Government Coalitions in Brazilian democracy

Introduction^*

The online encyclopaedia Wikipedia defines “government coalition” as “a cabinet of a parliamentary government in which several parties cooperate. The usual reason given for this arrangement is that no party on its own can achieve a majority in the parliament.” (Emphasis added.) This commonsense view predominated until very recently in academic circles. Studies of government coalitions were limited to countries with parliamentary systems and the comparative literature assumed that presidents, because they have fixed terms that are independent from the legislature, would not have incentives to form coalitions. The assumption that coalition governments are peculiar to parliamentary systems has, however, been demolished, both theoretically and empirically.¹

Coalition governments occur in 53.6% of the situations in which the president's party does not obtain a sufficient number of seats to govern alone (Cheibub, Przeworski and Saiegh 2004, 574). In other words, presidents also seek to form governments that increase their support in the legislature and the reasons leading them to do so are not much different from those that motivate prime ministers.

The literature on presidential coalition governments is small, though growing. Furthermore, its trajectory has been very different from that of studies on parliamentary coalitions. In 1990, in a book that aimed at consolidating the knowledge built up in the field, Laver and Schofield (1990, 7-11) pointed out the existence of a solid European tradition of empirical studies that, as well as containing important case studies, had also advanced in terms of comparisons between countries. Due to the greater availability of information and also because they benefit from the study of and models applied to parliamentary systems, comparative studies of government coalitions in presidential systems already display important results. However, individual case studies or comparative analyses of a few cases may still make an important contribution to our knowledge of the process of formation, functioning and performance of government coalitions in presidential systems.

But not even nowadays is the study of parliamentary governments able to do without case studies. On the contrary: recently, theoretically oriented studies have sought to organize, in as systematic a way as possible, aspects of coalition politics that would be very difficult to examine without a detailed examination of individual countries.²

This research note, which is part of a broader study on the performance of coalition governments in Brazilian democracy, aims to identify the governments formed under the last two democratic constitutions and to analyse their main characteristics. The first period begins with the promulgation of the 1946 Constitution and ends on 31 March 1964, with a military coup. The period beginning in October 1988, with the promulgation of the new constitution, is examined until the end of the 52nd Congress on 31 January 2007, thus including the first month of President Lula's second term, during which he made no changes to his ministry. This note also aims to contribute towards new studies and, who knows, to serve as an example, so that researchers from other countries may make similar efforts with regards to their own countries. In this way, it may contribute towards a more general objective, beyond the research project within which it is placed: the production of comparative studies based on more detailed and reliable information.

The formation of coalition governments: criteria and positions considered

The criteria used here to demarcate the end and the start of governments are the same defined by Müller and Strom for parliamentary systems: “1) any changes in the set of parties holding cabinet membership; 2) any change in the identity of the prime minister; and 3) any general election, whether mandated by the end of the constitutional inter-election period, or precipitated by a premature dissolution of parliament” (2000:12). These criteria are sufficiently general for them to be applicable to the formation of and changes in governments in the presidential system, with slight alterations that do not affect the comparability with parliamentary countries.

Presidents not only form government coalitions but alter them over the course of their terms, seeking to build a support base in the legislature. The constitutional rule that gives presidents the prerogative of choosing cabinet members does not grant them full autonomy in forming their government. If we take for granted the fact that their objective is obtaining support in the legislature, there is no reason to believe that the choice of future members of the government will not take into account the strength of parties in the legislature.

For Müller and Strom, cabinet members are “only those parties that have designated representatives with cabinet voting rights” (ibid.12), which excludes parties that support the government but do not hold cabinet posts. The point that must be emphasized is the difference between government coalitions and legislative coalitions. The formation of legislative coalitions results from ad hoc negotiations or simply from an identity of preferences. Government coalitions, on the other hand, imply access to jobs, to positions of power, in return for legislative support.

Criteria (2) and (3) are easily applied to the study of the formation of government coalitions in presidential systems. Changes in president, the head of the government, correspond exactly to changes in the identity of the prime minister in parliamentary systems. As for elections, although in presidential systems they cannot be brought forward, the regular ones already on the electoral calendar alter the correlation of forces in the government all the same. Thus the resulting changes in the number of seats must be taken into consideration in defining changes in government coalitions.³ However, certain peculiarities in the formation of a government in the presidential system make the identification of a government's party composition a little more complex, though not affecting the applicability of the first criterion.

