The presidential calculous in the appointment of ministers in Brazil

Abstract

This article aims to identify the factors that influence the selection of ministers in Brazil, considering the number of portfolios and their importance and saliency. We identify the regional influence on such selection, controlling for ministers’ personal characteristics and trajectories, including political aspects, expertise, and proximity, and loyalty to the president. We use an original dataset that covers the entire period known as the New Republic.

Keywords:
coalitional presidentialism; ministries; government; political elite; cabinet

Resumo

Este artigo procura identificar os fatores que influenciam a escolha dos ministros no Brasil, considerando a importância/saliência de cada um dos ministérios. Buscamos conhecer a influência dos aspectos regionais sobre tal escolha, controlando as características e trajetórias pessoais dos ministros, seja no que diz respeito aos aspectos políticos e técnicos, seja com relação à proximidade e à lealdade ao presidente da República. Para isso, utilizamos um banco de dados inédito, que abrange todo o período da Nova República.

Palavras-chave:
presidencialismo de coalizão; ministérios; governo; elites políticas; gabinete

Resumen

Este artículo busca identificar los factores que influyen la selección de ministros en Brasil, considerando la importancia y protagonismo de cada uno de los ministerios. Buscamos identificar la influencia de los aspectos regionales en esta selección, controlando las características y trayectorias personales de los ministros, ya sea en lo que respecta a aspectos políticos y técnicos o a la proximidad y lealtad al presidente de la República. Para ello, utilizamos una base de datos inédita, que cubre todo el período de la Nueva República.

Palabras clave:
presidencialismo de coalición; ministerios; gobierno; élites políticas; gabinete

1. INTRODUCTION

This article seeks to contribute to the discussion on the influence of regional political forces over the formation of the central government in Brazil. It evaluates the impact of the regional factors on the selection of federal ministers. Specifically, it explores whether there is a relationship between ministries’ political importance and the regional origin or political connection of ministers. Our hypothesis is that ministers from wealthy regions, especially from the Southeast and the state of São Paulo, tend to occupy the ministries most coveted by politicians.

In addition to the debate on regional influence, this study contributes to the political relevance of ministries. Thus far, a great majority of research in this area has considered only the number of ministries as if they were all equal. However, we know that it is not what happens in the real world (Abranches, 1988; Amorim, 2000).

Batista (2018) and Mauerber and Pereira (2020) are exceptions. They consider the extent to which ministries are attractive to politicians based on their budgets, regulatory capacities, and distribution of jobs or patronage potential. However, the focus of these authors is different than ours. They do not examine recruitment of ministers, regional factors, personal characteristics, or other aspects that can motivate the president’s choice. Instead, these studies analyze the distribution of cabinet portfolios among parties, the degree of coalition coalescence, and other possible consequences for legislative production. This study examines ministries’ attributes as the dependent variable whereas such characteristics are treated as the explanatory variables in aforementioned studies.

Another contribution of this study, which is rare in previous studies, is that we examine diverse factors that can affect ministerial selection simultaneously: politicians, experts, and proximity to the president. Upon selecting their ministers, presidents make complex calculations considering not only their relationships with the legislature but also regional political forces, organized groups in the society, voter satisfaction, the implementation of public policies, personal characteristics of ministers, and their own intuition and conviction, which gives personal characteristics to the administration. Still, it is important to control for the effect of other variables that can influence cabinet formation. Former President Fernando Henrique Cardoso’s (FHC) testimony on the choice of one of his ministers demonstrates how these elements are considered:

The question of the Ministry of Science and Technology had to be resolved. It seemed to me that appointing a deputy with no connection with the area would be forcing the issue, [...] then I remembered that, in Minas, Israel Vargas could remain. This solution pleased everyone I talked to in the state and whom I consulted later: President Itamar Franco and Francelino Pereira, who was very upset because the PFL had supported me and they had nothing there in Minas. I also had a conversation with Jorge Bornhausen [then president of PFL]. They wanted another name. I said: “Look, I really like Ivan Botelho, but I just can’t put him into the scientific community.” They understood. The next day, it appeared in the newspapers that Vargas would be the PFL’s indication. It’s not like that, but he had a really good relationship with Francelino, and it calmed many areas of Minas. On the eve of the announcement, I called Vargas and said: “Look, Zé” - he is my friend from so many years - I’ll need you, I don’t know where.” And he only learned on the radio that he was going to be Minister of Science and Technology. The Minas question was solved in this way (Cardoso, 2015).

Finally, another contribution that our study makes is the coverage of the period under investigation. Our data cover the entire period of the New Republic, totaling approximately 37 years. In the following section, we review the scholarly discussion on the subject.

2. THE DIVERSE FACTORS CONSIDERED IN THE PRESIDENT’S CALCULUOS

According to Camerlo and Pérez-Liñan (2015), in order to exercise their mandates well, presidents need aides that offer three principal resources: political support, expertise, and loyalty. Presidents attain political support primarily from those that have political and party experiences. The second, expertise or technical skills, is obtained through the appointment of technocrats who are less vulnerable to pressures from political parties and voters. Presidents secure loyalty by appointing closest allies from the same party or those with close personal ties, such as friends.

In Brazil, the discussion over cabinet appointment focuses on political support, involving the distribution of cabinet portfolios to political parties, coalition formation, and the resulting support by the coalition parties for the agendas of government interest in the National Congress. Although scholars generally agree on the executive dominance over the legislature, studies tend to concentrate on the legislature and not on the president or the presidency of the Republic. We know little about the internal dynamic of the executive branch (Batista 2013; Rennó 2016). Our study adopts a different strategy — we examine the selection of ministers through the eye of the president while controlling for individual characteristics of ministers and their proximity to the president.

With respect to the political context, Abranches (1988), in one of the most influential studies in Brazilian political science, already alerted to us that the selection of ministers in Brazil involved two pillars: partisan and state/regional. In his autobiography, former President Fernando Henrique Cardoso presented to us diverse evidence regarding how the two logics function in practice. We cite one of them below:

The two peemedebistas chosen were from the South, giving a margin to regional demands. How is the Northeast, where the PMDB and the PFL have so much presence? And here came the argument of regional inequalities to support yet another ministerial claim. In the conversation with the PFL, I had chosen a pernambucano for the Ministry and now, to please the PMDB, the proposal was to include a paraibano. But then, how would Bahia do, which was claiming the Ministry of Transport, if this were assigned to the PMDB of Rio Grande do Sul? The package was large (Cardoso, 2006).

