When the court is divided: minimum-winning coalitions in Brazil’s Supreme Court

Abstract

The aim of this article is to examine the decision-making behavior in Brazil’s Supreme Court (STF) judicial review cases, analyzing the dynamics of the collegial body in situations where "every vote counts", i.e., in cases that divided the court being decided by a margin of 1 or 2 votes. To do so, we conducted an exploratory study analyzing judicial review cases (ADIs) decided by the Supreme Court from 1988 to 2014, seeking to answer: i) how often and in which situations the court was divided in ADIs trials? ii) how compositions were formed, mapping the constitution and the fluidity of the minimum winning coalitions, and iii) how did its deliberative process flow? We answered to these descriptive questions, gathering empirical evidence to discuss the determinants of the minimum winning coalitions, to better understand the decision-making behavior of Supreme Court, dialoguing with arguments that understand the decision-making process of this court as personalistic, questioning its potential as a collegial body - which would pose concerns to democratic legitimacy of the institution. We conclude the Supreme Court was very consensual in the period analyzed, being divided into only 3% of all decisions. In terms of voting composition, we found much fluidity in coalitions, but even so we identified factors that make coalitions more predictable, like combination of the subject being questioned and the past career of Justices. We found strong evidence that Justices with career in the judiciary are more likely to vote together than to divide their votes. We also observed that the deliberative process in the court occurred with intense exchange of arguments, changes in vote direction and debates. The main contribution of this article is therefore the relativization of the personalism in decisions, presenting evidence of the centrality of the collegial game in the deliberative process and in the construction of decisions of the Brazil’s Supreme Court.

Keywords:
Supreme Court; Decision-making process; Minimum winning coalition

Resumo

O objetivo do artigo é discutir o comportamento decisório no Supremo Tribunal Federal (STF) no controle de constitucionalidade das leis, analisando a dinâmica de funcionamento do colegiado quando “cada voto conta”, ou seja, em casos decididos de forma apertada, por margem de 1 ou 2 votos. Realizamos, para isso, um estudo exploratório com base nas Ações Diretas de Inconstitucionalidade (ADIs) julgadas pelo colegiado do STF entre 1988-2014, buscando responder: i) com que frequência e em que situações o tribunal ficou dividido nos julgamentos de ADIs? ii) como os ministros se compuseram para votar nessas ações, mapeando a constituição e a fluidez das coalizões majoritárias mínimas, e iii) como se deu o processo deliberativo nesses casos? Respondemos a essas questões descritivas, reunindo elementos empíricos para discutir os determinantes das coalizões majoritárias mínimas, e melhor compreender o comportamento decisório do Supremo, no sentido de dialogar com argumentos que entendem o processo decisório dessa corte como personalista, questionando sua capacidade de deliberação colegiada – o que traria problemas de legitimidade democrática para a instituição. Concluímos que o Supremo foi bastante consensual no período analisado, ficando “dividido” em apenas 3% do total de decisões colegiadas. Em termos da composição de votação, houve bastante fluidez na corte, mas apesar dessa fluidez, identificamos fatores que tornam a constituição de coalizões mais previsíveis, como a combinação do tema em julgamento e a trajetória de carreira pregressa dos ministros, havendo indícios de que ministros oriundos da magistratura têm maior probabilidade de votar em conjunto do que dividir seus votos. Verificamos, ainda, que o processo deliberativo no tribunal se deu com intensa troca de argumentos, mudança de direção de votos e debates. A principal contribuição do artigo é, portanto, a relativização das teses do personalismo decisório, apresentando evidências da centralidade do colegiado no processo deliberativo e na construção das decisões do STF.

Palavras-chave:
Supremo Tribunal Federal; Processo decisório; Coalizões majoritárias mínimas

1. Introduction¹

The field of judicial behavior studies began to take shape in the United States in early decades of twentieth century as a descriptive enterprise, recognizing in judicial decisions the action of the political context in which they were taken, and these decisions were viewed not only as influenced by legal aspects, but also by personality, training, preferences, and values of judges (MAVEETY, 2003: 3).

It was only in the late 1940s, from the domain of behaviorist approach (“behavioral revolution”), that judicial behavior studies were consolidated, leaving researchers less concerned with the outcome of decisions and more with the dynamics of decision-making itself. At this point, taking Pritchett’s work (1948) as a landmark, a turning point took place in this field of studies, making the focus of discussion no longer the product of judicial decisions, that is, the meaning and content of the judges' decision, but in the decision-making process, that is, what makes the judges decide how they decide (GROSSMAN and TANENHAUS, 1969, apud MAVEETY, 2003: 11). Thus, the search for determinants of judicial behavior has become the central theoretical focus.

In Brazil, interest in judicial behavior earned academic relevance only in the 1990s, when pioneering work² about Supreme Court’s role in regulating the country’s economic, political, and social life began to emerge. And with this, the understanding of decision-making process in this court has become a topic of interest, occupying still small but growing space in the research agenda about Judiciary, not only in Law, but especially in Political Science and Sociology³.

Studies on STF decision-making process investigate factors that influence the decision of Justices in trial, proposing to map the determinants of judicial behavior and to understand court’s deliberative process and collegiate’ dynamics (see, as example, OLIVEIRA, 2012a and FERREIRA, 2013).

Recently, after a profound renewal of Supreme Court’s composition between FHC’s (1995-2002) and Lula’s (2003-2010) governments, researchers have also focused their attention to understand presidential appointment influence on how Justices organize themselves to vote (see, for example, OLIVEIRA, 2012b; FERREIRA and MULLER, 2014; DESPOSATO, INGRAM and LANNES, 2015 and ROSEVEAR, HARTMANN and ARGUELHES, 2015).

These works came together for an analysis on decision-making process especially in cases of constitutional control, highlighting judicial review cases (the Direct Actions of Unconstitutionality - ADIs). Researchers have used different time cutouts and criteria for case selection, some analyzing large volumes of cases with dissidence in the collegiate, and others based on cases considered complex, difficult or of great media repercussion.

In terms of theoretical and methodological approach, this work has been based on models developed to understand US Supreme Court’s decision-making process, adapted to Brazilian reality – especially the attitudinal and strategic models, influenced by neoinstitutionalism (following three dominant paradigms in American political science and sociology approach to judicial behavior⁴).

So far, the results achieved indicate that judicial decisions are, in fact, reducible to empirically observable concrete events. Therefore, STF research agenda has been increasingly seeking to translate the theoretical dimensions in judicial behavior discussion on tangible aspects.

Now we know, studying large volumes of cases decided by Supreme Court, that factors such as the origin of the law or norm being questioned (whether Federal or State, for example) and its thematic (tax, economic, social security, public servant, etc.) impact on decision-making process. And STF being more likely to reject federal legislation than state legislation, declaring the unconstitutionality of Federal legislation less often, and being more opposed to States occupying a larger space in the federation. We also know that STF has been more receptive to economic-tax issues and public administration, especially in the areas of public service, than to social rights issues (see, for example, OLIVEIRA, 2012a e FERREIRA, 2013).

We also know that the presidential appointment influences coalitions configuration in the Supreme Court, and different compositions, formed from presidential nomination blocs, result in different decision patterns (see, as example, OLIVEIRA, 2012b; DESPOSATO, INGRAM and LANNES, 2015 and ROSEVEAR, HARTMANN and ARGUELHES, 2015).

Other researchers have turned to more qualitative approaches, noting the decision-making process on issues politically costly, difficult cases, complex, or of great media impact. Kapiszewski (2011), for example, analyzed twenty-six cases, concluding that multiple political and institutional pressures influence judicial decisions, developing the tactical balance thesis. For this thesis, judicial behavior interpretation models (legal, attitudinal, strategic and neoinstitutional) are complementary, arguing that both legal and extralegal factors influence judicial decision making.

Kapiszewski states that when judging important cases, STF Justices tend to alternate the contestation of policies that interest the federal government with the endorsement of such policies, and when challenging or favoring the interests of federal government, Justices would have six considerations : (1) your own preferences; (2) institutional preferences; (2) federal government preferences; (4) economic potential or political consequences of the decision; (5) public opinion on the case and (6) legal aspects involved (2011: 472-473).

There are also researchers who discuss decision-making process from its normative and theoretical aspects, deducing from abstract reason the best decision-making model for Supreme Court’s collegiate, or discussing the adequacy of theoretical models developed in other national contexts to Brazilian case (see SILVA, 2009; 2013; RIBEIRO, ARGUELHES and PEIXOTO, 2009).

In this article, we enter this discussion considering a systematically unobserved aspect for any of these studies: Supreme Court’s decision-making behavior in cases of constitutional control in which "every vote counts"⁵. We analyze the totality of decisions about ADIs cases in the 1988-2014 period that divided the Court, that is, in which the margin of victory was given by only one or at most two votes⁶.

We deal directly with the Silva’s works (2009; 2013), considering the characteristics that the author highlights about deliberative process in STF, to analyze empirical elements and discuss what the author has constructed normatively.

We discussed mainly three aspects that, according to Silva (2013), decrease the deliberative Supreme Court’s quality, affecting consequently its democratic legitimacy: (i) the rapporteur’s irrelevance; (ii) the absence of a genuine exchange of ideas and arguments between Justices during trial, which according to Silva is evident since rarely a Justice mentions arguments presented by other Justices in his vote, turning court’s decision-making process purely aggregating, in which each Justice writes their own opinion and all opinions are published; and (iii) the possibility of interrupting plenary session, before each Justice had the opportunity to express their views on a particular case. For the author, this fact would be aggravated by the regimental disposition of each Justice voting in reverse order of seniority after the rapporteur, which would make it impossible or difficult for Justices to reconsider their votes, that is, to change their position after hearing the votes of their peers (SILVA, 2013: 569).

Besides dialogue with Silva’s work (2009; 2013), we rely on theoretical and methodological terms: Riggs’ (1993) exemplary study on decisions with a majority coalition in the U.S. Supreme Court.

In the following article, we present an overview of the works with which we dialogue, and analyze judicial review cases (ADIs) that divided the Supreme Court, decided by a minimum-winning coalition, with a margin of one or two votes, to answer the following descriptive questions: how often and in what situations STF was divided in ADIs trials? How were compositions of majority and minority blocs in those situations, i.e., who voted with whom? And how did the dynamics of deliberative process take place? To answer this last question, we mapped: (i) the frequency that a Justice changed its vote direction due another Justice vote; (ii) the frequency with which a Justice made explicit reference to an argument of another Justice; (iii) the frequency and justification of intervening recess throughout decisions, and (iv) the frequency with which rapporteur's vote was followed. With this, we seek dialogue with arguments of STF decision-making process being personalist, a sum of individual votes more than a collegial deliberation.

