institutionalization reinforced the coordination powers of the federal government.

Table 5
: Types of Vertical Intergovernmental Interactions

	Formal initiative to create the forum	Instrument of formal creation	Who proposes	Output of meetings	Binding decisions	Type of interaction
CNE	Central gvt	Presidential decree/ Complementary law	Education Ministry	Recommendations	No	Top-Down
CONFAZ	Central gvt	Presidential decree/ Complementary law	Finance Ministry	Agreements	Yes	Top-Down
CAF	Central gvt	Presidential decree/ Complementary law	Presidential Office	Agreements	Yes	Top-Down
CIT	Central gvt	Ministerial directive	Health Ministry	Agreements	Yes	Top-Down
CIB	Subnational gvt	Ministerial directive	States and Municipalities	Agreements	Yes	Bottom-Up

Source: Own elaboration based on the vertical intergovernmental analysis of this article.

As a general pattern, the National Executive Branch created four of the aforementioned five forums, either through a presidential decree or a ministerial directive. Also, institutions of the National Executive (for example, ministries, the presidential office) preside over the main meetings, which generate binding agreements. Except for CNE, which produces non-binding policy recommendations, these patterns appear in all the top-down forums – including both the general coordination forum (i.e., CAF) and the sector-specific forums (i.e., Confaz and CIT).

The only forum with a dominant bottom-up interaction is CIB, regarding health. Although it was created and regulated by the federal government, the meetings are managed by states and municipalities.

CONCLUSION

This article applied a temporal approach to territorial politics in Brazil and showed that politicians used temporality to shape the de/centralization processes from 1983 to 2007. In this approach, the modalities of temporality – namely, sequence and speed – became means to achieve political aims, and as such, temporal modalities contributed to recentralization as the final outcome of the de/centralization processes.

The temporal modalities of sequence and speed were critical for first advancing decentralization and then introducing centralization as a counter-reaction. However, temporal strategies alone were not sufficient to achieve recentralization. This outcome also relied on the advancement of the institutional strategy of building intergovernmental institutional structures.

To recapitulate, this article discussed three strategies to explain the interactions of intergovernmental players during the de/centralization processes. Subnational governments imposed foot-dragging as their dominant strategy to achieve decentralization (1983-1994) until the federal government introduced pace-setting , which ultimately dominated the foot-dragging strategy (1995-2007). The Federal Executive Branch also pursued fence-building simultaneously with pace-setting to maintain its own power in the centralization process, and this strategy increased the institutionalization of intergovernmental dynamics.

The conceptualization of these strategies made it possible to identify the mechanisms that drove the process of Brazilian recentralization. The initial mechanisms that reinforced decentralization were fiscal profligacy and the status quo bias . Then, the subsequent mechanisms of fiscal prudence and institutional layering eroded decentralization and incentivized centralization.

As the outcome of decentralization- cum -centralization, the recentralization process witnessed several distinct phases and patterns: (i) the Federal Executive Branch more actively used its agenda-setting powers that were closely linked with temporal strategies pursued over time; (ii) within the context of the party blocs’ strategic uses of temporality, this article holds that temporal strategies – foot-dragging and pace-setting – were used in an initial period of low institutionalization of intergovernmental relations; and (iii) temporal strategies eventually were offset by the institutionalization strategy – fence-building – which ensured that intergovernmental relations were driven primarily by top-down interactions .

Although the article leans heavily on the idea of the decentralization-centralization dichotomy, it provides a more complex picture by showing that there is a via media between both poles. First, the article proposes that federal intergovernmental relations changed gradually. Decentralization followed by centralization led to a mode of intergovernmental relations that differed from that of the pre-decentralization period. Second, although the main outcome of recentralization was stronger regulatory and coordination powers for the federal government, the subnational governments retained some autonomy. Third, the article suggests that the intergovernmental sphere has become more robust and institutionalized over time through the creation of forums, which, despite the predominance of top-down interactions , increasingly focus on technical issues.