In this paper, we put the party issue on the side and instead focus on regional influence. In the view of D’Araújo and Ribeiro (2018), “the ministry is one of the main resources used in search of the precarious federative balance. It permits the systematic search for alternative compensations to mitigate the abysmal inequalities among regions.” This discussion is long-standing and has been investigated in various ways. The literature has stressed the preponderance of the wealthy states and regions, especially that of the Southeast and particularly the state of São Paulo. Some scholars suggest the existence of a bargain between the legislature and the executive involving positions.

Soares (1973) observed that the limit through the 1969 constitutional amendment on the number of deputies from São Paulo “did not provoke any reaction, because the political forces of the state perceived that it would be better to abandon the problem of representation in the legislature and focus on the representation in the important organs and positions in the public administration, such as the Finance and Planning Ministries.”

In an empirical analysis, Abranches (1988) found that a combination of partisan and regional factors went into the appointments of ministers for politically and economically strategic ministries, which were, in his opinion, Finance and Agriculture. The state of São Paulo predominated the appointments to lead the Finance Ministry with 41% of all appointments as Finance Minister. Pernambuco dominated Agriculture by securing 59% of the appointments. The author suggested that the essentially regional criterion determined the allocation of the “spending” (or “clientelist”) ministries. The Ministries of Education and Health were captives of Bahia; São Paulo, Rio de Janeiro and Paraná also dominated the Ministries of Education and Culture; and the Ministry of Transport and Public Works had 43% of its ministers from Rio de Janeiro. However, when looking at a broader period (1889-2019), Neiva and Hiroi (2020) found that such assumptions did not hold. Souza (2006) went so far as to say that “the regional policy, at critical moments in Brazilian politics, has supplanted the parties as a focus of agglutination of political forces.” According to the author, the federative bargaining produced the over-representation of the least developed states in Congress and the predominance of the most industrialized states, especially of São Paulo, in the most salient economic policy areas. Souza argues that this arrangement resulted from a concession by the elites of the largest states to their allies in small states with the goal of fighting the common enemy — the emerging sectors of great urban and industrial centers. In exchange for the overrepresentation of small states in the legislative arena, the elites of São Paulo would have participation in the various bureaucratic-administrative niches, aside from parties and Congress. In other words, the author sustains that the representation of the economic interests of São Paulo was exercised through non-legislative representative channels, by direct articulation with the bureaucratic-public agencies. This would explain why the São Paulo elite has invested little in strengthening political parties.

The studies by Soares (1973), Abranches (1988) and Souza (2006) covered the periods of the New Republic (1945-1964) and the military rule (1964-1985). Ames (1986) argues that these practices were maintained into the most recent democratic period when affirming that São Paulo permitted other states to appropriate small clientelist projects while staying in total control of macroeconomic policy, investments, and fiscal and exchange rate policy. According to Ames, paulistas were neither underrepresented nor overrepresented in the ministerial positions. However, they dominated crucial organs, including the Finance Ministry, the Ministry of Public Works, and the Banco do Brasil.

Later studies have partially confirmed such hypotheses, by verifying the predominance of the wealthiest regions in the ministerial offices. Meneguello (1998) identified great influence of the Southeast region, which came to obtain 67% of the positions in the FHC government. Amorim (2004) observed that the three important states of the region (São Paulo, Rio de Janeiro, and Minas Gerais) administered half of President Lula’s first ministry in 2003. Like other authors, he suggests that such preponderance would be a form of compensating large states for their parliamentary under-representation.

However, in a subsequent study that analyzed a longer period, Amorim and Santos (2013) found that the frequency of Rio de Janeiro politicians in the executive branch fell from 15.7% (in the democratic period between 1946 and 1964) to 5.3% (during the military rule) and to 4.6% during the 1985-2010 period. The authors also identified the fall of São Paulo and the rise of Minas Gerais and Rio Grande do Sul during the military regime.

Neiva and Hiroi (2020) reinforced some of these findings by identifying that the South and Southeast regions jointly accounted for 68% of the federal ministers in the Republican period (1889-2017). In the case of the powerful Ministry of Finance, this number reached 73%. If we consider only the democratic periods, the South and Southeast combined held the Minister of Finance for 83% of the appointments to the position. With respect to the Ministry of Planning, another important ministry, Southeast politicians occupied the portfolio no less than 88% of the times. In contrast, the Northeast held the position only 6% of the times, and the North and Midwest never came to occupy such position. In regard to the military ministries, 78% were in the hands of the South and the Southeast, especially the states of Rio Grande do Sul (20.4%) and Rio de Janeiro (44.6%).

Obviously, the regional factors are not sufficient to explain presidents’ decisions on cabinet portfolio allocations. An ideal situation would be for each minister to have technical capacity and political articulation, as well as the loyalty to and personal trust of the president. According to Loureiro and Abrucio (1999), “more and more politicians are basing their actions on technical discourse and bureaucrats are reinforcing the political aspect in their calculation of performance.” However, as this is not always possible, the president must seek to maintain this balance collectively. In the end, the president’s team will have politicians, experts, expert politicians, personal friends, and representatives from various segments.

It was not possible to include in this study variables pertaining to ministers’ parties or their participation in Brazil’s coalition government. They are important, but our study emphasizes other aspects that can also influence the choice of the president. It is also worth noting that the number of partisan ministers in presidential systems tends to be much lower than that in parliamentary systems (Amorim & Samuels, 2010; Martínez-Gallardo & Schleiter, 2015). In fact, we calculated that one-third of the ministers had no party affiliation at the time of the appointment. If we consider that among party members, there are also those who are not necessarily indicated by the parties (Camerlo & Pérez-Liñan, 2015), the proportion of non-partisan appointments can be close to 40% or more. Nonetheless, we have included some variables of political nature: whether the minister had run for an election before the appointment, whether he held leadership positions in the legislature, and whether he was a legislator or governor.