Answering to these descriptive questions, we have gathered empirical elements to discuss the determinants of minimum-winning coalitions, and to better understand STF decision-making behavior.

2. “When every vote counts”

Riggs (1993) examined voting pattern in U.S. Supreme Court considering cases decided by minimum-winning coalitions (with margin of a single vote), a period of 90 years, from 1900 to 1990. In author's definition, a majority coalition is minimal if, given the total number of judges who participated in decision, the change in the vote direction of a single judge would have ability to change the outcome of the decision.

The author identified by applying the criterion above explained, 1,428 cases that divided U.S. Supreme Court, which corresponds to about 11% of total cases decided by this court during these ninety years. Riggs's interest is to understand how minimum-winning coalitions are formed, and whether and how they deteriorate.

A first point to note with Riggs is that, since the Judiciary Act of 1925, US Supreme Court had extended its discretion in selecting cases that would decide, no longer needing to take many cases considered less relevant, seen as less controversial, given the opportunity, as well, to choose judging more difficult cases of wide public interest. Therefore, according to the author, in this context, it was expected that dissent in the court would be increased, and that divisions became fiercest. He notes that over time issues decided by one vote margin have increased: they remained around 3% until 1934, rising to 6% in 1935-1940 period, and always being above 10% after the 1940s, reaching expressive marks above 20% in the 1985-1989 period (RIGGS, 1993: 674).

The author argues that among the main explanatory factors of this trend change, one would be the ideological profile of judges, with special attention to the leadership profile of Chief Justice Harlan F. Stone (RIGGS, 1993: 682). Riggs points out that it's possible to note and ideological affinity in a majority coalition, stating that although there is variation due to type and subject of cases, and other imponderable factors, there is a notorious and constant variation due ideological alignment of Justices in terms of liberalism and conservatism in their decisions.

He analyzes the formation of coalitions based on two measures: the fluidity of coalitions, noting the number of different winning coalitions as a percentage of the total number of 5-4 decisions, with full fluidity (100%) occurring when the number of decisions and the number of different coalitions were the same; and the level of paired agreement or polarization of the Court, obtained from the relative frequency with which each judge votes with each of the others.

For pairs and blocks concordance, Riggs (1993: 705) uses as threshold values 70% for high concordance and 30% for low concordance. Existence of absolute fluidity at the court is possible to affirm if no pair reached any of these thresholds. In the analyzed period, the author finds polarization only once, in 1936, with two constant groups (judges Brandeis, Cardozo, Hughes, Roberts, and Stone with 94% agreement, and Butler, McReynolds, Sutherland, and Van Devanter with 100%), and absolute fluidity was never achieved.

In the analysis of alignments in STF, we use these same thresholds to identify fluidity or polarization of coalitions, but remembering that, unlike the US Supreme Court, STF has far less discretion about the cases it will hear, being obliged to position itself in many routine and low complexity cases⁷.

Sunstein (2015) also focused on analyzing voting patterns in US Supreme Court over time. Starting from Justice John Roberts’ statement, made in an interview given in 2006, in which, concerned about the legitimacy of the court, it states that unanimous or with minimal dissenting are difficult to reverse and contribute to the stability of the law, while highly divided decisions, like 5-4, bring the court closer to an institution where partisan politics predominates, Sustein argues that there is no empirical evidence to support the Court's decision-making pattern presents legitimacy problems for the institution.

Analyzing the behavior of collegiate over time, from the 19th century to 2012, it shows that the profile change of the court occurred after 1941, when it stopped operating under the "consensus rule" (in which cohesion remained above 80%), greatly increasing dissent, which now represents most of decisions.

According to the author, since 1941 dissent has been below 50% only in four periods (terms): 1996, 1997, 2005 and 2013, with the overall rate of dissent turning around 60%, and "5-4" decisions reaching almost 17% of the annual total. And he states that there is no evidence of a reversing on dissent trend over last decade or that divisions at court compromise its role in US government or its public legitimacy (2015: 815-816).

Sustein points out that two main factors linked to ideological profile of the judges help to understand this change in dissent trend and the increase of divisions in court: a change in the court leadership profile, with the rise of Harlan Fiske Stone to Chief Justice position, and the change in ideological composition of the court, with the appointment of seven new justices between 1937 and 1941.

In this article, we analyze a significantly shorter period of the Supreme Court than that observed for the US court by Sustein, and even by Riggs (1993), covering only 25 years, which include 34 Justices in 21 different compositions, and 15 presidencies of the court. If we consider the total number of ADIs collegiate decision in this period (1,419), 72% of them were unanimous⁸, and there does not appear to be a well-designed or robust tendency of composition change in majority decisions proportion. In only 2 of 21 compositions majority decisions proportion was above 25%.

Perhaps this is the reason that in the studies of Supreme Court's decision-making process the concern about incidence and size of dissent in court has received little attention. In these studies, one of the theses that prevails is personalism. Following this thesis, the court would function as eleven islands, or as eleven independent offices, with the sum of individual votes and little or no deliberation (see SILVA, 2009; MENDES, 2010).

The justifications for this thesis, which add up to more conjectures than empirical elements, point to several factors, such as the vanity of the Justices, intensified by the excessive publicity of the audiences, and the volume of work, among others.

Silva (2013) is among the authors who discuss this thesis, turning their concern for the quality of deliberation in Supreme Court. He states that external deliberation would predominate in court, including public opinion, government and political actors, and other legal practitioners, when internal deliberation, the one aimed at influencing the collegiate to decide on a regular course of an action, should be the central focus. Thus, it criticizes several practices and institutional rules of the court that would contribute to fuel individualistic attitudes of Justices, compromising court's legitimacy, raising the publication of divergent opinions, since the decision-making process would be purely aggregating, in which each Justice writes his own opinion and all opinions are published (seriatim) (2013: 579-580).

As already presented in introduction, Silva defends the thesis that there is little or no deliberation in Supreme Court and that it undermines the quality of decisions and may even affect court’s legitimacy. He uses as evidence for his thesis the irrelevance of the rapporteur, since other Justices would only know the content of his argument in the plenary session, so each of the Justices would have produced his vote before even knowing what decision would be proposed by the rapporteur, so that Justices cannot "just agree" with opinions they do not even know (SILVA, 2013: 570).

Another aspect that hamper a quality deliberation would be, in author's view, the extreme publicity with which decisions are made, being televised, and broadcasted live. This would make it unlikely that a Justice who has already delivered his vote would come back after hearing the votes of his colleagues, since they would have made a public commitment to a certain position (2013: 571). And one last point brought by Silva, which we consider important to recall, is criticism based on intervening recesses. According to the author, this possibility of interrupting plenary session before all Justices had expressed their opinion would also undermine the quality of deliberation.

Klafke and Pretzel (2014) question some of Silva's propositions, pointing to "nuances" in the constructing process of final decisions, based on the analysis of a sample of 266 STF judgments, from decisions on abstract constitutionality control between 2006 and 2010. According to the authors, we need to consider a judgment which,

... there are two non-excluding ways of configuring the document: (i) by appending the written votes released by the Offices; (ii) by attaching the audio transcript of trial. Therefore, final judgment may consist of votes released by the Office and / or audio transcripts of the judgment. (...) the vote may be revised to contain positions that emerged in plenary debates, including changes in the outcome. And, if there is a conflict between one and the other, audio recording prevails, as court has already had opportunity to decide (KLAFKE and PRETZEL, 2014: 94).

This aspect highlighted by Klafke and Pretzel (2014) is extremely relevant, as it is in the debate that we manage to apprehend aspects of Supreme Court deliberative process. Although Justices write their votes most often before plenary session, in some cases these votes aren’t even fully read and are only annexed to decision.

The authors focused on verifying argumentative dispersion caused by aggregator decision-making process. They recorded in these 266 judgments the number of votes published in each trend (winner or loser), concluding that 29% of judgments contain a maximum concentration of fundamentals, that is, only the rapporteur presents a written vote with foundation to the winning trend; in 39% they classify as submaximal concentration, that is, there is more than one vote in current winning, but less than half of Justices who compose it; and in 32% there is maximum dispersion, when all Justices of winning trend have published their votes (2014: 98). Based on these data, they point out that the metaphor of 11 islands, of extreme personalism, is relative.

Here, we follow this line of relativizing the thesis of personalism, bringing elements of decision-making process that contribute to highlight the importance of Collegiate and deliberation process in the construction of Supreme Court decisions. Our objective is not to refute Silva's thesis, but rather to delimit its scope.

3. Tight disputes in the Supreme Court

The database we analyzed here was built from searching STF website for ADIs with final decision between 1988 and July 2014. By final decision we consider cases that have had trial of merit or cases taken with prejudice, dismissed or extinct, or archived. Applying these criteria, we identified 2,712 cases. Of this amount, 1,419 were decided by collegiate - 52% of the total, the other 48% being monocratic decisions taken by the rapporteur, and this expressive number of monocratic decisions is one of the elements in the debate to criticize court’s individualization (SILVA, 2013; ARGUELHES and RIBEIRO, 2015). Thus, we don’t question the argument of individual concentration of power in Justices. But our interest is in verifying collegiate decisions, what is the deliberative process dynamics, and whether in these decisions there are debates, exchanges of ideas and arguments, and other elements that allow us to relativize personalism thesis.

3.1. Frequency and situations that divided STF plenary on ADIs

Among the 1,419 collegiate decisions, 403 were majority (28%), and only 48 of them divided the court (3%), but since 13 of them were judged together with others, we have 35 judgments that divided the court, 21 ADIs were decided by a single vote and 14 by two-vote margin. We analyze here 35 judgments, but one of them the court was divided into two distinct points with different compositions for each point, so we consider a total of 36 decisions with tight division.

Thus, we answered part of the first question, identifying that Supreme Court was rarely divided, being quite consensual in terms of concentrated constitutionality control in the analyzed period.