Methodologically, the temporal approach presented here suggests that the chain of events that gradually shifted federal relations in Brazil is embedded in a political process in which temporal elements such as sequence and speed help territorial players to act strategically. As such, the federal changes in this article are portrayed as part of a flexible trajectory subject to shifts in direction and even outright reversals due to the strategies of political players. This interpretation of federal changes in Brazil post-1988 contributes methodologically to the Brazilian federal literature in the following manners: (i) it offers an alternative explanation for the ongoing yet gradual changes in Brazilian federalism; (ii) it provides a framework to understand counter-reactions to decentralization; and (iii) it explains how sequence and speed are influenced by national and subnational players that, over time, have exerted different influence on intergovernmental relations.

This article was unable to address several important questions that fell outside the scope of this research but that deserve attention in future research. For example, it did not systematically discuss how recentralization transformed modes of intergovernmental relations (for example, conflictive, non-hierarchical). By associating phases of de/centralization with interaction modes, it would become clearer whether, for example, there is more or less territorial conflict in Brazil over time. Another critical question is whether the framework above is applicable to policy processes in specific sectors, perhaps enabling the identification of the temporal strategies of different groups of policy players.

This article tries to add to the scholarly perspectives and debates concerning the changes in Brazilian federalism over the course of de/centralization. By introducing elements of temporality to the Brazilian recentralization trajectory, it lays the groundwork for an alternative explanation for the post-1988 developments in territorial politics in Brazil.

* Acknowledgement

the author would like to thank the anonymous reviewers for their contribution in providing insightful and detailed comments that were critical to improve the article’s argument. This work was supported by the Hankuk University of Foreign Studies Research Fund.