Nor are analyses focused solely on political aspects sufficient to understand the complexity associated with the appointment of ministers. Camerlo and Martinez-Gallardo (2017) warn that the president also needs the support of extra-parliamentary groups, technocratic elites, and public opinion. Camerlo and Pérez-Liñan (2015) state that to ensure governance, presidents rely on the support of organized groups such as labor unions, churches, and the military, which can be crucial at certain historical times. In the same vein, Inácio (2013) recalls that, although the parties interfere directly in the selection of ministers, they do so under constraints imposed by the president.

Camerlo and Pérez-Liñan (2015) also point out that the data on party affiliation are fluid, not always reliable, difficult to determine, and do not necessarily reflect the political alliances between the Executive and the Legislative. Frequently, partisan ministers are only sympathizers, with no political weight in their party. Therefore, they are not necessarily nominated by their parties.

It is also necessary to keep in mind that ministers may have interests of their own which differ from the interests of the president. Ministers, as representatives of the party, may have ideological and political preferences, electoral projects, and their own agenda (Batista, 2018, 2013). As such, the risks and costs of monitoring and controlling their agents (ministers) by the principals (presidents) increase as the distance of preferences between them increases. In this manner, the appointment of ministers without party affiliation can contribute to decreasing the risks of agency (Codato & Franz, 2018; Martínez-Gallardo & Schleiter, 2015; Palloti & Cavalcante, 2019). Moreover, the complexity of the governmental apparatus, the heterogeneity among ministries, the presence of non-partisan and technocrat ministers have increased, and the legitimacy of parties declined (Camerlo & Pérez-Liñan, 2015; Inácio, 2013). All this has made room for new approaches. In Latin America as a whole, scholars have become increasingly interested in the topics of formation, stability, and duration of ministers in office (Rennó, 2016).

Such changes also brought a renewed interest in other characteristics of ministers: personal/individual characteristics, such as gender, educational level, profession, and technical competence; representativeness of the political and social forces that helped elect the president but are not necessarily tied to political parties, such as ethnic, religious, professional, student, business, and social movement groups, among others; and “outsiders,” who are close and loyal to presidents (Carmelo & Pérez-Liñan 2015).

In a similar vein, Palotti and Cavalcante (2019) consider that the appointment of ministers can expand the representative character of the government from different perspectives by including the political and social forces that supported the government and are not necessarily tied to the alliances in the National Congress or the implementation of public policies. The authors maintain that regional, ethnic, and religious aspects, pressure and interest groups, social movements, personal and professional attributes, and the characteristics of the state bureaucracy also matter in this complex equation.

Research on this subject in Brazil is rare but exists. Codato and Franz (2018) found that both “political” and “non-political” ministers had extensive experience, including administrative experience. The authors found little difference between the PSDB and PT governments, contradicting somewhat the recurrent view that the FHC government promoted more managerial offices, while Lula (PT) politicized the Executive (Inácio, 2013).

Palotti and Cavalcante (2018) sought to identify the impact of variables related to political affiliation, education, and professional experience. The authors observed that, in general, ministers were experienced and had good professional training. Loureiro, Abrucio, and Rosa (1998) evaluated the selection of ministers in the José Sarney, Fernando Collor de Mello, Itamar Franco, and FHC governments based on five criteria: party, technical qualification, affiliation with interest groups, and public and federal opinion. We present our data in the following section which include several variables discussed in this section.

3. PRESENTATION AND INTERPRETATION OF DATA

Our study covers nearly all federal ministers appointed in the New Democratic Republic from 1985 to 2021 based on the data collected from the Center for Research and Documentation of Contemporary History (CPDOC) of the Getúlio Vargas Foundation (FGV), newspaper collections, the official websites of the Chamber of Deputies and the Senate, among others. Since our unit of analysis is the appointment as a minister, the data refer to attributes, experiences, and events before the appointment. It does not, therefore, matter which office the minister held after the mandate.

In addition to the heads of ministries, our data also include those who held the “status” of minister in the Brazilian government, with the exception of Central Bank (BC) governors and Attorney Generals of the Union (AGU) because their selections are essentially technical. We also did not consider interim ministers and military ministers except those appointed to the Ministry of Defense.

During the period of this study, Brazil was governed by eight presidents, who altogether exercised a total of eleven terms. In total, the country had 94 different ministries, which can be reduced to 39 if we disregard the mere name changes or if we aggregate them by area of activity. Our unit of analysis is appointments. Our dataset identified 578 appointments, independent of the durations of office, over a period of almost 37 years. We counted as a new observation (appointment) every time a minister was sworn in, including cases in which ministers were maintained when presidents or presidential terms changed.

The total number of individuals is lower at 419 persons. This is because our coding rule counts as a new observation the appointment of a minister to another ministry, or to the same ministry at another time. Similarly, following previous studies (Codato & Franz, 2018; Palotti & Cavalcante, 2018), we count as a new appointment the maintenance of a minister when the president is re-elected. Former President Dilma Rousseff, for example, appears in the dataset three times, as Minister of Mines and Energy and twice as Presidential Chief of Staff, in the first and second terms of Lula.

The dependent variable reflects the saliency and relevance of ministries, as proposed by Zucco, Batista and Power (2017). The measure is based on the responses of 142 parliamentarians to a questionnaire in which they assessed the attractiveness of each of the ministries to politicians. The authors used the “Bradley-Terry” probability model based on paired comparison of ministries. This method allows to avoid more difficult ordinal classifications of a large number of options. The index varies from 0 to - 4.21. The lower the number (i.e., the more negative number), the less attractive the ministry is. The value “0” therefore indicates the ministry most valued by politicians.

The authors applied the same questionnaire to a group of 283 specialists affiliated to the Brazilian Political Science Association (ABCP). The two indices showed high correlation between them, with a linear coefficient of 0.87. Our study utilizes the index based on the survey of politicians because we believe that politicians, especially parliamentarians, have a more accurate view of the context that surrounds them in their daily lives. This is also the evaluation of Mauerberg and Pereira (2020), who affirmed that “legislators’ views truly reflect what really matters when someone discusses ministries, since these politicians are usually responsible for appointing their party fellows to occupy such positions.”