Dissension began to increase with the arrival of new Justices to court, especially since the appointments of Carlos Velloso and Marco Aurélio in 1990, in the fifth composition, which also corresponds to the increase of the activation of the Supreme Court by the legitimate ones, with several new cases arriving. The Court follows this trend until Fernando Henrique Cardoso’s nominations in the 10th and 11th compositions, which made the Tribunal more cohesive again. And with the court's major renewal in the Lula administration, dissent has reappeared more frequently. In 2004, dissension started to increase again (the 13th composition).

Considering only collegiate decisions, the compositions that presented the greatest dissent, being above 35% of decided cases, were 5a, 7a, 9a, 13a, 17a and 19a (see table 1).

Table 1
Percentage of majority decisions in accordance with STF composition

Period		Composition	Total				Collegiate
			Unanimous decisions	Majority decisions	Decisões monocráticas	(N)	Unanimous decisions	Majority decisions	(N)
1	10/03/87 13/04/89	Mayer, Falcão, Alves, Silveira, Correa, Passarinho, Rezek, Sanches, Gallotti, Madeira, Borja	75%	0%	25%	4	100%	0%	3
2	05/04/89 16/05/89	Mayer, Alves, Silveira, Correa, Passarinho, Rezek, Sanches, Gallotti, Madeira, Borja, Brossard	50%	0%	50%	2	100%	0%	1
3	17/05/89 16/08/89	Mayer, Alves, Silveira, Passarinho, Rezek, Sanches, Gallotti, Madeira, Borja, Brossard, Pertence	50%	0%	50%	6	100%	0%	3
4	17/08/89 12/06/90	Alves, Silveira, Passarinho, Rezek, Sanches, Gallotti, Madeira, Borja, Brossard, Pertence, Mello	59%	11%	30%	27	84%	16%	19
5	13/06/90 25/06/91	Alves, Silveira, Passarinho, Sanches, Gallotti, Borja, Brossard, Pertence, Mello, Velloso, Aurélio	43%	26%	31%	42	62%	38%	29
6	26/06/91 20/05/92	Alves, Silveira, Sanches, Gallotti, Borja, Brossard, Pertence, Mello, Velloso, Aurélio, Galvão	52%	21%	27%	67	71%	29%	49
7	21/05/92 14/12/94	Alves, Silveira, Sanches, Gallotti, Brossard, Pertence, Mello, Velloso, Aurélio, Galvão, Rezek	47%	30%	23%	91	61%	39%	70
8	15/12/94 14/04/97	Alves, Silveira, Sanches, Gallotti, Pertence, Mello, Velloso, Aurélio, Galvão, Rezek, Corrêa	68%	18%	13%	76	79%	21%	66
9	15/04/97 13/12/00	Alves, Silveira, Sanches, Gallotti, Pertence, Mello, Velloso, Aurélio, Galvão, Corrêa, Jobim	41%	22%	38%	282	65%	35%	176
10	14/12/00 19/06/02	Alves, Silveira, Sanches, Pertence, Mello, Velloso, Aurélio, Galvão, Corrêa, Jobim, Gracie	26%	5%	69%	341	83%	17%	106
11	20/06/02 24/06/03	Alves, Sanches, Pertence, Mello, Velloso, Aurélio, Galvão, Corrêa, Jobim, Gracie, Mendes	49%	6%	44%	311	88%	12%	173
12	25/06/03 29/06/04	Pertence, Mello, Velloso, Aurélio, Corrêa, Jobim, Gracie, Mendes, Peluso, Britto, Barbosa	39%	9%	52%	198	81%	19%	95
13	30/06/04 15/03/06	Pertence, Mello, Velloso, Aurélio, Jobim, Gracie, Mendes, Peluso, Britto, Barbosa , Grau	25%	15%	61%	354	62%	38%	140
14	16/03/06 20/06/06	Pertence, Mello, Aurélio, Jobim, Gracie, Mendes, Peluso, Britto, Barbosa, Grau, Lewandowski	39%	14%	47%	49	73%	27%	26
15	21/06/06 04/09/07	Pertence, Mello, Aurélio, Gracie, Mendes, Peluso, Britto, Barbosa, Grau, Lewandowski, Lúcia	47%	22%	31%	116	69%	31%	80
16	05/09/07 22/10/09	Mello, Aurélio, Gracie, Mendes, Peluso, Britto, Barbosa, Grau, Lewandowski, Lúcia, Direito	40%	16%	44%	366	72%	28%	205
17	23/10/09 02/03/11	Mello, Aurélio, Gracie, Mendes, Peluso, Britto, Barbosa, Grau, Lewandowski, Lúcia, Toffoli	28%	24%	48%	148	53%	47%	77
18	03/03/11 18/12/11	Mello, Aurélio, Gracie, Mendes, Peluso, Britto, Barbosa, Lewandowski, Lúcia, Toffoli, Fux	41%	18%	40%	99	70%	31%	59
19	19/12/11 28/11/12	Mello, Aurélio, Mendes, Peluso, Britto, Barbosa, Lewandowski, Lúcia, Toffoli, Fux, Weber	7%	22%	71%	69	25%	75%	20
20	29/11/12 25/06/13	Mello, Aurélio, Mendes, Britto, Barbosa, Lewandowski, Lúcia, Toffoli, Fux, Weber, Zavascki	0%	14%	86%	7	0%	100%	1
21	26/06/13 30/07/14	Mello, Aurélio, Mendes, Barbosa, Lewandowski, Lúcia, Toffoli, Fux, Weber, Zavascki, Barroso	30%	7%	63%	57	81%	19%	21
Total			38%	15%	48%	2.712	72%	28%	1.419

Considering all dissent in court, of 403 cases with majority decision, 47% of them (n = 189) had only one Justice in dissidence, being Marco Aurélio isolated minority in 76% of these decisions with single dissidence (n = 143). Justice Britto was the second most frequent isolated dissident, being alone in 5% of these decisions and Justice Pertence, the third, being isolated dissident in 4% of majority decisions with a single minority vote.

With this behavior Justice Marco Aurélio gained the reputation of intentional dissident, "defeated vote", leading many researchers, even, to exclude him from his models of understanding of judicial behavior (see DESPOSATO et al, 2015).

Here, we do not follow this orientation because we believe that we learn more about judicial behavior by keeping Justice Marco Aurélio in the analysis, since he is a constitutive part of the Court⁹.

But the isolated dissent of Justice Marco Aurélio is noteworthy and seems to have a standard in his ratio decidendi - the Justice tends to consider as legitimate, associations or confederations that present for him potential of having national representation, and even in some cases recognizes legitimacy to "associations of associations”¹⁰.

Sometimes he tends to disagree with terminology employed in the decisions, stating that one must consider the lack of demand and not the illegitimacy¹¹. At other times, his dissent is motivated by understanding that Supreme Court should admit cases that other Justices think is not up to the court decide¹². To investigate more systematically Justice Marco Aurélio’s dissident behavior is therefore necessary, to verify a plausible hypothesis that he would tend to recognize and seek an extension of the Supreme Court's authority.

Considering all ADI judgments with majority decisions that he participated, Justice Marco Aurélio voted with majority only in 24%, and was an isolated minority in 37%. Francisco Rezek was isolated minority in two judgments with majority decision that he participated.

Another important factor to understand decision-making behavior is judicial fluidity - that is, how often Justices have positioned themselves with majority or minority (see table 2).

Table 2
Frequency with which Justices voted with majority

Justices *	Total majority ADIs they voted for	% of times that they voted with majority	% of times that they voted isolated in minority	Total majority ADI in which they voted (minority >1)	% of times that they voted with majority (minority >1)	Total ADI with tight division in which they voted	% of times that they voted for the minimum winning coalition
Weber	18	100%		8	100%	1	100%
Zavascki	5	100%		2	100%	1	100%
Alves	150	94%		89	90%	17	82%
Sanches	153	93%		88	88%	16	81%
Gracie	198	92%		109	85%	18	78%
Corrêa	128	90%	2%	73	85%	16	75%
Gallotti	125	86%	2%	79	81%	16	75%
Mello	310	88%		169	78%	26	69%
Jobim	150	94%		84	89%	16	69%
Grau	145	87%	1%	73	77%	14	64%
Lúcia	143	92%		60	82%	11	64%
Lewandowski	158	94%		71	87%	13	62%
Galvão	134	87%	1%	79	80%	16	56%
Britto	210	74%	4%	109	58%	16	56%
Peluso	221	89%	1%	110	81%	18	56%
Passarinho	10	80%		9	78%	2	50%
Brossard	48	81%	2%	32	75%	6	50%
Barbosa	195	80%	3%	100	66%	18	50%
Velloso	205	77%	2%	123	65%	22	50%
Pertence	274	69%	3%	155	50%	27	41%
Mendes	233	91%	0%	115	82%	15	40%
Direito	37	81%		19	63%	5	40%
Aurélio	384	24%	37%	202	28%	35	34%
Rezek***	29	69%		20	65%	3	33%
Borja	27	52%	7%	18	39%	4	25%
Silveira	135	73%	1%	81	57%	17	24%
Barroso	4	100%		1	100%	0
Fux	33	91%	3%	14	86%	0
Toffoli	57	91%	2%	22	82%	0
Rezek**	2	100%		0		0
Madeira	3	100%		0		0

• ^*
  Justices Rafael Mayer, Djaci Alves Falcão e Oscar Dias Correa did not vote in any of the majoritarian decisions in judicial review cases in the analysed period, therefore the analysis take into consideration the amount of 31 Justices
• ^**
  First appointment period: 3/24/83 to 3/15/90
• ^***
  Second appointment period: 5/21/92 to 5/2/97

We note that Justices Barbosa, Velloso, Rezek, Direito, Britto, Silveira, Pertence and Borja are the ones that present more fluid behavior (voting with majority in a percentage between 30% and 69% of the time), being Marco Aurélio quite consistent in minority position (voting in minority 76% of the time).

Justices Alves, Passarinho, Gallotti, Sanches, Brossard, Galvão, Mello, Corrêa, Jobim, Gracie, Mendes, Peluso, Grau, Lewandowski, Toffoli, Lucia, Fux, Weber, Zavascki and Barroso, consistently voted with majority in 70% of times or more.