REFERENCES

• ABRUCIO, Fernando. (1998), Os barões da federação: os governadores e a redemocratização brasileira. São Paulo: Hucitec.
• ABRUCIO, Fernando. (2005), “A coordenação federativa no Brasil: a experiência do período FHC e os desafios do governo Lula”. Revista de Sociologia e Política, n^o24, pp. 41-67.
• ALMEIDA, Maria Hermínia Tavares. (2005), “Recentralizando a Federação?”. Revista de Sociologia e Política, n^o24, pp. 29-40.
• AMORIM NETO, Octavio. (2002), “Presidential cabinets, electoral cycles, and coalition discipline in Brazil”. In: S. Morgenstern; B. Nacif (eds.), Legislative Politics in Latin America. Cambridge: Cambridge University Press, pp. 48-77.
• AMORIM NETO, Octavio. (2019), “Cabinets and coalitional presidentialism”. In: B. Ames (ed.), Routledge Handbook of Brazilian Politics. New York and London: Routledge, pp. 293-312.
• ARRETCHE, Marta. (2004), “Federalismo e políticas sociais no Brasil: problemas de coordenação e autonomia”. Revista São Paulo em Perspectiva, vol. 18, n^o2, pp. 17-26.
• ARRETCHE, Marta. (2007), “The veto power of sub-national government in Brazil: political institutions and parliamentary behaviour in the post-1988 period”. Brazilian Political Science Review, n^o2, pp. 40-73.
• ARRETCHE, Marta. (2009), “Continuidades e descontinuidades da federação brasileira: de como 1988 facilitou 1995”. DADOS – Revista de Ciências Sociais, vol. 52, n^o2, pp. 377-423.
• ARRETCHE, Marta. (2012), Democracia, centralização e federalismo no Brasil. Rio de Janeiro: FGV/ Fiocruz.
• ARRETCHE, Marta. (2013), “Demos-constraining or demos-enabling federalism? political institutions and policy change in Brazil”. Journal of Politics in Latin America, vol. 5, n^o2, pp. 133-150.
• BEHNKE, Nathalie; MULLER, Sean. (2017), “The purpose of intergovernmental councils: a framework for analysis and comparison”. Regional & Federal Studies, 27, n^o5, pp. 507-527.
• BINDER, Sarah A.; LAWRENCE, Eric D.; SMITH, Steven S. (2002), “Tracking the Filibuster, 1917 to 1996”. American Politics Research, vol. 30, n^o4, pp. 406-422.
• BORGES, André. (2011), “Nacionalização partidária e estratégias eleitorais no presidencialismo de coalizão”. DADOS – Revista de Ciências Sociais, vol. 58, n^o3, pp. 651-688.
• BORGES, André; LLOYDS, Ryan. (2016), “Presidential coattails and electoral coordination in multilevel elections: comparative lessons from Brazil”. Electoral Studies, vol. 43, pp. 104-114.
• BORGES, André. (2019), “Federalism, party politics, and coalition dynamics”. In: B. Ames (ed.), Routledge Handbook of Brazilian Politics. New York and London: Routledge, pp. 175-201.
• BORZEL, Tanja. (2002), “Pace-setting, foot-dragging, and fence-sitting: member state responses to europeanization”. Journal of Common Market Studies, vol. 40, n^o2, pp. 193-214.
• BROSCHEK, Jörg; PETERSOHN, Bettina; TOUBEAU, Simon. (2018), “Territorial politics and institutional change: a comparative-historical analysis”. Publius, vol. 48, n^o1, pp. 1-25.
• CARREIRÃO, Yan. (2014), “O sistema partidário brasileiro: um debate com a literatura recente”. Revista Brasileira de Ciência Política, n^o14, pp. 255-295.
• CHEIBUB, José Antônio; FIGUEIREDO, Argelina; LIMONGI, Fernando. (2009), “Political parties and governors as determinant of legislative behavior in the Brazilian Chamber of Deputies, 1988-2006”. Latin American Politics and Society, vol. 51, n^o1, pp. 1-30.