In the parliamentary system, the leader of the party that obtains the most votes, in general is the person who becomes prime minister and takes on the role of forming a government, which emerges after a period of negotiation when the parties invited to be part of the government express their agreement in filling cabinet posts. The government is only considered to have been formed when the parties’ position is defined and — even where there is no formal vote of or official taking of office — the government is approved by the legislature. At this moment, the government's party makeup is also given. In the presidential system, negotiations between the president and the parties invited by her/him to participate occur in similar fashion. However, since the last word is the president's, the appointment of ministers affiliated to parties that did not offer a definitive position, or even that refused to support the government, can occur. In the presidential system, major party leaders often take up ministerial posts without the formal agreement of their party.

Studies on coalition governments in presidential systems generally determine the party composition of governments by means of ministers’ party affiliation, even if they lack their party's endorsement. In studies comparing a large number of countries there is no alternative to this, at least for now. The information on the stance taken by parties is not available. However, case studies not only can but should check whether the minister participates in the government on behalf of her/his party; in other words, whether the participation is of a party nature or of an individual nature.

If we take as our starting point the fact that the presidents seek to form alliances with a view to gaining legislative support, the strategy of co-opting members of parties that do not accept taking part in the government cannot be considered the most efficient. Even though they have the prerogative of choosing their ministry unilaterally, if presidents are rational, they will seek party support in the legislature. Otherwise, we would be making different behavioural assumptions for the same position, that of head of government. Additionally, the argument suffers from a sort of hyper-institutionalism that fails to consider the political conditions under which heads of government operate. This being the case, the study of the functioning and performance of coalition governments in presidential systems requires one to check the commitment of parties whose members accept taking part in government. This is the only possibility of checking the returns for the government in terms of parliamentary support.

Lastly, a few comments on the positions considered in the composition of the government. Participation in government was defined considering only ministerial positions. There are, however, relevant differences between the two periods. Firstly, one observes a significant increase in the number of ministries. On the one hand, this indicates the existence of a more complex governmental structure. On the other, considering the variation from one government to the next, it shows their importance in the political negotiations for the formation of the government. Between 1946 and 1964, the number of ministries remains largely stable. It rises from 10 to 13 with the division of the Ministry of Education and Health and of the Ministry of Labour, Industry and Trade. Furthermore, there are few extraordinary ministries, just two for the whole period, both in 1963: the Extraordinary Ministry of Planning, whose main task was preparing the Triennial Plan, and the Extraordinary Ministry of the Administrative Reform. In the current period, on the contrary, different governments have created a profusion of extraordinary ministries and/or sub-ministerial departments (Secretarias de Estado), some later transformed into ministries.

The second difference is that the Presidency of the Republic has also gained an organizational structure that has grown and become institutionalized, with posts carrying a status equivalent to that of a minister. For example, during the 1946-64 period, the Gabinete Civil (office of the civilian presidential staff), created in 1938, was a group of almost personal aides to the president. The head of the Gabinete Civil joins the ministerial structure during the Geisel government in the mid-1970s. The Gabinete Civil is transformed into Casa Civil (civilian household) after the government of President Collor (who abolished it), thus remaining until today. The high-level posts of the Presidency also include the General Secretary's Office, the departments of state for public administration, communication, institutional relations and others. All these departments with ministerial status were taken into consideration. Beyond these departments, the structure of the Presidency also includes military staffers at, for instance, the Department of Strategic Affairs and the Casa Militar (military household), and other departments — most of which created in the Cardoso government — such as the Chief Advocate's Office and the Chief Controller's Office, which were not taken into consideration. In fact, the military ministries were not considered either, given the non-political nature of these positions. However, the Ministry of Defence, created in 1999, was included in the calculations of the composition of coalitions identified here.⁴

Tables 1 and 2 summarize the number of posts per government in each period considering only the office holder.⁵