But when considering only tight divisions, we realize that these positions changes. Voting consistently with majority the Justices Alves, Sanches, Gracie, Correa and Gallotti. And Justices Mello, Jobim, Grau, Lúcia, Lewandowski, Galvão, Britto, Peluso, Passarinho, Brossard, Barbosa, Velloso, Pertence, Mendes, Direito, Rezek and Aurélio stand with greater fluidity between majority and minority. Only Borja and Silveira have voted consistently with minority. Justices Weber and Zavascki only participated in a tight dispute with majority.

This data would be sufficient evidence to weigh the outvoted label that Justice Marco Aurélio carries, but when "difficult" cases are at stake, he tends to present a more fluid behavior between majority and minority.

The second part of the question concerns situations that most divided Supreme Court. In this point, two aspects are relevant: origin of legislations questioned and its subject matter. We noticed that federal legislation were the ones that provoked the greatest divisions in court. Although they corresponded to only 29% of total legislation questioned in concentrated control (ADI), federal regulations were responsible for 47% of tight divisions in Supreme Court.

Concerning this theme, most of them are about public agents career (40%), especially judicial career organization (19%) and judicial prerogatives (11%), followed by civil society regulation (14%), involving social rights (8%) and civil rights (6%).

Figure 1
Origin of legislation questioned in cases, according to the result of the decision

Federal regulations relate mainly to public officials and civil society (29% of federal regulations each), followed by norms that seek to regulate other subjects of public administration (23%), mainly institutional organization.

An example of a case that divided the court in the subject of federal civil service is ADI 3,772, requested by the Attorney General in 2006, questioning Law 11,301/06 definition of teaching functions for special teacher retirement, allowing to compute as service period the time spent outside classroom, like administration activities. Court ruled partially in favor, with an interpretation that exclude special retirement only to experts in education, the majority was made up of Justices Lewandowski, Peluso, Aurelio, Mello and Grau, and the minority by Justices Britto (rapporteur), Carmen Lúcia and Barbosa, who had judged the case fully upheld, and Gracie, whom considered to be unfounded.

Figure 2
Subject of the legislation questioned in the case, according to the result of the decision

Regarding ADIs that question civil society issues, we have, for example, ADI 1,755, in which the Liberal Party questioned the sole paragraph of Article 1 of Federal Law 9,294 / 96, with restrictions on use and advertising of alcoholic beverages and cigarettes, banning advertising of drinks when its alcohol content exceeds 13 degrees on Gay Lussac scale. Plaintiff argued that the law could not leave out drinks with alcohol content less than thirteen degrees.

The court didn't accept the case, in the terms of rapporteur Justice Nelson Jobim's vote, considering that STF would be called to act as "a true positive legislator", the majority were with Justice Jobim (Justices Alves, Pertence, Galvão and Corrêa), defeating Justices Aurélio, Silveira and Velloso who accept the case. During this trial debate, it's possible to perceive the dispute of positions between minority and majority trends, in a sense of extending or restricting Court’s authority, that is, whether Supreme Court could decide this question. And we noticed that three of the four magistrates present to the decision were positioned for expanding STF competence, being defeated.

MIN. VELLOSO (President): Mr. Justices, I don’t give the slogan "The Judiciary, in constitutional control, is, simply, negative legislator", the relevance that it seems to have, but which doesn’t (...). This question was given to the Supreme Court two hundred years ago, and the Court faced it. MIN. ALVES: Mr. President, in concentrated control, we cannot extend what is alleged to have hurt isonomy. We can withdraw what has been given, but we cannot extend to others what has not been given to it, when it should have been. MIN. VELLOSO (President): in Brown x Board of Education, 1954, US Supreme Court acted as a positive legislator. This is a case I can mention, at once, but there are others. Supreme Court thus proceeded, recently, in the “public servants’ 28%” case. It didn't take this “slogan” to orthodoxy as some want (Min. Carlos Velloso and Min. Moreira Alves, ADI 1,755, pp. 26-27).

Other examples on civil society subject that divided the court are ADI 3,112¹³, in which Federal Law No. 10,826/03 - the Disarmament Statute - is questioned, and ADI 3,510, in which Law 11,105/05, about stem cell research, which we will discuss later.

About public administration, we have as example ADI 3,290, in which PSDB questions Provisional Measure No. 207, dated August 13, 2004, that equates the special post of president of the Central Bank to the position of Minister of State. Most of Justices, composed of Jobim, Gracie, Mendes, Peluso, Barbosa and Grau, dismissed the case, defeating Justices Britto, Velloso, Aurelio and Pertence that considered it appropriate, and in part Justice Mello, who considered partially founded.

At state level, most of diplomas refer to public agents’ subject: 60% of state legislations that divided the court deal with the civil service, being largely related to the careers of justice, such as ADI 139, in which the Association of Brazilian Magistrates questions Article 82 of the transitional provisions of the Constitution of the State of Rio de Janeiro, ensuring that judicial and extrajudicial services holders have the right to retire with 60% of judges' salaries. In this case, the majority, made up of Justices Alves, Sanches, Gallotti, Mello, Velloso and Aurélio, declared Article 82 unconstitutional, defeating Justices Pertence, Passarinho, Brossard, Borja and Silveira.

Only 13% of state-level legislation that divided the court concern civil society issues, such as ADI 386, proposed by three industry associations (ABRASSUCOS, ABECITRUS and ANIC), questioning Article 190 of the State Constitution and Article 41 of the transitional provisions of the Constitution of the State of São Paulo, about safety in rural and urban workers transport. Most of the court, Justices Sanches, Alves, Silveira, Gallotti, Passarinho and Brossard, didn't accept the case, due to the active illegitimacy of its proponents, defeating Justices Velloso, Aurélio, Mello, Pertence and Borja, who receive part of it.

We note, therefore, that some of tight divisions in court were due to whether recognize the legitimacy of applicants (8% of total), and that on all occasions Justice Aurelio remained in minority, recognizing greater legitimacy to these applicants. For example, in ADI 23, proposed by ADEPOL (Brazilian Association of Police Chief), arguing the unconstitutionality due Governor of São Paulo's omission of article 241 of the Federal Constitution, which would equate salaries of police chief careers and other legal careers, due the resemblance of their functions. Majority coalition didn’t recognize ADEPOL's legitimacy as it is an association of associations, defeating Justices Marco Aurélio, Galvão, Pertence and Silveira, who recognized its legitimacy.

Justice Marco Aurélio also tends to position himself by extending STF competence when receiving incomplete or "deficient" petitions, arguing that STF can accept such requests (examples are ADIs 2,174 and 2,980) - what reinforces the need to analyze in detail his dissident behavior, and not simply to exclude him from analysis models.

3.2. Minimum-winning coalitions

The second question that we are interested in answering concerns the formation of majority blocs, or minimum-winning coalitions, and its counterpart, the losers’ blocs. We use the degree of cohesion of Justices in pairs, working with the cut of 70% agreement among Justices for high cohesion, and below 30% agreement for low cohesion, following the same parameters adopted by Riggs (1993).

The first point that stands out is that since 1988, in all ADI trials where the court was divided, only two "cliques"¹⁴ with more than two Justices were identified, one composed of Justices Alves, Sanches and Corrêa, and another by Justices Alves, Corrêa and Jobim, who maintained a high cohesion on the same side in more than 80% of times they voted together.

Justices Alves, Sanches and Correa have in common being perceived as having a more technical behavior, coming from different presidential appointments, and with different careers paths. On the other hand, both Corrêa and Jobim occupied the position of Minister of Justice before being Justices of the Supreme Court.

There are several cases of high cohesion pairs, such as Justices Alves and Brossard, Gallotti and Galvão, and Gallotti and Jobim, Borja and Brossard and Borja and Pertence, Mello and Velloso, Velloso and Britto, Peluso and Law, Grau and Lewandowski, Lúcia And Law, among others - high cohesion pairs are highlighted in yellow in table three.

Noteworthy is that the Justice Marco Aurélio showed high cohesion with only one Justice, Carlos Velloso (71% agreement). What these two Justices have in common is the appointment given by President Fernando Collor de Mello and the fact that they come from a career in magistracy¹⁵. On the other hand, Marco Aurélio systematically disagreed with Alves, Gallotti and Sanches, the latter also coming from the magistracy - pairs of disagreement are highlighted in red in Table three.

Despite the great fluidity in minimum-winning coalitions formation, we have noticed some situations that place Justices more frequently in the same bloc. There are indications that, in general, Justices with a previous career in magistracy tend more often to vote in same direction than to divide their votes to contribute to greater court cohesion - these evidences were obtained from a cohesion indicator, close to cohesion measures used in parliamentary decision-making studies, as the division of the difference between the total of Justices in majority and the total of Justices in minority, by the total of Justices voting in case.

The cohesion indicator varies from 0 to 1, although here it never reaches extremes, as it would be equal to 1 if there was unanimity (all Justices voted together) and equal to 0 if Justices were equally divided between minority and majority. Therefore, the closer to zero, the lower the cohesion, and the closer to 1, the greater the cohesion - this indicator was constructed based on Rice (1928).

The average cohesion in these 36 decisions was 14% and the minimum 9% - since we are dealing with cases that most divided the court. Correlating this cohesion indicator with the proportion of Justices with a previous career in magistracy voting in each decision, we have a positive coefficient of 0.31, in a 90% confidence interval.

It’s noted that Justices with a previous career in magistracy tended to vote more frequently together on demands involving the Association of Brazilian Magistrates (AMB). An example is the ADI 139, in which the AMB questions an article of the transitional provisions of the Constitution of the State of Rio de Janeiro that equates the salaries of the judicial and extrajudicial services holders to judges’ salaries. STF upheld the case, leaving three of the five judges present in the decision in majority block: Sanches, Aurélio and Velloso, and in minority block were Silveira and Passarinho.

Another example is ADI 3,362, in which all career magistrates voted in the interests of AMB. It questioned an article of the Constitution of the State of Bahia that established a maximum limit of 35 appellate judges in the TJBA. Justices Aurelio, Gracie, Peluso and Lewandowski, all four Justices with a previous career in magistracy, voted for the case, arguing an initiative vice, in violation of courts of justice’s authority to propose to the State Legislature to change the number of lower courts.