We used as a basis the 37 existing ministries between 2014 and 2015, the last two years of Dilma Rousseff’s administration. Since the ministries are not exactly the same throughout the whole period, we classified those that were missing by approximation.¹ This strategy is not perfect, since in addition to the differences in structure and attributes, different governments prioritize ministries differently. Nevertheless, we believe that this method of matching ministries by approximation does not invalidate our study. Rather, we have two advantages. First, Rousseff’s second term had the largest number of ministries in the current democratic period, which reduced the issue of matching ministries to the ones under other governments. Second, political saliency of ministries has been relatively consistent over time. The Ministry of Finance, the Ministry of Planning, and the Office of the Presidential Chief of Staff, for example, have been powerful in any government. In contrast, the Ministries of Culture and the Environment have been undervalued by all presidents.

Our main explanatory variables relate to regional factors. They will be tested in three different ways: across the region (the Southeast, compared to all other regions) and the state (São Paulo, compared to all other states), and using an index that reflects the disproportionality of representation of each state in the Senate (the federative house and more malapportioned). In the first two cases (Southeast and São Paulo), the expected sign is positive, and in the latter, the expected sign is negative as economically developed states tend to be underrepresented in the legislature. To some extent, the disproportionality tests the hypothesis raised by some authors (Amorim, 2004; Soares, 1973; Souza, 2006) that the wealthy states ceded in the overrepresentation of the poor states and their underrepresentation in the legislature in exchange for their preponderance in the federal government, especially in the more expressive ministries.

The literature has indicated that there is a preponderance of ministers from the Southeast, mainly from São Paulo (D’Araújo & Ribeiro, 2018; Palotti, 2017). D’Araújo and Ribeiro (2018) also observed that the region exercised a greater participation in the ministries related to the economy, finance, management, and planning, being responsible for 85% of these positions. Our data partially confirm these findings: More than half of ministers come from that region. They also show that the region prevails in some relevant ministries, such as Finance (82%), Planning (67%), Office of the Presidential Chief of Staff (59%) and Health (62%).

The control variables are based on the profiles of the respective ministers, involving four main aspects: personal, political, technical, and loyalty to the president. In the first aspect, the values found are close to those identified by other authors (Codato & Franz, 2018; D’Araujo & Ribeiro, 2018; Palotti & Cavalcante, 2019). Ministers of the New Democratic Republic are, on average, 55 years old, 92% are men, 93.4% have higher education, 24.2% have master’s degrees, and 20.6% hold doctorates.

We found 58 different professions, which we aggregated into 14 categories to facilitate the analysis. While it is common for ministers to have more than one profession or occupation, we consider only the main one, often connected to their training area. Although it is not a decisive characteristic for appointments, the presidents seem to take it into account, as the data show. Economists predominated in the Ministries of Finance (36%) and Planning (48%), while teachers had an important participation in the Ministry of Education (57%). In Justice, we found half of appointments to be lawyers. In Foreign Affairs, 53% were diplomats. In Health, 62% were doctors. To avoid excessive information, we only consider the most important professions in our regression analysis.

We classify ministers as “political” or “non-political” on the basis of whether they had participated in an election or had been affiliated to parties prior to their cabinet appointments. These are two clear forms to demonstrate interest in politics. Approximately 67% of the ministers in our dataset met at least one of these criteria, a number close to 66% found by Palotti and Cavalcante (2018). It shows that despite the importance of the parties, presidents have a margin of maneuver since they fill a reasonable part of the cabinet without regards to parties. This margin is even greater if we consider that not all ministers with party affiliation are in fact nominated by their party. The appendix provides the percentage of politician ministers in each of the most important ministries.

With regard to the main elective positions held before ministerial appointments (federal deputy, senator, and state governor), 53% of ministers had held at least one of those positions prior to their appointment. This percentage is slightly higher than the 45% found by Palotti and Cavalcante (2018): 38% were federal deputies; 14% were senators; and 10% were governors.

According to Inácio (2013), party and legislative leadership is a relevant aspect for the selection for a minister. According to the author, parties tend to privilege the internal elite and professional politicians who tend to have strong ties with their organization. Indeed, 122 of the 578 ministers in our sample (21%) had experienced at least one party or legislative leadership position. It remains to be seen whether such positions are politically the most valued. We created a variable for political leadership which takes the value of one if a ministerial appointee held at least one of the following leadership positions: the president of the Chamber of Deputies or the Senate; the leader of government or of a party in the National Congress; the president or secretary-general of a party; a founder of a medium-size or a large party; and the chair of a legislative caucus (bancada).

Still in the political “dimension,” the analysis considers the exercise of social leadership of two types. One reflects an administration position in an organized societal group in diverse areas, such as student, professional, artistic, sport, religious, and minority group organizations. Thirteen percent of ministers fall into this classification. The other type refers to business groups and includes 45 ministers. The category covers those who were prominent in the society, either because they participated in the administration of business associations — such as the Federation of Industries of the State of São Paulo (Fiesp), National Confederation of Industry (CNI), and Brazilian Federation of Banks (Febraban), or because they ran large companies. Their participation in most ministries is low or none. The Ministries of Agriculture (29%), Finance (32%), and Industry and Trade (40%) are exceptions.

We also consider variables of “technical” or “technocratic” nature, which involve the exercise of positions that are not strictly political in nature. In accordance with the literature, ministers in general have high technical qualifications, almost all recieved higher education, and 30% holds master’s and/or doctorate. Technical expertise also includes administrative experience. Codato and Franz (2018) found that most of the ministers had administrative experience when they took office. They found little variation among ministries’ functional areas (political, social, or economic), with the exception of social ministries in the PT governments. The ministers appointed by the PT governments to lead social ministries had less administrative experience.