TABLE 1
NUMBER OF MINISTERIAL POSITIONS, BY TYPE AND BY GOVERNMENT COALITION – 1946-1964

COALITION/ PRESIDENT	MINISTRIES	SUB-MINISTERIAL DEPARTMENTS (SECRETARIAS)	EXTRAORDINARY MINISTRIES	TOTAL
DUTRA I	7	-	-	7
DUTRA II	7	-	-	7
DUTRA III	7	-	-	7
VARGAS I	8	-	-	8
VARGAS II	8	-	-	8
VARGAS III	8	-	-	8
CAFÉ FILHO	8	-	-	8
CAFÉ FILHO	8	-	-	8
N. RAMOS	8	-	-	8
JK I	8	-	-	8
JK II	8	-	-	8
QUADROS	10	-	-	10
GOULART I (NEVES I)	10	-	-	10
GOULART II (NEVES II)	10	-	-	10
GOULART III (B. ROCHA)	10	-	-	10
GOULART IV (H. LIMA)	10	-	1	11
GOULART V	10	-	2	12
GOULART VI	10	-	1	11
GOULART VII	10	-	1	11

• * Excluding military ministries.

TABLE 2
NUMBER OF MINISTERIAL POSITIONS, BY TYPE AND BY GOVERNMENT COALITION – 1988-2007

COALITION/PRESIDENT	MINISTRIES	SUB-MINISTERIAL DEPARTMENTS (SECRETARIAS)	EXTRAORDINARY MINISTRIES	TOTAL*
SARNEY 2	18	3	1	22
COLLOR 1	9	7	-	16
COLLOR 2	9	7	-	16
COLLOR 3	9	7	-	16
COLLOR 4	11	7	1	19
FRANCO 1	18	3	-	21
FRANCO 2	18	3	-	21
FRANCO 3	18	3	-	21
CARDOSO I 1	18	2	1	21
CARDOSO I 2	18	2	3	23
CARDOSO II 1	19	7	3	29
CARDOSO II 2	21	5	-	26
LULA I 1	24	6	1	31
LULA I 2	24	7	-	31
LULA I 3	24	7	-	31
LULA I 4	24	7	-	31
LULA I 5	24	6	-	30

• ^*
  Excluding military ministries; includes Ministry of Defence since 1999.

Party composition and duration of governments

In this section, the aim is to identify the party composition of the ministries, taking as criteria not only ministers’ affiliation, but also whether they took part in the government representing their party. This task is not always easy, for it is often in the party's interest to maintain some ambiguity, especially when it is divided. However, before presenting the most controversial cases and the decisions made, a few comments on the sources are called for.

The identification of ministers’ party affiliation was based on official sources available in publications and/or websites containing electoral and legislative data, as well as publications of biographical summaries and academic studies. Other sources used to check parties’ participation in government were testimonies and, mainly, the politics section of daily newspapers, especially for the 1946-64 period, which demanded more systematic research (see ^{Appendix 1}).

In the periodicals, news items on and journalistic analyses of negotiations preceding the formation of ministries were examined, both at the start of governments and when changes took place during the presidential term. These news reports are frequent in the national press in the two periods. The expression “coalition presidentialism”, coined in 1988 by Abranches (1988), has been incorporated into academic studies and become current in the press. In the 1946-64 period, political analyses and news reports used the expression “coalition government”.⁶ These sources allowed one to determine the nature of parties’ participation in the various governments, since they dealt not only with the process of negotiation, identifying the agents that participated, but also with the final result.

In a large number of cases, the party's agreement implies the very recommendation of a name to represent it in the government. When the president revealed a preference for a specific name, it had to obtain the party's endorsement anyway. When party members participate against the party's wishes, two situations arise. In the first, the party expels or temporarily suspends its member. In the second, the party does not impose any kind of punishment, but continues in active or overt opposition to the government. In these cases, there is no doubt as to the individual nature of the participation in the ministry. But a third situation also exists: in spite of not formally endorsing the participation of one of its members, the party does not define an unequivocal or unitary position in relation to the government. In this case, the decisions made about the inclusion of parties in coalitions are reported in more detail further on.

Charts 1 and 2 show the party composition of the government coalitions identified in the two periods, also indicating the event that gave rise to it and its duration, as well as the percentage of seats held by the parties in the coalition and by the president's party.