• DESPOSATO, Scott. (2004), “The Impact of Federalism on National Political Parties in Brazil”. Legislative Studies Quarterly, no. 29, pp. 259-285.
• DICKOVICK, J. Tyler. (2007), “Municipalization as central government strategy: central-regional-local politics in Peru, Brazil, and South Africa”. Publius, vol. 37, n^o1, pp. 1-25.
• DO VALE, Helder F. (2013), “The judicialization of territorial politics in Brazil, Colombia and Spain”. Brazilian Political Science Review (BPSR), vol. 7, n^o2, pp. 88-113.
• DO VALE, Helder F. (2015a), “Comparing and measuring subnational autonomy across three continents”. LexLocalis, vol. 13, n^o3, pp. 741-764.
• DO VALE, Helder F. (2015b), “Territorial polarization in Brazil’s 2014 presidential elections”. Regional & Federal Studies, vol. 25, n^o3, pp. 297-311.
• FALLETI, Tulia. (2005), “A sequential theory of decentralization: Latin American cases in comparative perspective”. American Political Science Review, vol. 99, n^o3, pp. 327-346.
• FALLETI, Tulia; LYNCH, Julia F. (2009), “Context and causal mechanisms in political analysis”. Comparative Political Studies, n^o42, pp. 1143-1166.
• FALLETI, Tulia; LYNCH, Julia F. (2016), “Process tracing of extensive and intensive processes”. New Political Economy, vol. 21, n^o5, pp. 455-462.
• FALLETI, Tulia; LYNCH, Julia F; MAHONEY, James. (2015), “The comparative sequential method”. In: J. Mahoney; K. Thelen (eds.), Advances in comparative-historical analysis. Cambridge: Cambridge University Press, pp. 211-239.
• FENWICK, Tracy Beck. (2009), “Avoiding governors: the success of Bolsa Família”. Latin American Research Review, vol. 44, n^o1, pp. 102-131.
• FIGUEIREDO, Argelina; LIMONGI, Fernando. (2000), “Presidential power, legislative organization, and party behavior in the legislature”. Comparative Politics, vol. 32, n^o2, pp. 151-170.
• FIGUEIREDO, Argelina; LIMONGI, Fernando. (2007), “Government coalitions in Brazilian democracy”. Brazilian Political Science Review, vol. 1, n^o2, pp. 182-216.
• FILIPPOV, Mikhail; ORDESHOOK, Peter; SHVETSOVA, Olga. (2004), Designing federalism. A theory of self-sustainable federal institutions. Cambridge: Cambridge University Press.
• FIORETOS, Orfeo; FALLETI, Tulia; SHEINGATE, Adam. (2016), “Historical institutionalism in political science”. In: O. Fioretos; T. Falleti; A. Sheingate (eds.), The Oxford Handbook of Historical Institutionalism. New York: Oxford University Press.
• GRIN, Eduardo; ABRUCIO, Fernando. (2017), “O comitê de articulação federativa no governo Lula: os percalços da cooperação territorial”. Revista Brasileira de Ciências Sociais, vol. 33, n^o97, pp. 1-24.
• GRZYMALA-BUSSE, Anna. (2011), “Time will tell? Temporality and the analysis of causal mechanisms and processes”. Comparative Political Studies, vol. 44, n^o9, pp. 1267-1297.
• HIROI, Taeko. (2008), “The dynamics of lawmaking in a bicameral legislature: the case of Brazil”. Comparative Political Studies, vol. 41, n^o12, pp. 1583-1606.
• HIROI, Taeko; RENNO, Lucio. (2017), “Meaning of time: legislative duration in the Brazilian Congress”. IPEA Discussion Paper 228, Brasília, IPEA.
• HOWLETT, Michael; GOETZ, Kaus H. (2014), “Introduction: time, temporality and timescapes in administration and policy”. International Review of Administrative Sciences, vol. 80, n^o3, pp. 477-492.