Table 3
Matrix of similarity indices, corresponding to the combinations of Justices 2 x 2, in cases with a tight division (in%)

	Alves	Silveira	Passarinho	Sanches	Gallotti	Borja	Brossard	Pertence	Mello	Rezek	Velloso	Aurélio	Galvão	Corrêa	Jobim	Gracie	Mendes	Peluso	Britto	Barbosa	Grau	Lewand.	Lúcia	Direito	Weber
Alves		18	50	86	56	50	83	31	55	33	38	29	40	82	88
Silveira			50	36	50	50	50	50	36	67	44	53	53	27	0
Passarinho				50	50	0	50	50	50		50	50
Sanches					57	25	67	36	58	33	29	29	39	89	67	0
Gallotti						50	67	40	46	67	40	13	71	40	71
Borja							75	75	25		25	50	0
Brossard								50	0	0	0	33	25
Pertence									58	33	45	52	47	36	36	46	50	36	67	36	33	29	60
Mello										67	75	48	40	67	50	50	27	43	50	50	27	50	38	40	100
Rezek											33	67	33	0
Velloso												71	47	50	36	17	50	20	80	20	0
Aurélio													40	33	27	24	60	56	56	39	21	39	18	60	100
Galvão														33	56	100
Corrêa															73	60	50	0	50	25
Jobim																50	50	57	14	100	50	25
Gracie																	43	53	44	65	69	50	80	40
Mendes																		67	31	47	33	20	22	100	0
Peluso																			25	50	57	54	27	80	0
Britto																				44	42	58	60	0
Barbosa																					29	46	36	0	100
Grau																						70	90	75
Lewand.																							50	25	100
Lúcia																								50	0
Direito
Weber

In ADI 314, proposed by the Attorney General, questioning paragraph 2 of article 58 of the Constitution of the State of Pernambuco, which attributed to the state governor the power to appoint appellate judges, through the promotion of a career judge, also allows us to visualize the same tendency of magistrates to vote together, to defend interests of judiciary.

Most of the court, composed of Justices Sanches, Silveira, Velloso, Galvão, Mello and Gallotti, decided for the unconstitutionality of the article, since it would violate Judiciary prerogative in appellate judges’ appointment, damaging their autonomy. We note that four of the five magistrates in this composition vote together, ensuring Judiciary independence. Justice Marcus Aurelio, the fifth magistrate, was in minority, along with Justices Brossard, Pertence, Borja and Alves, considering the constitutional norm.

The rapporteur, Justice Carlos Velloso, brought the principle enshrined in article 96, paragraph I, line "c", of the Federal Constitution, establishing the exclusive authority of the courts for appointing appellate judges. He quotes himself in a speech given on 05/29/1985, in which he indicated the need to reform the Judiciary, criticizing the form of appointment and promotion of state judges,

In some Member States, judges’ appointment or promotion depends on indication and support of local political leaders - mayors, councilors, deputies. This requires the judge to seek indications of these politicians, which, of course, significantly reduces his independence, resulting in damage to courts. (Min. Carlos Velloso, ADI 314, pp. 13-14).

Velloso finished his vote stating that "in good time the 1988 Constitution eliminated the illness and gave more independence to the Judiciary, for claimants’ benefit” (pp. 14), concluding on the unconstitutionality of the provision, given the danger that this form of appointment represents to the autonomy achieved by Judiciary since the 1988 Constitution.

Justice Paulo Brossard countered Velloso argument, stating that the fact that the Governor appointed a judge would not remove his independence, and that any process of choice has merits and limitations, thus voting for the constitutionality of the questioned rule, he brought his experience as a politician, in the state government of Rio Grande do Sul, as an additional element that formed his conviction, in a strong indication of how previous experience of STF influences the way Justices vote.

Courts are composed of men, and these, with or without a black cloak, look and sometimes are not immune to feelings that are not the best. Neither are courts exempted from the formation of groups within them, which, once constituted, are above any correction or sanction. (...) I was also Secretary of the Interior and Justice in my State, for a short time, that's right, but enough to be able to say that, while I was a secretary, I never received a request for a judge to be promoted. And in the gaucho magistracy I had classmates, my friends, therefore. Given this testimony, I would like to add that the fact that a judge is engaged in his promotion through petitions does not mean that he loses his independence (Min Paulo Brossard, ADI 314, pp. 22-23).

Even when they run counter to AMB's interests (which is rare), Justices with a career in magistracy tend to hold more in same bloc. An example is ADI 1,289, requested by AMB against a decision of the Superior Council of Labour Public Prosecutor, in which the council established criteria for the composition of lists with candidates to fill judges’ positions in Regional Labor Courts, via “Quinto Constitucional” (or "Constitutional Fifth"). Four of the five career magistrates present at the session voted against the case: Velloso, Aurelio, Silveira and Galvão. Justice Sanches was in majority coalition, judging the case to be appropriate.

During debate, when countering the argument of Justice Sepúlveda Pertence, Justice Marco Aurélio exposed the argument that oriented minority:

Your Excellency does not honor the precedent established in the ADI 581 judgment. What is more interesting is that it was a judgment that also dealt with a normative act linked to Labor Justice. We gave this interpretation about the career positions; now, regard to the Fifth positions, we will lend another approach to the Federal Constitution. Where is the harmony? Where is the celebrated, well-defined preservation of the jurisprudence of the Court? What is the importance of this unanimous decision? We had no divergent votes. (...) And now we do not want to acknowledge the paternity of this son; We are saying that he is an ugly son and that therefore we are not the parents. (Min. Marco Aurélio, ADI 1.289, pp. 65-66).

Here we see, beside the formation of a minority around Justices with previous career in magistracy, the discourse towards valorizing Court's precedents and jurisprudence. Another aspect that interests us in this ADI is the declaration of a change of understanding made by Justice Velloso. When voting on merits, the Justice said that he changed the way he voted in the preliminary trial, believing that he didn’t have "the best interpretation of constitutional text", reconsidering his position after hearing other Justices' votes. This aspect of the change in voting orientation will be explored further.

Continuing on the impact that the previous career has on the way Justices position and group to vote, we can observe a greater division between those who have worked in the Public Prosecutor's Office and those who came from advocacy, when what is on trial matters to the respective professional bodies.

An example of this influence is the debate between Justices Sepúlveda Pertence and Ayres Britto in the "ADI 3,028" trial, in which the Attorney General questions article 28 of complementary law 166/99 of the State of Rio Grande do Norte, which establishes additional revenue collection on all state extrajudicial procedures to compose the state Public Prosecutor's reequipping fund.

MIN. BRITTO: Madam President, as I have already been rapporteur for some of the ADIs, including in favor of the Office of the Public Defender of Rio de Janeiro, I will accompany the eminent rapporteur, asking, however, not to endorse the fundamentals of His Excellency's decision. MIN. PERTENCE: Was the Public Defender Fund declared constitutional?? MIN. BRITTO: It was. MIN. PERTENCE: Why, however, do you deem it unconstitutional, following the rapporteur, in this case? MIN. BRITTO: I am following the Rapporteur, but not his fundamentals, because I fear that… MIN. LÚCIA: In this case, it seems that there are other characteristics. MIN. BRITTO: That's right, they are not compatible with the fundamentals we approve in ADI 3,643. (…) MIN. PERTENCE: Yes, that is why I asked why Your Excellency, despite voting for the constitutionality of the Public Defender Fund, agrees with the eminent Rapporteur. MIN. BRITTO: Because - I quote from memory and I do not know if you remember - we understand that the Public Defender's Office was a necessary instrument for the democratization of access to justice. This was the line, that we followed. MIN. PERTENCE: Is the Public Prosecutor's Office useless? MIN. BRITTO: Absolutely. (...) I’d like to do the following, Your Excellency: I want to request an intervening recess so I would not contradict myself. We started with the consideration that the Public Defender's Office was an indispensable mechanism for the democratization of access to the justice. MIN. AURÉLIO (rapporteur): the most interesting is the antagonism: Public Prosecution X Public Prosecutor - but it reveals that this is a bit extravagant: it involves a "plus" and charges for a service that has nothing to do with this service. (Min. Ayres Britto, Min. Sepúlveda Pertence, Min. Marco Aurélio and Min. Carmen Lúcia, ADI 3,028, pp. 08-10).

When voting after the intervening recess, Justice Britto reconsidered his position, inaugurating the dissidence that became winner, considering constitutional the creation of the fund, dismissing the case.

Another decision in which we perceive the Justices bringing their professional experience to justify their views is the ADI 1.194, in which the National Confederation of Industry questions articles of the Federal Law nº 8,906 / 94, disposing on the Statute of Advocacy and the Order of Attorneys of Brazil (OAB), the imposition of compulsory legal counsel in the execution of a contract, and the payment of legal fee.

During the debate, Justice Pertence brought his experience as a lawyer to substantiate his position for partially dismissing the case in relation to article 21 of the law (saying that in causes which the employer is part, the attorney's fee would be paid to the employed lawyers). Justice Pertence argued that by his experience as lawyer it’s possible, if both parties wish, to include another stipulation, as it is a right they own. Justice Aurelio countered this argument, claiming to have also experience in the subject, given his experience in Labor Justice. Justice Britto was another to use his years of advocacy to support his position.

MIN. PERTENCE: Mr. President, my vote coincides with Your Excellency's, only to contradict learned opinions, with a little more experience of advocacy, knowing that it will be a statistical rarity, for example, in labor law, to the poor worker receive these fees. MIN. AURÉLIO: Excellency, in this case I can speak using the experience I have in Labor Court. MIN. PERTENCE: Your Excellency has an experience in the Labor Court always at the presidential seat. MIN. AURÉLIO: Just as Your Excellency has brought earlier, including overlapping others in general advocacy. I'm following you, Your Excellency. MIN. PERTENCE: It was only a lack of employment, Your Excellency. The one I had was taken. MIN. AURÉLIO: The experience is privilege of Justice Sepúlveda. MIN. PERTENCE: Justice Marco Aurelio, I was also a Public Prosecutor, but instead of being promoted, they took my job. That's why I had so much experience in advocacy. MIN. AURÉLIO: Do not hold resentment, Your Excellency. MIN. BRITTO: Mr. President, I was a lawyer for 33 years, mainly of collective cases. Very rarely have I made a fee contract to receive something in advance. Almost every time we put an ad exitum contractual clause. MIN. AURÉLIO: Your Excellency, as a lawyer, was a unique Christian. (Min. Sepúlveda Pertence, Min. Marco Aurélio and Min. Ayres Britto, ADI 1,194, pp. 27-29).