CHART 1
GOVERNMENT COALITIONS – OCTOBER 1946 TO MARCH 1964

Coalition / President	Party of the President	Parties in the Coalition	Date of the Coalition		Event Leading to he New Coalition	Duration		% of seats in the lower house
			Start	End		Days	Months	Party of the President	Coalition
Dutra I	PSD	PSD-PTB	31/1/1946	13/10/1946	Presidential & Legislative Elections	253	8	52,8	60,5
Dutra II	PSD	PSD-PR-UDN	14/10/1946	30/3/1950	PTB leaves, UDN and PR join	1246	42	52,0	81,3
Dutra III	PSD	PSD-UDN	1/4/1950	30/1/1951	PR leaves	299	10	52,0	78,0
Vargas I	PTB	PSD-PTB-PSP	1/2/1951	15/6/1953	Presidential & Legislative Elections	854	28	15,6	57,4
Vargas II	PTB	PSD-PTB	15/6/1953	2/6/1954	PSP leaves	347	12	15,6	50,0
Vargas III	PTB	PSD-PTB-PSP	3/6/1954	24/8/1954	PSP joins	81	3	15,6	57,4
Café Filho	PSP	PSD-PR-UDN	25/8/1954	31/1/1955	Vargas’ Suicide	156	5	0,0	78,2
Café Filho	PSP	PSD-PR-UDN	1/2/1955	11/11/1955	Legislative Election	280	9	0,0	80,7
N. Ramos	PSD	PSD-PTB-PSP-PR	11/11/1955	31/1/1956	Politico-Military Crisis	80	3	35,0	67,8
JK I	PSD	PSD-PTB-PSP-PR	1/2/1956	31/1/1959	Presidential Election	1080	36	35,0	67,8
JK II	PSD	PSD-PTB-PSP-PR	1/2/1959	31/1/1961	Legislative Election	720	24	35,3	68,4
Quadros	PTN	UDN-PSP-PR	1/2/1961	24/8/1961	Presidential Election	204	7	0,0	34,4
Ranieri Mazzili*		Interim president	25/8/1961	7/9/1961	Politico-Military Crisis	13	0	0,0	0,0
Goulart I (Neves I)	PTB	PSD-PTB-PDC	8/9/1961	12/10/1961	Prime Minister takes office	34	1	20,2	57,7
Goulart II (Neves II)	PTB	PSD-PTB PDC-PSP	13/10/1961	12/7/1962	PSP joins	269	9	20,2	65,3
Goulart III (B. Rocha)	PTB	PSD-PTB-PSB	13/7/1962	17/9/1962	Change in Prime Minister	64	2	20,2	58,3
Goulart IV (H. Lima)	PTB	PSD-PTB-PSB	18/9/1962	23/1/1963	Change in Prime Minister	125	4	20,2	58,3
Goulart V	PTB	PSD-PTB PSP-PSB	24/1/1963	7/6/1963	Change to Presidential system /Legislative Election	133	4	28,4	63,6
Goulart VI	PTB	PSD-PTB-PDC	8/6/1963	19/12/1963	PSB leaves, PDC joins	191	6	28,4	62,1
Goulart VII	PTB	PSD-PTB	20/12/1963	31/3/1964	Coup d’État	101	3	28,4	57,2
Total (for the period)						6530	216	26,6	60,1

• ^*
  Speaker of the Chamber of Deputies (lower house of Congress) who held the Presidency of the Republic in an interim capacity until a constitutional change to a parliamentary system was approved, in order to allow Vice-president João Goulart to take office. Henceforth, Mazzilli's government will be excluded from the government coalitions.