• HOOGHE, L. et al. (2016), Measuring regional authority: a postfunctionalist theory of governance, vol. 1. Oxford: Oxford University Press.
• JONES, Mark; MAINWARING, Scott. (2003), “The nationalization of parties and party systems: an empirical measure and an application to the Americas”. Party Politics, vol. 9, n^o2,pp. 139-166.
• LIMONGI, Fernando. (2008), “O Poder Executivo na Constituição de 1988”. In: R. G. Oliven; M. Ridenti; G. M. Brandão (eds.), A Constituição de 1988 na Vida Brasileira. São Paulo: Aderaldo & Rothschild Editores and Anpocs, pp. 23-56.
• MAHONEY, James; RUESCHEMEYER, Dietrich. (2003), “Comparative historical analysis: achievements and agendas”. In: J. Mahoney; D. Rueschemeyer (eds.), Comparative historical analysis in the social sciences. Cambridge: Cambridge University Press.
• MAHONEY, James; RUESCHEMEYER, Dietrich; THELEN, Kathleene. (2010), “A theory of gradual institutional change”. In: J. Mahoney; K. Thelen (eds.), Explaining institutional change. New York: Cambridge University Press, pp. 1-37.
• MAINWARING, Scott. (1997), “Multipartism, robust federalism and presidentialism in Brazil”. In: S. Mainwaring; M. Shugart (eds.), Presidentialism and democracy in Latin America. Cambridge: Cambridge University Press, pp. 55-80.
• MEDEIROS, Antônio Carlos. (1986), Politics and intergovernmental relations in Brazil 1964-1982. New York and London: Garland Publishing.
• MELO, Carlos Ranulfo; CÂMARA, Rafael. (2012), “Estrutura da competição pela presidência e consolidação do sistema partidário no Brasil. DADOS – Revista de Ciências Sociais, vol. 55, n^o1, pp.71-117.
• MELO, Carlos Ranulfo; CÂMARA, Rafael. (2015), “The 2014 elections and the Brazilian Party System”. Brazilian Political Science Review, vol. 9, n^o1, pp. 93-114.
• MELO, Marcus. (2005), “O sucesso inesperado das reformas de segunda geração: federalismo, reformas constitucionais e política Social”. DADOS – Revista de Ciências Sociais, vol. 48, n^o4, pp. 845-889.
• MONTERO, Alfred. (2001), “After decentralization: patterns of intergovernmental conflict in Argentina, Brazil, Spain, and Mexico”. Publius, vol. 31, pp. 43-64.
• NEIVA, Pedro. (2011), “Disciplina partidária e apoio ao governo no bicameralismo brasileiro”. Revista de Sociologia e Política, vol. 19, n^o39, pp. 183-196.
• PALOTTI, Pedro L. de Moura; MACHADO, José Angelo. (2014), “Coordenação federativa e a ‘armadilhada decisão conjunta’: as comissões de articulação intergovernamental das políticas sociais no Brasil”. DADOS – Revista de Ciências Sociais, vol. 46, n^o4, pp. 399-441.
• PIERSON, Paul. (2016), “Power in Historical Institutionalism”. In: O. Fioretos; T. Falleti; A. Sheingate (eds.), The Oxford Handbook of Historical Institutionalism. New York: Oxford University Press.
• POWER, Timothy. (2010), “Optimism, pessimism, and coalitional presidentialism: debating the institutional design of Brazilian Democracy”. Bulletin of Latin American Research, vol. 29, n^o1, pp. 18-33.
• RIBEIRO, Pedro Floriano. (2013), “Organização e poder nos partidos brasileiros: uma análise dos estatutos”. Revista Brasileira de Ciência Política, vol. 10, pp. 225-265.
• RIBEIRO, Pedro Floriano. (2014), “Em nome da coesão: parlamentares e comissionados nas executivas nacionais dos partidos brasileiros”. Revista de Sociologia e Política, vol. 22, n^o52, pp. 121-158.