Although Codato and Franz (2018) consider the secretaries of ministry more political than technical, we regard ministry secretaries as more technical in nature. We coded as secretaries individuals who held ministry secretary positions in specific functional areas and those who served as executive secretaries (akin to deputy minister). Ministry secretaries are often those who know the ministry well and who serve as an alternative in situations where the president seeks to give a more technical character to the ministry, when the minister’s position is vacant, or at the end of a government’s term. Pereira, Praça, Batista and Lopes (2015) suggest that presidents use ministry secretaries to monitor ministers, especially when they are from a party far from the president’s preferences. We found 147 former ministry secretaries, equivalent to a quarter of our sample.

Presidents have some room to choose whomever they would like to appoint, within the so called “personal quota (cota pessoal).” Under the personal quota, presidents appoint individuals whom they trust, such as close friends, advisors, and long-time aides who have accompanied them in their political careers. Their close relationships could be independent of their party or political connections, although some of these could come from the president’s own party. We identified 70 ministers, or 12% of the sample, in this situation. Also enter the “personal quota” ministers who participated in the election campaign for the presidency, as coordinators, treasurers or special advisers in a certain area. In our sample, 86 ministers (15%) are in this situation.

We also include in the president’s personal quota celebrities appointed to direct a ministry as political “outsiders.” Celebrities lack in prior political experience and are often chosen to bring greater popularity and visibility to the government. We identified 12 ministers in this category. In the group are soccer players Pelé and Zico, volleyball player Bernard, actress Regina Duarte, and singer Gilberto Gil, among others.

4. MULTIVARIATE ANALYSIS

The variables presented in the previous section are not sufficient, individually, to explain the factors considered by presidents when making their choices. A more appropriate analysis should consider them together so that we can evaluate the influence of each variable while controlling for the influence of all others.

In the following table, we present three different linear regression models. In this way we can evaluate separately the behavior of the variables that relate to regional factors: the fact of being from the Southeast or São Paulo and the disproportionate representation in the Senate, the federative house. They also allow for testing supposedly close variables in separate models, such as being an entrepreneur and having a business connection. With the three models, we can have greater confidence in our conclusions since it enables us to evaluate their robustness in different contexts.

TABLE 1
DETERMINANTS OF SELECTION OF MINISTERS BASED ON THE RELEVANCE OF MINISTRIES

Independent variables	Model 1	Model 2	Model 3
Southeast	0.11 (0.87)	-	-
Disproportionality of state in the senate	-	-0.00 (-0.12)	-
São Paulo	-	-	0.28** (2.08)
Age (years)	-0.00 (-0.63)	-	-
Woman	-0.82*** (-3.90)	-0.83*** (-4.09)	-0.80*** (-3.96)
Layer	-0.08 (-0.52)	-	-
Economist	0.98*** (4.86)	0.98*** (5.04)	1.02*** (5.23)
Entrepreneur	0.29 (1.53)	-	0.27 (1.55)
Business connection	-	0.46*** (2.29)	-
Doctorate	0.31** (2.05)	0.33** (2.33)	0.29** (2.01)
Político (politician)	-0.03 (-0.16)	-	-
Federal deputy	0.05 (0.34)	-	-
Senator	0.35 (0.86)	-	-
Governor	0.25 (1.23)	0.29 (1.51)	0.26 (1.39)
Political leadership	0.24 (1.57)	0.25** (1.87)	0.23* (1.68)
Ministry secretary	0.09 (0.66)	-	-
National technical position	0.17 (1.18)	-	-
Social leadership	-0.48*** (-2.72)	-0.53*** (-3.18)	-0.55*** (-3.23)
Celebrity	-1.23*** (-3.14)	-1.24*** (-3.27)	-1.25*** (-3.31)
Personal proximity to president	0.23 (1.29)	0.18 (1.04)	-
Participation in electoral campaign	0.04 (0.27)	-	-
Constant	-2.52 (-6.62)	-2.79 (0.12)	-
Observations	576	578	578
R-squared	0.18	0.18	0.18

• *** p<0.01, ** p<0.05, * p<0.1. T-values in parentheses. Obs.: Although not shown in the table, regressions were estimated with dummy variables that control for fixed effects for all governments. Source: Elaborated by the authors.

We begin the interpretation of the results with our main hypothesis which involves the regional question. The variables related to the regional question present varying results. The literature has indicated that being from the Southeast increases the probability of occupying a ministry (Amorim, 2004; D’Araújo & Ribeiro, 2018; Meneguello, 1998; Neiva & Hiroi, 2020). However, our analysis shows that being from the Southeast is not related to the political weights of ministries. The t-value of 0.87 for the variable “Southeast” (model 1) reveals that the region is far from statistically significant.

Indeed, the presence of ministers from the Southeast in important ministries is significant, but they also have a strong presence in the ministries coveted less by politicians, such as Sports (87%), Women’s Rights (83%), and Racial Equality (83%). These data suggest that the Southeast leads not only an economic agenda, but also a “post-materialist” agenda, which prioritizes individual values of autonomy and self-expression over materialistic, physical, and economic values (Inglehart, 2015).

In any case the results for the Southeast region are not conclusive. It should be remembered that although it facilitates the analysis, the aggregation of very different states into the same region can generate distortions. In the case of the Southeast, for example, it covers both the state of São Paulo, the most populous and the economic center of the country, and Espírito Santo, a much smaller state with less than 2% of the national population.

The variable that identifies the disproportionality in the representation of states in the Senate also did not present statistical significance (model 2). It weakens the hypothesis advanced by several authors that the wealthier states have accepted their underrepresentation in the legislature in exchange for their preponderance in the most important positions in the Executive. This hypothesis has found support with regard to the number of ministries occupied, but not with regard to the saliency of ministries.

It is important to note, however, that the disproportionate representation in the legislature is very concentrated in the under-representation of São Paulo, including in the Chamber of Deputies — the house in which representation, in principle, should be proportional to states’ population size. This is due to the constitutional amendment that limits the number of seats in the lower house to the maximum of 70, which affects only the state of São Paulo. Thus, we cannot totally rule out the possibility of bargaining between large and small states. Perhaps the measure that considers all states does not exactly capture the pact that may have been struck between São Paulo and the smaller states, and not among all states of the federation. Even for this reason, it is important to reserve a variable that identifies that particular state.