CHART 2
GOVERNMENT COALITIONS – OCTOBER 1988 TO JANUARY 2007

Coalition / President	Party of the President	Parties in the Coalition	Date of the Coalition		Event Leading to The New Coalition	Duration		% of seats in the lower house
			Start	End		Days	Months	Party of the President	Coalition
Sarney 2	PMDB	PMDB-PFL	6/10/1988	14/3/1990	Promulgation of new Constitution	518	17	41,4	63,03
Collor 1	PRN	PRN-PFL	15/3/1990	12/10/1990	Presidential Election	207	7	5,1	33,94
Collor 2	PRN	PRN-PFL-PDS	13/10/1990	31/1/1991	PDS joins	108	4	6,1
Collor 3	PRN	PRN-PFL-PDS	1/2/1991	14/4/1992	Legislative Election	433	14	8,0	34,59
Collor 4	PRN	PRN-PFL-PDS PTB-PL	15/4/1992	30/9/1992	PTB and PL join	165	6	6,2	43,54
Franco 1	No Party	PFL-PTB-PMDB-PSDB-PSB	1/10/1992	30/8/1993	Impeachment of the president	329	11	-	60,0
Franco 2	No Party	PFL-PTB-PMDB-PSDB-PP	31/8/1993	24/1/1994	PSB leaves, PP joins	144	5	-	59,64
Franco 3	No Party	PFL-PMDB-PSDB-PP	25/1/1994	31/12/1994	PTB leaves	336	11	-	55,27
Cardoso I 1	PSDB	PSDB-PFL-PMDB-PTB	1/1/1995	25/4/1996	Presidential & Legislative Elections	474	16	12,5	56,14
Cardoso I 2	PSDB	PSDB-PFL-PMDB-PTB-PPB	26/4/1996	31/12/1998	PPB joins	965	32	16,6	77,19
Cardoso II 1	PSDB	PSDB-PFL-PMDB PPB	1/1/1999	5/3/2002	Presidential & Legislative Elections	1144	38	18,3	73,88
Cardoso II 2	PSDB	PMDB-PSDB-PPB	6/3/2002	31/12/2002	PFL leaves	295	10	18,2	45,22
Lula I 1	PT	PT-PL-PCdoB-PSB-PTB-PDT-PPS-PV	1/1/2003	22/1/2004	Presidential & Legislative Elections	381	13	18,0	42,88
Lula I 2	PT	PT-PL-PCdoB-PSB-PTB-PPS-PV-PMDB	23/1/2004	31/1/2005	PMDB joins, PDT leaves	368	12	17,7	62,38
Lula I 3	PT	PT-PL-PCdoB-PSB-PTB-PV-PMDB	1/2/2005	19/5/2005	PPS leaves	108	4	17,7	57,7
Lula I 4	PT	PT-PL-PCdoB-PSB-PTB-PMDB	20/5/2005	22/7/2005	PV leaves	62	2	17,7	58,28
Lula I 5	PT	PT-PL-PCdoB-PSB-PTB-PP-PMDB	23/7/2005	31/1/2007	PP joins	548	18	17,7	69,59
Total (for the period)						6586	220	13,0	55,9

Generally speaking, identifying the level of parties’ commitment to the governments formed in the two periods was not very difficult. However, certain cases are controversial, the main ones being that of the UDN in the 1946-64 period and that of the PMDB in the current period.

The UDN was the main party in opposition to the PSD-PTB alliance that governed Brazil for almost the whole of the 1946-64 period, only reaching power in the brief Jânio Quadros government. Nominally, however, it took part in almost every government in the period and with significant figures. As shown by Chart 1, however, the UDN's participation as a party was much more restricted. Let us see why.

With the elections for the presidency and the national legislature in 1945, the government and the Congress had as their main task the restoration of constitutional order. At the same moment when the government was being formed, the National Constituent Assembly (NCA) was taking office. One of the first decisions that the NCA had to make referred to its own scope: would it take on ordinary legislative functions as well or be only a constituent assembly? The UDN, as an opposition party and, back then, harbouring in its ranks leftwing deputies who had fought against the Vargas dictatorship, proposed that the constituent congress should also play an ordinary legislative role. The rationale behind this position was the following. If its function were strictly a constituent one, the president would carry on ruling by decree, in line with the still valid 1937 Constitution. The government parties, the PSD and the PTB, the former in particular (the president's party), defended the opposite position. And since they had the majority, they prevailed. Congress would play only the constituent role, with no participation in the tasks of government. This being the case, the UDN leadership refused to take part in the first coalition of the Dutra government, in spite of his invitation, remaining in opposition until the promulgation of the new constitution.

With the re-establishment of constitutional democracy, the NCA elected the vice-president of the Republic, also from the PSD, and got organized to participate in government policymaking from then on. A ministerial reform was expected for soon after the promulgation of the constitution. But it occurred slowly, beginning in October and ending only in December with the incorporation of two UDN ministers. The vice-president was a central figure in the ministerial reform. He was seen as “the virtual chief of Brazilian politics”, who was “making an effort to re-articulate the negotiations with a view to reestablishing the coalition”, suggesting “new possibilities for the UDN to cooperate in future with the government's administrative work, in two or three ministries” (Mundo da Política, Correio da Manhã, October 1,1946).

For the opposition party, the main issue at stake now was the election for state governments scheduled for October 1947. The state governments were still in the hands of interventores (appointed governors), many of whom had been replaced by the new president. The UDN, which took part in some of these appointed governments after democratization, foresaw difficulties in ensuring legitimate elections if it remained completely outside the administrative machine of the states. If it maintained its opposition to the federal government, it might find itself in a position of having to abandon its posts in state governments as well.