• RICH, Jessica A.; GÓMEZ, Eduardo J. (2012), “Centralizing decentralized governance in Brazil”. Publius, vol. 42, n^o4, pp. 636-661.
• RICCI, Paolo. (2003), “O conteúdo da produção legislativa brasileira: leis nacionais ou políticas paroquiais?”. DADOS – Revista de Ciências Sociais, vol. 46, n^o4, pp. 699-734.
• RIKER, William. (1964), Federalism: origin, Operation. Boston: Significance Little Brown).
• RODDEN, Jonathan A. (2006), Hamilton’s paradox: the promise and peril of fiscal federalism. Cambridge: Cambridge University Press.
• RODRIGUES-SILVEIRA, Rodrigo. (2011), “Descentralización y focalización del gasto social en los municipios brasileños”. Latin American Research Review, vol. 46, n^o3, pp. 60-92.
• SAMUELS, David; ABRUCIO, Fernando Luis. (2000), “Federalism and democratic transitions: the ‘new’ politics of the governors in Brazil”. Publius, vol. 30, n^o2, pp. 46-61.
• SAMUELS, David; ABRUCIO, Fernando Luis, David. (2003), Ambition, federalism, and legislative politics in Brazil. New York: Cambridge University Press.
• SELCHER, Wayne. (1989), “A new start toward a more decentralized federalism in Brazil?” Publius, vol. 19, n^o3, pp. 167-183.
• SOUZA, Celina. (1998), “Intermediação de interesses regionais no Brasil: o impacto do federalismo e da descentralização”. DADOS – Revista de Ciências Sociais, vol. 41, n^o3, pp. 569-592.
• SOUZA, Celina. (2001), “Federalismo e descentralização na Constituição de 1988: processo decisório, conflitos e alianças”. DADOS – Revista de Ciências Sociais, vol. 44, n^o3, pp. 513-560.
• SOUZA, Celina. (2015), “Breaking the boundary: pro-poor policies and electoral outcomes in Brazilian Sub-national Governments”. Regional & Federal Studies, vol. 25, n^o4, pp. 347-363.
• SOUZA, Celina. (2019), “Coordenação, uniformidade e autonomia na formulação de políticas públicas: experiências federativas no cenário internacional e nacional”. Cadernos de Saúde Pública, vol. 35, pp. 2-14.
• STEPAN, Alfred. (2000), “Brazil’s decentralized federalism: bringing government closer to the citizens?” Daedalus, vol. 129, n^o2, pp. 145-169.
• SUGIYAMA, Natasha Borges. (2011), “Bottom-up policy diffusion: national emulation of a conditional cash transfer program in Brazil”. Publius, vol. 43, n^o1, pp. 25-51.
• THELEN, Kathleen. (2003), “How institutions evolve: insights from comparative historical analysis”. In: J. Mahoney; D. Rueschemeyer (eds.), Comparative historical analysis in the social sciences. New York: Cambridge University Press, pp. 208-240.
• THORLAKSON, Lori. (2013), “Dynamics of change in federal representation”. In: A. Bezns; J. Broschek (eds.), Federal dynamics: continuity, change, and the varieties of federalism. Oxford: Oxford University Press, pp. 229-248.
• TRENCH, Alan. (2006), “Intergovernmental relations: in search of a theory”. In: S. L. Greer (eds.) Territory, democracy and justice. London: Macmillan, pp. 224-256.
• ZUCCO, Cesar; LAUDERDALE, Benjamin E. (2011), “Distinguishing between influences on brazilian legislative behavior”. Legislative Studies Quarterly, vol. XXXVI, n^o3, pp. 363-396.
• WAWRO, Gregory J.; SCHICKLER, Eric. (2004), “Where’s the pivot? Obstruction and lawmaking in the pre-cloture senate”. American Journal of Political Science, vol. 48, n^o4,pp. 758-774.