Model 3 presents the results of the same test, but with the variable São Paulo instead of Southeast or disproportionality in representation. The variable São Paulo is statistically significant at the level of 5%, suggesting that the state predominates not only in the number of ministries it occupies, as identified by several authors (D’Araújo & Ribeiro, 2018; Figueiredo, 2007; Palotti, 2017) but also in the relative weights they have. The results show that there is an association between the state of São Paulo as a minister’s home state and the saliency of the ministry to which the individual is appointed.

These findings reinforce the old hypotheses (Abranches, 1988; Ames, 1986; Soares, 1973; Souza, 2006), which have never been effectively tested, suggesting the prevalence of São Paulo in the federal government, both in numerical terms and in relation to the federal agencies it controls. The state, in fact, occupied some of the most important ministries in the New Democratic Republic.

Obviously, the regional factor alone also does not explain the occupation of ministerial posts. Thus, it is essential to control for other factors that supposedly influence ministerial appointments. With regard to personal characteristics, the results show that there is no relationship between the age of the minister and the importance of the ministry the individual occupies. Being a woman is inversely related with the level of ministerial saliency in all three models. This is because ministries linked to social policy, such as welfare and minority policy, are much less valued by members of Congress and are more directed toward women. These include Racial Equality (67%), Social Development (36%), and Human Rights (35%). In important ministries, such as Finance and Planning, women led those ministries only 4.6% and 4.7% of the time, respectively. This finding indicates that the idea that the women’s area of activity is that of family, care, reproduction, and welfare persists in Brazil, a concept that shows in the government’s actions. It also shows how misleading it is to evaluate the participation of women in politics only by numbers.

With regard to the professions of ministers, we can observe that the most frequent of them (lawyer) is distributed more or less equally regardless of the saliency of the ministries. The variable for lawyer, with a t-value of -0.52, is far from statistically significant. On the other hand, economist ministers tend to be allocated to more important ministries — the variable presented high statistical significance in all three models.

In general, entrepreneurs may have had some space in the most relevant ministries. The variable for entrepreneurs was close to achieving statistical significance, with a p-value of 0.12 (not shown in the table) in two models. However, this is a very wide-ranging variable, involving all types of entrepreneurs. The importance of the segment appears more clearly in the variable that indicates if the minister had any connection with large companies, either by occupying managerial positions or participating in business associations. In this case, it presented high statistical significance at the level of 1%.

The results show that the most expressive ministries received a greater proportion of ministers with good academic training, measured here by the conclusion of a doctoral course. The variable was significant at the level of 5% in all three models.

In relation to technical qualification, we observed that the fact that the minister had occupied ministerial secretariats shows no impact on our dependent variable. This does not mean that the occupation of positions of this kind does not matter in the conquest of a ministry. As we have seen, the frequency of former ministry secretaries among ministers is large - a quarter of them went through such positions. However, this happens in both strong and weak. We also evaluated the occupation of what we call “national technical positions,” which was not significant. However, the diversity of positions considered — ranging from ministers of higher courts to legislative advisers, presidents and directors of state organs, such as banks, National Institute of Social Security (INSS) etc. — may make it difficult to reach more definite conclusions.

We analyze the political dimension in various forms. The first one, with the “politico” variable, informs if the minister had any political ties before being sworn in. It was not significant in the model in which it was included. Another way we assessed the political dimension was through the elective positions held. Having an experience as a federal deputy or a senator did not present statistical significance, that is, it did not show clear correlation with the importance of the ministry the individual occupied. One possible explanation is that legislators tend to specialize in certain public policies, which entitles them to occupy ministries in related areas. While those ministries may be electorally interesting to such ministers, they may not necessarily be the most coveted by the majority. This may help explain the findings by Batista (2018) that “coalition partners do not respond differently in terms of legislative support in light of the different assets’ distribution within the portfolio allocation.”

The variable “political leadership” presented inconsistent results - no statistical significance in the first model and significant at 5% in the second and 10% in the third. It seems that party elites tend to occupy the most important ministries. It is worth remembering that leaders always have a prominent role. If they themselves do not become ministers, they often take part in the nominations.

The variable “social leadership” was unequivocally significant in all three models (at the level of 1%). The negative sign of the coefficients suggests that having social leadership experiences favors the selection of the individuals by presidents for less politically prestigious ministries. These are generally in the areas where social leaders are active: minorities, human rights, social inequality, and the environment. According to Codato and Franz (2018), these are actors who become “minister-able” without an intense career, be it technical or political.

The three final variables are more linked to the personal decisions of presidents. In the case of celebrities, the variable presented negative sign and high statistical significance in all three models. The results are not surprising, as most of them are concentrated in the Ministries of Sports and Culture. These are well-known individuals among the public, such as musicians, athletes, and actors.

The variables “personal proximity” and “participation in the campaign” reflect greater closeness and loyalty to the president. However, neither of them was significant in the models. Most people close to the president end up participating in the government, but not necessarily in the most salient ministries.

During the period under examination, the country experienced several different governments. We also controlled for the fixed effect of each of the governments with dummy variables. These variables are not included in the regression table because they are not variables of interest and we wished to avoid excessive information for the reader.

Nevertheless, it is worth making a special note on the Jair Bolsonaro government. Bolsonaro’s choices of ministers differed from all his predecessors. He has been in power for two and a half years (as of this writing), and his appointments account for only 7% of the ministers appointed throughout the New Democratic Republic. This makes it difficult to have robust conclusions from a statistical point of view. However, we were able to see some differences. The Bolsonaro government allocated the largest share of cabinet portfolios to entrepreneurs: 24.4%, against the general average of 11.6%. Bolsonaro was the second president to hire more politicians from his home state: 29.3%, against the average of 19.4%. He was the second in terms of selecting fewest ministers with doctorates: 9.8%, compared to 20.6% of the general average.

The Bolsonaro government was also known for the space it gave to the military. At the beginning of the term, Veja magazine reported that Bolsonaro was the second president in the number of military personnel recruited to the cabinet (Romano, 2019). At that time, Bolsonaro’s cabinet had eight military ministers. It was only behind Castelo Branco, who appointed twelve military personnel to his government. As of this writing in August 2021, Bolsonaro has already surpassed this mark, if we consider the absolute number of appointments: no less than thirteen of the forty-one ministers who passed through his government were military.