Whether experience in political office, experience in magistracy, or experience in advocacy, what these votes illustrate is that Justices’ previous career before occupying the post in Supreme Court permeates the construction of their positions, as well as influence with whom they group to vote.

3.3. Deliberative process in tight divisions

The third question that we want to answer is how was the deliberative process in cases decided by minimum-winning coalitions. We're interested in observing frequency of a Justice's vote changing its direction, the frequency with which a Justice made explicit reference to another Justice's argument, the frequency and justification of intervening recesses, and the influence that rapporteur's vote has in these decisions, noting the incidence of the Justice s' adherence to his arguments.

The first point to note is that in 28 percent of times that the Court was divided into ADIs trials, the majority decided by dismissing the case. As already pointed out, Justice Marco Aurelio tended to remain in the minority coalition in almost these decisions, either because he recognized a larger scope of acting for Supreme Court or because he considered associations legitimate against most of the court. In 20% of the time, the Supreme Court held to be unfounded, other 20% to be founded and in 22%, to be partially founded.

The second point is that for 32 of 35 ADIs here analyzed is available a transcript of debates, and it’s possible to analyze the process of interaction between Justices during trial. And in these debates sometimes appears reference of a Justice who brought written vote, but only to attached it or read excerpts, and not to read everything during the session. An example of this is Justice Gilmar Mendes in ADI 3,112, stating that he "brought long considerations - I will not read them, you don’t have to be worried - about this issue of criminalization mandates" (pp. 124). And Justice Ellen Gracie, at ADI 3,510 judgment, that after an intervening recess asked by Justice Menezes Direito, when requesting an advance of her vote, she expressed her position by briefly informing how she would vote, following the rapporteur, and that she would attach her written vote, which detailed the foundations of her vote.

I ask Justice Carlos Alberto Direito, with Court's permission, if I can state my vote immediately. (...) I am sure that Your Excellency, with your diligence, soon will bring the process. However, this seat brings me, unfortunately, the task of reminding the Colleagues we have, in line, to be called to judgment by this Plenary, no less than 565 other cases. So, I ask again to Justice Carlos Alberto Direito and Colleagues to excuse me for advance my vote to follow the eminent Rapporteur. I have a few reasons for my conviction - and I will bring them together later - that largely coincide with those that were brilliantly developed by Justice Carlos Britto. (...) For these reasons, which will be well explained in the words I have written, I conclude that the case is inadmissible, as the Rapporteur has voted. (Gracie, ADI 3,510, pp. 79-80)

Addressing central questions about deliberative process that interests us, especially dialoguing with Silva's (2013) arguments, the absence of ideas and arguments exchange between Justices during trial, and the very low probability of a Justice who has already given his vote backtrack after hearing his colleagues’ votes, as they would have made a public commitment in a certain position, we note that a Justice 's vote change of direction due to another Justice 's vote occurred 22% of the time – which is not negligible.

The Justices who most rectified the vote direction were Ayres Brito, three times, and Cezar Peluso, twice. And Justices Barbosa, Mendes, Pertence, Corrêa, Jobim and Rezek once each.

An example is ADI 1,648, requested by the National Confederation of Commerce, questioning the law of the State of Minas Gerais that the incidence of ICMS on divestment of a rescued by the insurer. Most of the court held to be partially founded, defeating Justices Jobim, Lewandowski, Barbosa and Britto. These last two changed their understanding after hearing the position and vote of other Justices.

Mr. President, at the trial that I participated in ADI 1,648/MG, reported by Justice Gilmar Mendes, I followed His Excellency, the Rapporteur, thus upholding the case. But I continued to reflect on the subject and had the opportunity to read Justice Nelson Jobim's vote in this same ADI that seemed to me precious and convincing, leading me to a repositioning of my technical judgment. Justices Ricardo Lewandowski and Joaquim Barbosa, in this session, summarized what I think today, but my inspiration for this vote is, above all, the vote of Justice Nelson Jobim. (Min. Ayres Britto, ADI 1,648, pp. 77).

Another example is ADI 2,586, in which the National Confederation of Industry questions Law no. 9,314, dated November 14, 1996, and Administrative Rule no. 503, dated December 28, 1999, of the Ministry of Mines and Energy, which regulate mining research. Divergence that appeared in decision-making process referred to the technique of judgment. Justice Carlos Velloso, the rapporteur, voted for not receiving the case regarding the ordinance and the dismissal of the case regarding the law. From its vote, a debate was opened on implications of not knowing the ordinance, or dismissing it with prejudice.

MIN. CORRÊA: I'm in full agreement with the eminent Justice -Rapporteur, although this question of dismissing with prejudice and not receiving the case bears great similarity. I'll prefer, technically, in this hypothesis, to say that, regarding the ordinance, the case is prejudiced. MIN. VELLOSO: Mr. Justice, we're not saying that the Ordinance is illegal. MIN. PERTENCE: It is alleged that, since the law is unconstitutional, because it delegated to the Justice the rate fixing, consequently, the ordinance, whatever its content that fixed the rate, will also be unconstitutional. Otherwise, in legal system, when not receiving the case regarding the ordinance: the law is unconstitutional. But the ordinance, which says that the tax rate is 10%, is constitutional, or regarding it, we didn't receive the case MIN. VELLOSO: Would be perfect if we were declaring the unconstitutionality of the law, but we are declaring the constitutionality of this. MIN. PERTENCE: If the issue is over because of the decision taken in the case of unconstitutionality of the law, what happened was injury about more. We don’t need to discuss the Ordinance anymore. MIN. VELLOSO: I think we're in an academic discussion of the best level. MIN. PERTENCE: No. Data venia, this is not academic, it's trial technique. If we don't receive the case, it would not be received in any case. What may happen because of one decision on the other is the loss. (Min. Maurício Corrêa, Min. Sepúlveda Pertence, Min. Carlos Velloso, ADI 2.586, pp. 26-27).

The debate continues around this technical question, and in the end, Justices Pertence, Corrêa and Jobim rectified their votes, to form a majority and dismiss the case both in relation to the law and in relation to the ordinance. In the words of Justice Corrêa, "Mr. President, it seems that better solution would be to dismiss the case in one issue, but I agree that, to end this discussion, all two issues should be dismissed” (pp. 32).

Besides observing vote rectification, it was common to find situations which a Justice outlined an understanding in a way during the debate, before delivering his vote, but when voting changed his understanding, attributing this change to the arguments of his colleagues, as in ADI 3,028, in which Justice Sanches claims to have changed his judgment after hearing the arguments of Justice Pertence.

I confess that at first was inclined to follow the eminent Rapporteur, but from the vote of the Justice Sepúlveda Pertence, I found that it is possible to reconcile the text of Article 273 with the Constitution, once it is free of the expressions "of the Public Prosecutor's Office". As for the other contested devices, I also adopt the foundations of the votes of Justices Sepúlveda Pertence and Moreira Alves and those who are in the same trend, to conclude in the same way. (Min. Sidney Sanches, ADI 171, pp. 84).

We also note in more than half of the decisions (56%), the explicit mention of a Justice to other Justices’ arguments in justifying their vote, bringing delivered arguments, either to refute them, to complement them, or to request further explanations. That is, in more than half of these decisions there was discussion and debate, with Justices forming their convictions from exchange of ideas throughout the trial. And in these decisions, we find several passages in which Justices indicate that they didn’t come to plenary with a finished vote, and that they were waiting for their colleagues’ position to form their understanding.

ADI 236 is an example, the Governor of the State of Rio de Janeiro required it, questioning the article 180 of the State Constitution, in your item II, which established the penitentiary police. In this case, Justice Marco Aurélio, commenting on Justice Moreira Alves's statement that the rapporteur hadn’t touched on the formal aspect of discussion, he said he was awaiting "with some anxiety" Justice Borja’s vote, "who requested an intervening recess in a process which he is well-versed", only then to decide whether the Constitution itself may decide on certain subjects (ADI 236, pp. 40-41).

Another example, in ADI 3,112 about the Disarmament Statute, Justice Marco Aurélio also makes an intervention that reinforces the existence of a deliberative process in plenary,

Madam President, I insist on the question I asked initially. In my view, we the others who are on the bench and do not examine the process, can’t, in a conscious way, express understanding about various devices. Hence the need to retake the tradition of, when the Direct Action of Unconstitutionality attack several legal precepts, submit to the Plenary device to device. I realize that we are picking up, by memory, certain articles. (...) Lately the practice has been this - the rapporteur really examines the whole of the diploma, causing a great difficulty for others that vote just by hearing. (Min. Marco Aurélio, ADI 3.112, pp. 68).

Similarly, the debate between Justice s Pertence, Britto and Aurélio in ADI 3,833, in which the Popular Socialist Party questions federal Legislative Decree n. 444, dated December 19, 2002, about the National Congress members’ remuneration during the 52nd legislature,

MIN. PERTENCE: Today, subsidies can only be fixed by legislative decree, not by law, in accordance to article 49, line VII, of the Constitution. MIN. BRITTO: There we have another disagreement, Justice. MIN. PERTENCE: We are here to disagree. MIN. AURÉLIO: The sum of different forces is the beauty of the Collegiate. (Min. Sepúlveda Pertence, Min. Ayres Britto, Min. Marco Aurélio, ADI 3,833, pp. 45).

We also record the frequency of citing precedents in the justification of the Justices' vote, and 78% of the decisions mention the "consolidated jurisprudence of the Court".

We also observed the frequency and justification of intervening recess by Justices throughout these decisions, since Silva (2013) also criticizes this regimental possibility as an obstacle to deliberative process. We agree that this may be one of the consequences, but we must consider others - for example, the possibility of an intervening recess to be motivated by a Justice's need to revise his vote, or to deepen the examination of a case once he has had contact with an argument different from his in plenary.

In 53% of decisions with tight division there was no intervening recesses, in 39% of cases there was one intervening recess, and in 8% two intervening recesses.

Justices Borja, Britto, Grau, Mendes and Alves requested intervening recess in two of these cases each, and Justices Aurelio, Direito, Gracie, Jobim, Mello, Pertence, Rezek, Velloso, Peluso and Silveira requested intervening recess in one of these cases, each one.