NOTES

• *
  Acknowledgement: the author would like to thank the anonymous reviewers for their contribution in providing insightful and detailed comments that were critical to improve the article’s argument. This work was supported by the Hankuk University of Foreign Studies Research Fund.
• 1
  . In the context of this research, decentralization is here understood as the process of distributing power downward across increasingly lower levels of government ( Falleti, 2005 ), while centralization is the process through which the federal government regulates and coordinates decentralized policies.
• 2
  . It is important to mention that although centralization entails limitation on the exercise of subnational autonomy, Brazilian subnational governments retained a high degree of subnational autonomy despite centralization of power and authority (Do Vale, 2015a; Souza, 2019 ).
• 3
  . De/centralization measures are laws that concern the decentralization or centralization of fiscal and administrative issues and were approved in the House of Representatives. These laws were proposed by either the national executive branch or a political party represented in the Brazilian House of Representatives.
• 4
  . Brazil stands out among federal democracies in that de/centralizing legislation addresses both levels of subnational government in the same law. In several other countries, including Spain, South Africa, Mexico, and India, de/centralizing legislation is created specifically in reference to either meso/regional governments or local governments.
• 5
  . Other possible dichotomies pertaining to territorial politics include federalism-unitarism and diversity-harmonization. However, these dichotomies are not as significant in Brazilian territorial politics as decentralization-centralization.
• 6
  . Parties within the same bloc do not necessarily act in a coordinated manner, since their grouping pertains only to their territorial policy agenda regarding the decentralization-centralization dichotomy.
• 7
  . Territorial players (national or subnational) use multiple strategies to establish vertical intergovernmental forums. Strategies include attempts to influence intergovernmental interactions, protect their own autonomy, coordinate policy, and exchange information (Behnke and Mueller, 2017: 514-515).
• 8
  . Out of these 32 legislative measures, three were approved as laws before the 1988 Constitution, 26 became laws after 1988, and three measures (PEC0169, PEC0175, PEC0222) failed to pass.
• 9
  . Based on these criteria, most legislative measures analyzed in this article were selected from a list of 58 decentralization laws that Arretche (2013) studied from 1989 to 2006.
• 10
  . One of the central propositions of historical institutionalism is that institutions have a long-lasting, constraining effect on the decisions of players, and, as a consequence, players themselves try to influence these institutions to advance their own preferences ( Thelen, 2003 ).
• 11
  . Falleti (2016:4-6) distinguishes between causal and consequential processes that contribute to an outcome. In a causal, or “extensive,” process, variables directly affect an outcome. In a consequential, or “intensive,” process, events linked in a sequence by causal mechanisms indirectly affect an outcome.
• 12
  . There is strong indication that territorial regional divide in Brazil has been politicized in times of elections (Do Vale, 2015b).
• 13
  . Speed is operationalized as the number of days necessary for a law to be approved in the House of Representatives. The approval process starts from the moment that the drafted law is first presented at a plenary session in the National Congress, and the process ends with final approval of the House of Representatives.
• 14
  . This law is considered a landmark case because, among other reasons, it was the first law intended to regulate finances at different levels of government after the 1988 Constitution, which represented an exceptionality in the sequence of decentralizing laws at the time ( Arretche 2012: 104). From its proposal in 1989 to its final approval in 1995, the text of the Camata Law experienced several changes. Rita Camata, the PMDB Congresswoman who proposed the law, faced opposition from her own party, and the law was approved only with the support of President Fernando Herinque Cardoso in 1995.
• 15
  . None of the de/centralization laws analyzed in this article were proposed or approved during Franco’s government.
• 16
  . The FSE was a temporary measure that was valid for a two-year period. Its purpose was to free up resources constitutionally earmarked for subnational governments and invest them in social services.
• 17
  . Since the mid-1990s, the federal government managed to overcome political deadlock and to rely on large government coalitions to approve legislations ( Figueiredo and Limongi, 2000 ). Specifically, Brazilian presidents overcame deadlock by issuing presidential decrees (Figueiredo and Limongi, 2006) and distributing cabinet positions (Amorim Neto, 2002).
• 18
  . Party systems in federal countries are considered essential for reducing territorial conflicts and fostering territorial cooperation (Filipov, Ordershook and Shvetsova, 2004) as they influence territorial politics through party representation across government levels and within their internal organization ( Riker, 1964 ).
• 19
  . Party nationalization causes an increase in policy interdependence. Parties become interlocked through a more congruent political agenda and a more integrated representation system. On the other hand, decentralized party systems have greater competition for agenda setting, since policy issues are more localized ( Thorlakson, 2013 ).
• 20
  . Indeed, party incongruence is attributable to two factors that raise the cost of cooperation: the autonomy of party branches at the state level and the fragmented party system (Borges and Lloyd, 2016).
• 21
  . Ideological elements in Brazilian territorial politics are minimized here because parties from different positions might share similar strategies in the recentralization process ( Carreirão, 2014 ). In effect, the ideological divide in the Brazilian National Congress does not always determine the voting behavior of legislators ( Zucco and Lauderdale, 2011 ).
• 22
  . Nonetheless, PT has been recognized for its establishment of clientelistic networks in places where traditional clientelistic parties used to have important influence ( Souza, 2015 ).
• 23
  . Even though the powers of the Executive Branch per se do not guarantee approval of its proposed legislation, the 1988 Constitution conferred important institutional mechanisms on the Brazilian executive to strengthen its powers vis-à-vis the legislative branch ( Limongi, 2008 ).
• 24
  . The creation of both CNE and Confaz took place during an authoritarian regime and were partially transformed in structure and function under the new democracy. While Confaz maintained its original name, CNE was formally known as the Federal Education Council (CFE).
• Datasets: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/HKSTMU

Publication Dates

• Publication in this collection
  14 Mar 2022
• Date of issue
  2022

History

• Received
  12 June 2019
• Reviewed
  21 July 2020
• Accepted
  8 Dec 2020