The small number of paulistas in Bolsonaro’s government is also noteworthy. Only 7.3% of his ministers are paulistas, compared to 24.4% in the general average. It is curious that the most ‘radical’ president of the entire New Democratic Republic period thus far is also the one with the least ministers from São Paulo. Perhaps it is one more way to oppose the parties that predominated the country’s government in recent decades (the PT and PSDB), which hold a strong base in São Paulo. It is difficult to say precisely what his strategy is, but it seems that the regional factor also enters his political calculation.

5. FINAL CONSIDERATIONS

This article seeks to evaluate the factors considered by presidents in selecting their ministers. Unlike most studies in the area, we are concerned not only with the number of ministries, but above all with their political weights. Another difference from previous research is the use of multivariate models in order to simultaneously, and thus more appropriately, examine the behavior of each variable. It is worth emphasizing that presidents make choices taking into account various factors. They are present in the individual nominations and collective results. In the end, the cabinet must reflect a complexity of predominant political forces in society, the need for political support and to produce efficient public policies, electoral prospects, and the style and the intuition of the president.

Overall, we found that factors related to the economic environment and the business class seem to contribute to the occupation of the most important ministries. There are also some indications that less salient ministries have connections with the representation of minorities, either by the greater presence of women or by the presence of social leaders.

Our principal question concerns regional influence. Our findings partially support previous hypotheses. It was not possible to confirm any relationship between the alleged bargaining (or compensation) over the disproportionate representation in the legislative and executive branches. That is, contrary to the hypotheses presented by the literature, legislatively underrepresented states do not always occupy the most salient ministries, and legislatively overrepresented states do not necessarily take less salient ministries.

The data, however, suggest that something happens in the wealthiest and most underrepresented state. São Paulo holds not only the largest share of ministerial positions but also the most important ones. Such positions are related to some corporate representation. To complicate the situation, ministries linked to the “post-materialist agenda,” which are not necessarily prestigious, receive a large contingent of ministers from São Paulo.

It is necessary to consider that São Paulo was the state most affected by the under-representation in both houses of Congress, mainly due to the limited number of deputies. Perhaps political bargaining took place not between regional blocs, but between São Paulo and poor states.

Today, we experience a profound dispute between the Bolsonaro government and the governors over the policies related to the pandemic. This dispute appears to be taking place in the absence of parties. Political science seems “half lost” in understanding the phenomenon. Obviously, we do not claim to resolve this now; we merely draw attention to other relevant factors that we should address. The debate on relations between regional political forces and the central government is one of them.

We believe that the discussions around the presidency, as well as those relating to federalism, need to be reopened. There is much work to be done. Future studies may take other paths. It is important to investigate the role of other states in this game. It is very curious, for example, the predominance of the military from Rio Grande do Sul and Rio de Janeiro. The role of Minas Gerais, which has been dubbed “the faithful balance of the federation,” is also worth investigating.