Data presented in the third report "Supremo em Números" (see FALCÃO, HARTMANN and CHAVES, 2014), account for a total of 481 ADIs with intervening recess, which corresponds to approximately 10% of the total ADIs in the period covered by the report (1988-2013). Average duration of ADIs with finished intervening recess period (94%) was 1.2 years, and the average duration with unfinished intervening recess period, at that time, was 3.7 years.

The mean duration of intervening recess in tight decisions didn't vary much in relation to these data, being 1.3 years in the first intervention recess, with median of 6 months, and 1.2 years in the second intervention recess, with median of 3 months. Intervening recess with the longest duration in these cases was 3.6 years - requested by Justice Nelson Jobim at ADI 1,648, noting that this decision had a second intervening recess, by Cezar Peluso, who had the second longest duration (3, 4 years) to return the case to plenary.

It was common in these decisions for Justices to focus on the importance of intervening recess, to better weigh the arguments and study the implications of the case, searching for a more balanced and wise decision. And, the observation, made by some Justices inclined to ask for an intervention recess, that the development of other Justices’ arguments and positions made them dispense this resource.

The amount of time that has passed since the vote of Justice Carlos Britto has enabled us to ponder prudently arguments as well as access to texts and information free of emotion. The request for an intervening recess made by Justice Carlos Alberto was wise. Without this space of time, necessary for the exercise of their own reflection at prhonesis, I wouldn't have managed to align the reasons that conform the vote that i shall formulate. Time is essential to the exercise of prudence, even if it causes inconvenience to the unwary ones. We will spend, in the explanation of our votes, the necessary hours for the correct performance of our profession. Noble profession, especially when faced with such complex subjects as this. It does not exist, or should exist, time limit for the delivery of our votes. I am sure that I speak on behalf of the whole Court, which is here to respect the Constitution, not to the convenience of those concerned. (Min. Eros Grau, ADI 3,510, pp. 316).

I was going to ask for an intervening recess, given the absence of controversy, but with the question posed by Justice Marco Aurelio and now by Justice Cezar Peluso, I do not see how to follow the Justice -Rapporteur. I could even adhere to the thesis, but I would not feel comfortable if I followed the premises of your Excellency's vote regarding the dismiss of the direct action for technical-legal reasons ... (Min. Gilmar Mendes, ADI 2,885, pp. 26).

With the split vote - five to five - that requires me to vote in a tie-breaker, I would normally ask for an intervening recess to deepen my study. But from the outset, with the permission of those who think otherwise, I became convinced to uphold this case. (Min. Sidney Sanches, ADI 139, pp. 35).

Last two aspects that we consider relevant to understand deliberative process are the frequency of direct adherence of one Justice to another’s vote, without adding justifications, and the role of the rapporteur in decision-making process.

Regarding the adhesion of one Justice to another’s vote, in 75% of these decisions at least one of the Justices directly adheres to the vote of another Justice, without adding arguments. An example is the full text of Justice Eros Grau’s vote in ADI 3,833, transcribed above.

Madam President, regarding the competence of the Supreme Court, I wish to reaffirm that it is a political court because it looks after the viability of the polis and provides it. It is a political Court because it must understand the uniqueness of each situation within the polis. Pedro Lessa's lesson, remembered by Justice Carmen Lúcia, is simply anthological. I would just like to regret the fact of voting after the demonstrations of those who preceded me, especially the vote of Justice Carmen Lúcia, which exhausts what I would have to say on the subject. I follow the rapporteur's vote. (Min. Eros Grau, ADI 3,833, pp. 28).

Regarding the role of the rapporteur, considering the majority of ADIs decided by a majority, in 85% of cases he was the winner, and considering only cases decided with tight division, the rapporteur was the winner 67% of the time.

Considering all the majority decisions, the Justice who was most defeated as rapporteur was Marco Aurélio, followed by Britto, Silveira and Corrêa (see figure 3). And in the ADIs with tight division, Marcus Aurelio was defeated 3 times, and Justices Velloso, Gracie, Grau, Galvão, Corrêa, Silveira, Gallotti, Pertence and Sanches were defeated once each.

Figure 3
Frequency of the rapporteur being defeated

That is, even in those decisions in which each vote counts, the rapporteur has significant influence in determining the outcome of the decision, and very often Justices follow their vote. Therefore, the "irrelevance" of the rapporteur pointed out by Silva (2013), can be questioned, since the position of the rapporteur, about the north and the fundamentals of the vote, is followed most often, and on several occasions, without addition of new arguments.

4. Final considerations

Our objective in this article was to analyze decision-making behavior of Supreme Court in cases of constitutionality control, examining the dynamics of collegiate functioning in cases that divided the court.

There were three guiding questions: (i) how often and in what situations Supreme Court was divided in ADIs trials? (ii) How did the composition of majority and minority blocs in those situations, i.e. who voted with whom? And (iii) how did Justices decide, that is, how did the deliberative process take place?

We note that the Supreme Court was quite consensual in terms of concentrated control over this period, with only 28% of total decisions made in ADI generating dissenting votes. Even more infrequently was the division of court, and in these 25 years only 3% of collegiate decisions caused tight divisions, in which each vote had considerable power.

In terms of composition of voting blocs, there was a lot of fluidity in court, with only two "cliques", one with Justices Alves, Sanches and Corrêa and another with Justices Alves, Corrêa and Jobim, both achieving more than 80% cohesion. And although Justice Marco Aurelio was a recurring minority vote, we observed when dealing with more complex issues in which each vote counts, the Justice tended to have a more fluid behavior, showing high cohesion with Justice Carlos Velloso.

We also observe, despite the great fluidity, there are some factors that make coalitions more predictable, such as the combination of the ADIs subject and the Justice 's career before STF, and there are strong indications that Justices who came from magistracy are more likely to vote together than to split their votes.

We found deliberative process in court was intense. Reading judgments of decisions in which each vote counts, that is, those more complex cases that divided court, we found enough elements to relativize decision-making personalism theses, and to affirm the importance of collegiate deliberation in the construction of STF decisions.

Also, we found many references to other elements that seem to influence the decision in addition to the legal factors of the case, as the public opinion and the actors and experts called to stand in court, either through public hearings, as amici curie, or by mediatic manifestation. Justice Gilmar Mendes’ vote in ADI 3,510, about stem cell research, exemplifies these diverse influences, positioning the collegiate as a space for democratic deliberation,

Certainly, the alternative of the passive attitude of self-restraint - or, in some cases, of greater restraint, using the Garcia de Enterria's expression - would have been more detrimental or less beneficial to our democracy. Supreme Court demonstrates, with this judgment, that it can be a People's House, just like the parliament. A place where multiple social aspirations and the political, ethical, and religious pluralism find their place in procedural and argumentatively organized debates in previously established norms. Public hearings, in which experts are heard on the subject under discussion, the intervention of the amici curiae, with their juridical and socially relevant contributions, as well as the intervention of the Public Prosecutor, as representative of the whole society before the Court, and public and private advocacy groups, in the defense of their interests, make this Court a democratic space as well. A space open to reflection and legal and moral argumentation, with wide repercussion in community and democratic institutions (Min. Gilmar Mendes, ADI 3,510, pp. 465-466).

It’s evident that our study has an exploratory character, and there is a need for new researches focusing on investigating these elements. But the evidence gathered here allows us to conclude that when cases are complex, with tight votes in which each vote counts, the role played by collegiate is central to the construction of Supreme Court decisions.

In this study, we found evidence pointing possibilities for relativizing theses of decision-making personalism, demonstrating that when the court divides, decisions result from a deliberative process with an intense exchange of ideas and arguments, even generating changes in the direction of votes already pronounced.

We can say, with Ferejohn and Pasquino (2010:372-373), that in the cases analyzed here, the Justices of the Supreme Court worked with persuasive speech, acting both internally, among themselves, to reach an agreement on their decisions, and outside, providing reasons for their decisions to broader audience. Thus, in decisions in which every vote counts, we could understand Supreme Court configuration as a deliberative space.

• 2
  Among the pioneers are CASTRO (1993), VIEIRA (1994), SADEK (1995), ARANTES (1997), TEIXEIRA (1997), and VIANNA et al (1999).
• 3
  OLIVEIRA, FALAVINHA and BRAGHIN (2015) mapped the annual meetings of Associação Nacional de Pós Graduação em Ciências Sociais (ANPOCS), researching working groups (GTs) that dealt with the Judiciary between 2010-2015, finding 18 articles related to the STF, in a total of 73 articles presented in these GTs. In 61% of the 18 articles the thematic focus is the result of the decisions; in 22%, the determinants of the decision-making process, and 17% bring a doctrinal-jurisprudential analysis of cases of great media repercussion, these being the three thematic categories with higher incidence. The authors also mapped the Scielo journal portal, with no temporal clipping, but searching articles using three keywords ("Supreme Federal Court," "judicialization of politics" and "judicial activism"), finding a total of 46 articles (exclusives) on the STF. Of these, 26% of the articles focus on the institutional design of the Supreme Court, 20% discuss the determinants of the decision-making process, 20% give the results of decisions on cases of great media repercussion, and 6% bring jurisprudential analysis, these being the four themes with higher incidence.
• 4
  See MAVEETY, 2003.
• 5
  We use here, purposely, the expression in the title of Riggs’s’ (1993) article: "When every vote counts".
• 6
  Unlike Riggs (1993) who considered only cases decided by a margin of one vote , we need to highlight we included cases decided by two-vote margin, extending the concept of "tight decision". This choice was basically because Brazilian court is more cohesive compared to American court.
• 7
  EC 45/2004 increased Supreme Court's discretion, adopting the "general repercussion", yet Supreme Court must judge as constitutionality control all cases that reach it. Thus, over time, Justices have developed procedural mechanisms to deal with the massive volume of cases. VERÍSSIMO (2008), for example, points to what he calls "informal certiorari", highlighting "the possible existence of a kind of Brazilian certiorari, that is, a procedural filter that may be allowing the court to manage, informally, its expressive workload "(2008: 416). Among the mechanisms of this informal certiorari it would be, for example, monocratic decisions.
• 8
  Analyzing the number of cases decided by US Supreme Court in 50 years, between 1953 and 2003, in a total of 12,004, SEGAL and SPAETH (2002) found only 39% of these decisions were taken unanimously.
• 9
  We believe that excluding an outlier of a maximum composition of 11 Justices in each decision, out of a total of 34 different Justices who took part in the court in that period, has different implications from the usual practice of excluding outliers from a sample in demographic and population studies, for example. Thus, as a control for this type of outlier, we chose to observe in the configuration of voting blocs only those cases in which more than one Justice voted in the minority. We do not question the validity of the studies that make the option to exclude Justice Marco Aurélio, but we believe that for the present study, given the interest presented, the form of control adopted is more appropriate.
• 10
  Examples are ADI 335, in which the Court found the request for active illegitimacy "ad causam" of the Central Única dos Trabalhadores prejudiced, and Justice Marco Aurélio disagreed, acknowledging it not as a member of the trade union system, but as a national class entity of and giving it legitimacy for the request; ADI 912 in which the Justice suggests the lack of national coverage of the Associação Brasileira dos Professores do Ensino Público should not be immediately contested, suggesting the association proves to have this characteristic, but he's defeated by the majority voting for its active illegitimacy, and he conclude for "framing in the permissive constitutional allusive to the legitimacy"; ADI 1,788, in which it recognizes the legitimacy of the Associação Nacional dos Registradores de Pessoas Naturais, stating that they are also notaries in a specific segment association compatible with the most comprehensive association of Notaries and Registrars of Brazil; And ADI 2,353, in which it recognizes the legitimacy of the Associação Nacional do Ministério Público in Courts of Accounts (of States, Federal District and Municipalities).
• 11
  Examples are ADI 17, proposed by the National Federation of Engineers, and ADI 466, proposed by PSB.
• 12
  See, for example, ADI 2,387, regarding the questioning of a law that regulates the functioning of private pension entities. The judgment of the decision stated that "The case-law of the Supreme Court is firm that the decree’s subject, which, under the pretext of regulating a particular law, goes beyond its scope of incidence, is a matter that lies at the level of legality, not of constitutionality. In this case, the decree in question does not have an autonomous nature, being circumscribed in an area that, pursuant to Law No. 6,435 / 77, is subject to regulation, regarding the determination of adequate minimum standards of economic and financial security for the benefit plans or for the preservation of liquidity and solvency of the benefit plans alone, and of private pension fund as a whole”. Justice Marco Aurélio’s vote recognized the autonomy of the decree and could therefore be attacked, justifying by stating that "everything recommends the granting of the injunction, thus avoiding the filing of countless cases, which would have an imaginable outcome, including preliminary injunctions, only serving to block up, even more, judicial machine."
• 13
  This case was judged together with the ADIs 3,137, 3,198, 3,263, 3,518, 3,535, 3,586, 3,600, 3,788 and 3,814.
• 14
  The concept of clique (panelinha) is used here according to Oliveira (2012a).
• 15
  In this article, we do not differentiate type of admission to the magistracy. Thus, we consider as coming from a career in magistracy Justices who (i) have exercised position of magistrate before his appointment to STF, (ii) have entered the career through public tendering or (iii) through the constitutional fifth.

Referências bibliográficas

• ARANTES, Rogério Bastos. (1997), Judiciário e política no Brasil. São Paulo, Idesp/Sumaré.
• ARGUELHES, Diego W. e RIBEIRO, Leandro M. (2015). “O Supremo Individual: mecanismos de atuação direta dos Ministros sobre o processo político”. Revista Direito, Estado e Sociedade, n.46, p. 121-155.
• CASTRO, Marcus Faro de. (1997), “O Supremo Tribunal Federal e a judicialização da política”. Revista Brasileira de Ciências Sociais, 12 (34): 147-155.
• DESPOSATO, Scott W., INGRAM, Matthew C., and LANNES, Osmar P. (2015). "Power, Composition, and Decision Making: The Behavioral Consequences of Institutional Reform on Brazil's Supremo Tribunal Federal." Journal of Law, Economics, and Organization 31. DOI: 10.1093/jleo/ewu018
  » https://doi.org/10.1093/jleo/ewu018
• FALCÃO, Joaquim, HARTMANN, Ivar, CHAVES, Vitor. (2014). III Relatório Supremo em Números: O Supremo e o tempo. Rio de Janeiro: Escola de Direito da Fundação Getulio Vargas. Disponível em: http://hdl.handle.net/10438/12055 Acesso em 07dez2015.
  » http://hdl.handle.net/10438/12055
• FEREJOHN, John, & PASQUINO, Pasquale. (2010). The Countermajoritarian Opportunity Journal of Constitutional Law. Vol 353. Disponível em: http://scholarship.law.upenn.edu/jcl/vol13/iss2/6 Acesso em: 07dez2015.
  » http://scholarship.law.upenn.edu/jcl/vol13/iss2/6
• FERREIRA, Pedro Fernando de Almeida Nery (2013). Como decidem os ministros do STF: pontos ideais e dimensões de preferências. Dissertação (Mestrado em Economia). Universidade de Brasília, Brasília. Disponível em: http://repositorio.unb.br/handle/10482/13565 Acesso em: 27nov2015.
  » http://repositorio.unb.br/handle/10482/13565
• FERRREIRA, Pedro e MUELLER, Bernardo. (2014). “How judges think in the Brazilian Supreme Court: Estimating ideal points and identifying dimensions”. Economia, Vol 15 (3), pp. 275- 293. DOI:10.1016/j.econ.2014.07.004.
  » https://doi.org/10.1016/j.econ.2014.07.004
• KAPISZEWSKI, Diana (2011). “Tactical Balancing: High Court Decision Making on Politically Crucial Cases”. Law & Society Review, Vol. 45 (2), pp. 471–506. DOI: 10.1111/j.1540-5893.2011.00437.x.
  » https://doi.org/10.1111/j.1540-5893.2011.00437.x
• KLAFKE, Guilherme F. e PRETZEL, Bruna R. (2014). “Processo Decisório no Supremo Tribunal Federal: aprofundando o diagnóstico das onze ilhas”. Revista de Estudos Empíricos em Direito, Vol. 1 (1), pp. 89-104.
• MAVEETY, Nancy. (Ed.). (2003). The Pioneers of Judicial Behavior. Ann Arbor: The University of Michigan Press. 433pp.
• MENDES, Conrado Hübner (2010). Onze Ilhas. Folha de São Paulo: Tendências e Debates, São Paulo, 01.02.2010, disponível em: http://avaranda.blogspot.com/2010/02/conrado-hubner-mendes.html Acesso em: 07dez2015.
  » http://avaranda.blogspot.com/2010/02/conrado-hubner-mendes.html
• OLIVEIRA, Fabiana Luci de (2012a). Processo decisório no Supremo Tribunal Federal: coalizões e 'panelinhas'. Revista de Sociologia e Política (UFPR), v. 20, p. 139-153. DOI: http://dx.doi.org/10.1590/S0104-44782012000400011
  » http://dx.doi.org/10.1590/S0104-44782012000400011
• _________(2012b). “Supremo relator: processo decisório e mudanças na composição do STF nos governos FHC e Lula”. Revista Brasileira de Ciências Sociais, Vol. 27 (80), pp. 89-115. DOI: http://dx.doi.org/10.1590/S0102-69092012000300006
  » http://dx.doi.org/10.1590/S0102-69092012000300006
• OLIVEIRA, Fabiana Luci de. FALAVINHA, Diego H. S. BRAGHIN, Simone. (2015). “Processo decisório no STF e o caso da Reforma do Judiciário”. Revista Direito e Práxis, vol. 6, p. 365-394. DOI: 10.12957/dep.2015.18739.
  » https://doi.org/10.12957/dep.2015.18739
• PRITCHETT, Herman. (1948). The Roosevelt Court; A Study in Judicial Politics and Values, 1937-1947. New Orleans: Quid Pro, LLC.
• RIBEIRO, Leandro Molhano; ARGUELHES, Diego Werneck & PEIXOTO, Vitor. (2009), "Processo decisório, Judiciário e políticas públicas". Trabalho apresentado no 33º Encontro da Anpocs.
• RICE, S. A. (1928). Quantitative methods in politics. New York: Knopf.
• RIGGS, Robert E. (1993) "When Every Vote Counts: 5-4 Decisions in the United States Supreme Court, 1900-90", Hofstra Law Review: Vol. 21(3), pp. 667-724.
• ROSEVEAR, Evan; HARTMANN, Ivar & ARGUELHES, Diego (2015). Disagreement on the Brazilian Supreme Court: An Exploratory Analysis. (October 31, 2015). Available at SSRN: http://ssrn.com/abstract=2629329 Acesso em 10jun2016.
  » http://ssrn.com/abstract=2629329
• SADEK, Maria Tereza Aina. (1995). A Crise do Judiciário Vista Pelos Juízes: resultados da pesquisa quantitativa. São Paulo: Sumaré, 1995. v. 1. 250p.
• SILVA, Virgílio Afonso da. (2009). O STF e o controle de constitucionalidade: deliberação, diálogo e razão pública. Revista de Direito Administrativo, São Paulo, n. 250, pp. 197-227, jan./abr.
• _________(2013). “Deciding without deliberating”. International Journal of Constitutional Law, Vol. 11 (3), pp. 557-584. DOI: 10.1093/icon/mot019.
  » https://doi.org/10.1093/icon/mot019
• SUNSTEIN, Cass R. (2015). “Unanimity and Disagreement on the Supreme Court”. Cornell Law Review, Vol. 100 (4), pp. 769-823. Disponível em: http://cornelllawreview.org/files/2015/05/Sunsteinfinal.pdf Acesso em 28dez2015.
  » http://cornelllawreview.org/files/2015/05/Sunsteinfinal.pdf
• TEIXEIRA, Ariosto. A judicialização da política no Brasil (1990-1996). (Dissertação). Universidade de Brasília. Brasília, 1997.
• VERISSIMO, Marcos Paulo (2008). A constituição de 1988, vinte anos depois: suprema corte e ativismo judicial "à brasileira". Revista Direito GV, vol.4, n.2, pp. 407-440.
• VIANNA, Luiz Werneck et al. (1999). A Judicialização da política e das relações sociais no Brasil Rio de Janeiro, Iuperj/Revan.
• VIEIRA, Oscar Vilhena. (1994). O Supremo Tribunal Federal: jurisprudência política São Paulo, Malheiros Editores.

1

The research was supported by FAPESP. This author thanks the two anonymous reviewers for their careful reading, criticism, and suggestions for this article.

Publication Dates

• Publication in this collection
  July 2017

History

• Received
  10 July 2016
• Accepted
  17 Oct 2016