REFERÊNCIAS

• Abranches, S. (1988). Presidencialismo de coalizão: o dilema institucional brasileiro. Dados, 31(1), 5-38.
• Amorim, O. Neto. (2000). Gabinetes presidenciais, ciclos eleitorais e disciplina legislativa no Brasil. Dados, 43(3), 479-519. Recuperado de https://doi.org/10.1590/S0011-52582000000300003
  » https://doi.org/10.1590/S0011-52582000000300003
• Amorim, O. Neto,. (2004). O Poder Executivo, centro de gravidade do sistema político brasileiro. In L. Avelar, & A. O. Cintra (Orgs.), Sistema político brasileiro: uma introdução São Paulo, SP: Fundação Editora Unesp.
• Amorim, O. & Samuels Neto, D. (2010). Democratic regimes and cabinet politics: a global perspective. Revista Ibero-Americana de Estudos Legislativos, 1(1), 10-23. Recuperado dehttps://doi.org/10.12660/riel.v1.n1.2010.4123
  » https://doi.org/10.12660/riel.v1.n1.2010.4123
• Amorim, O Neto & Santos, F. (2013). O Rio de Janeiro e o Estado Nacional (1946-2010). Dados, 56(3), 467-496. Recuperado dehttps://doi.org/10.1590/S0011-52582013000300001
  » https://doi.org/10.1590/S0011-52582013000300001
• Batista, M. (2013). O poder no Executivo: uma análise do papel da presidência e dos ministérios no presidencialismo de coalizão brasileiro (1995-2010). Opinião Pública, 19(2), 449-473. Recuperado dehttps://doi.org/10.1590/S0104-62762013000200009
  » https://doi.org/10.1590/S0104-62762013000200009
• Batista, M. (2018). Who gets what and how does it matter? Importance-weighted portfolio allocation and coalition support in Brazil. Journal of Politics in Latin America, 10(3), 99-134. Recuperado de https://doi.org/10.1177/1866802X1801000304
  » https://doi.org/10.1177/1866802X1801000304
• Camerlo, M., & Martinez-Gallardo, C. (2017). Government formation and minister turnover in presidential cabinets: comparative analysis in the Americas Nova York, NY: Routledge.
• Camerlo, M., & Pérez-Liñan, A. (2015). The politics of minister Retention in presidential systems technocrats, partisans, and government approval. Comparative Politics, 47(3), 315-333. Recuperado dehttps://doi.org/10.5129/001041515814709310
  » https://doi.org/10.5129/001041515814709310
• Cardoso, F. H. (2006). A arte da política: a história que vivi Rio de Janeiro, RJ: Civilização Brasileira.
• Cardoso, F. H. (2015). Diários da Presidência 1995-1996 São Paulo, SP: Companhia das Letras.
• Codato, A., & P. Franz(2018). Ministros-técnicos e ministros-políticos nos governos do PSDB e do PT. Revista de Administração Pública, 52(5), 776-796. Recuperado de https://doi.org/10.1590/0034-7612174301
  » https://doi.org/10.1590/0034-7612174301
• D’Araujo, M. C. S., & Ribeiro, G. L (2018). Trajetória socioeducacional dos ministros brasileiros na Nova República (1985-2014). Revista de Sociologia e Política, 26(65), 39-61. Recuperado de https://doi.org/10.1590/1678-987317266503
  » https://doi.org/10.1590/1678-987317266503
• Inácio, M. (2013). Escogiendo ministros y formando políticos: los partidos en gabinetes multipartidistas. América Latina Hoy, 64, 41-66. Recuperado dehttps://doi.org/10.14201/alh.10231
  » https://doi.org/10.14201/alh.10231
• Inglehart, R. (2015). The silent revolution: changing values and political styles. Nova Jersey, NJ: Princeton University Press.
• Loureiro, M. R., & Abrucio, F. L. (1999). Política e burocracia no presidencialismo brasileiro: O papel do Ministério da Fazenda no primeiro governo Fernando Henrique Cardoso. Revista Brasileira de Ciências Sociais, 14(41), 69-89. Recuperado de https://doi.org/10.1590/S0102-69091999000300005
  » https://doi.org/10.1590/S0102-69091999000300005
• Loureiro, M. R., Abrucio, F. L., & Rosa, C. A. (1998). Radiografia da alta burocracia federal brasileira: o caso do Ministério da Fazenda. Revista do Serviço Público, 49(4), 46-82. Recuperado de https://doi.org/10.21874/rsp.v49i4.400
  » https://doi.org/10.21874/rsp.v49i4.400
• Martínez-Gallardo, C., & Schleiter, P. (2015). Choosing whom to trust: agency risks and cabinet partisanship in presidential democracies. Comparative Political Studies, 48(2), 231-264. Recuperado dehttps://doi.org/10.1177/0010414014544361
  » https://doi.org/10.1177/0010414014544361
• Mauerber, A. Jr & Pereira C. (2020). How valuable is a presidential cabinet? Measuring ministries’ political attractiveness in Brazil. Latin American Politics and Society, 62(1), 25-45. Recuperado de https://doi.org/10.1017/lap.2019.44
  » https://doi.org/10.1017/lap.2019.44
• Palotti, P. L. D. M., & Cavalcante, P. L. C. (2018). Does one size fit all? An analysis of portfolio allocation in the Brazilian multiparty presidential system. Opinião Pública, 24(2), 427-455. Recuperado de https://doi.org/10.1590/1807-01912018242427
  » https://doi.org/10.1590/1807-01912018242427
• Palotti, P. L. D. M., & Cavalcante, P. L. C. (2019). Estratégias de nomeações ministeriais: entre a política e as políticas públicas. Revista de Sociologia e Política, 27(70), e001. Recuperado dehttps://doi.org/10.1590/1678987319277001
  » https://doi.org/10.1590/1678987319277001
• Pereira, C., Praça, S., Batista, M., & Lopes, F. (2015). A nomeação de secretários-executivos e o monitoramento da coalizão no presidencialismo brasileiro. In F. Lopes (Org.), Cargos de confiança no presidencialismo de coalizão brasileiro (pp. 139-164). Brasília, DF: IPEA.
• Romano, G. (2019, fevereiro 18). Bolsonaro supera Geisel, Médici e Figueiredo em ministros militares. Veja Recuperado de https://veja.abril.com.br/politica/bolsonaro-supera-geisel-medici-e-figueiredo-em-ministros-militares/
  » https://veja.abril.com.br/politica/bolsonaro-supera-geisel-medici-e-figueiredo-em-ministros-militares/
• Soares, G. A. D. (1973). Desigualdades eleitorais no Brasil. Revista de Ciência Política, 7(1), 25-48. Recuperado de https://bibliotecadigital.fgv.br/ojs/index.php/rcp/article/view/59226
  » https://bibliotecadigital.fgv.br/ojs/index.php/rcp/article/view/59226
• Souza, M. C. C. (2006). Federalismo no Brasil: aspectos político-institucionais (1930-1964). Revista Brasileira de Ciências Sociais, 21(61), 7-40. Recuperado de https://doi.org/10.1590/S0102-69092006000200002
  » https://doi.org/10.1590/S0102-69092006000200002
• Zucco, C Jr Batista, M., & Power, T.. (2019). Measuring portfolio salience using the Bradley-Terry model: an illustration with data from Brazil. Research & Politics, 6(1), 1-8. Recuperado de https://doi.org/10.1177/2053168019832089
  » https://doi.org/10.1177/2053168019832089

• 1
  Because we did not find an adequate equivalent, we left out the Ministry of De-bureaucratization and the Extraordinary Ministry for Irrigation Affairs.

• [Translated version] Note: All quotes in English translated by this article’s translator.

APPENDIX

TABLE A
NUMBER OF “POLÍTICOS (POLITICIAN)” MINISTERS IN PRINCIPAL MINISTRIES

Ministry	Políticos Ministers (%)	N	Importance according to legislators
City	90,0	10	0,00
Agriculture	87,5	24	-2,17
Institutional Relations	85,7	14	-2,90
Health	80,8	26	-1,69
Communication	80,0	15	-2,75
Agricultural Development	80,0	15	-3,98
Regional Integration/Development	78,3	23	-2,17
Secretary-General of the Presidency	75,0	12	-3,06
Education	73,9	23	-1,25
Transport	73,9	23	-3,64
Social Security	72,7	22	-2,61
Labor	72,7	22	-2,33
Defense	71,4	14	-3,22
Human Rights	70,6	17	-4,36
Justice	70,0	30	-2,45
Presidential Chief of Staff	68,2	22	-0,61
Sports	66,7	15	-4,82
Environment	64,7	17	-3,67
Social Development	64,3	14	-2,90
Science and Technology	63,6	22	-2,17
Tourism	61,5	13	-4,35
Planning	57,1	21	-0,15
Mines and Energy	55,6	18	-1,40
Industry and Commerce	55,0	20	-2,71
Foreign Relations	47,1	17	-2,40
Culture	43,5	23	-4,69
Finance	40,9	22	-0,24

• Source: Elaborated by the authors.

Publication Dates

• Publication in this collection
  21 Jan 2022
• Date of issue
  Sep-Oct 2021

History

• Received
  29 Dec 2020
• Accepted
  05 July 2021