The new ministry would mark in fact the start of the first democratic government, with the participation of the legislature in government policymaking. The decree no longer existed in the new constitution. The uncompromising opposition that had characterized UDN's actions during the NCA now threatened the party's electoral objectives. The position of the party's state deputies was a pragmatic one. They wanted to remain in the state governments and increase their electoral potential. In mid-October, the president invited Clemente Mariani for the Ministry of Education and Health. The latter linked his acceptance to the solution of the “Bahia case”, i.e., the support of all the parties to the candidacy of Octavio Mangabeira, president of the UDN, for the governorship of the state of Bahia (Jornal do Brasil, October 17, 1946). The deal was done. But Mariani and his party colleague Raul Fernandes only take up the portfolios of Education and Health and of Foreign Relations, respectively, in December. In between, negotiations with the UDN had their comings and goings. Despite all the practical reasons that led it to joining the government, the UDN felt obliged to keep up its political principles. In reference to the invitations of Clemente Mariani and Raul Fernandes, in mid-November the UDN president stated that:

The hypothesis of the invitation in itself does not include any commitment on the part of the UDN to cooperate with the government, since it is only a personal honour bestowed upon Mr. Fernandes and Mr. Mariani. However, out of party discipline and ethics, these two UDN militants will only be able to receive and eventually accept the invitation of Mr. President of the Republic after being authorized for such by the UDN. The consultations within and without the party sphere aim at establishing an average opinion regarding the matter (…) (Correio da Manhã, November 14, 1946, emphasis added).

The National Directorate of the UDN finally approved the participation of its members in the government, by 26 votes to 16, but not the party's participation, as explained by its president. For that, he states, it would be necessary for there to be a “previously agreed programme”; he would also need to know “in which ministries these policies would be implemented and, lastly, an environment that would guarantee the policies in every state without distinction is essential” (Correio da Manhã, December 4, 1946). With this decision, the UDN directorate sought to set out the difference between the party's participation, on the basis of the commitment to a programme, and the participation of party members. Anyway, the participation of the new ministers was approved by a decision-making body of the party, which would certainly reap the benefits. For this reason, the participation of UDN ministers in the Dutra government coalitions was considered a party-sanctioned one.

In fact, this ministerial reform was a turning point or, simply put, defined the Dutra government's line of action. Note that the PTB ceased to have a representative from among its ranks in the government. The Ministry of Labour went to Morvan Dias de Figueiredo, an important figure in the São Paulo business community and member of the leadership of FIESP (Federation of Industry of the State of São Paulo). Furthermore, a new party, the PR, was brought into the coalition.

When the cabinet was reshuffled in April 1950, the PR representatives left their posts in the Executive to stand in the elections but the UDN ministers were kept on. That month, the UDN put forward the presidential candidacy of Brigadier Eduardo Gomes. News reports forecast the exit of the UDN ministers. The latter formally consulted the party president, who authorized their permanence arguing that the party's presidential candidacy did not represent an act of hostility to the government. The president of the Republic agreed with the argument and asked the two ministers to stay (Jornal do Brasil, April 22, 1950). Therefore, the situation that led me to consider the UDN's participation as a formal one in the first coalition is not altered.

In the case of the Vargas government, the situation is different. Vargas invited the UDN formally, but the party placed itself clearly in opposition to the government from the start (Jornal do Brasil, January 23, 1951 and February 2, 1951). The participation of João Cleofas, from the UDN of Pernambuco state, had a regional character, for it counted on the support of both the party's state section and of the PSD's state section. At national level, the party remained in opposition to the government that came to a close with Vargas’ suicide. For these reasons, it is not possible to identify the UDN as a pro-government party during the Vargas government. For this same reason, the participation of UDN ministers in the Goulart government is considered individual. In the Café Filho government it is the PTB that withdraws from the coalition, given his approximation to the UDN. As vice-president, Café Filho had publicly split with Vargas at the height of the crisis, spearheaded by the UDN, which led the latter to commit suicide.⁷ The brief government of Nereu Ramos, who replaced Café Filho as a result of a preventive military coup that sought to ensure the respect for the results of the 1955 presidential election, brought the PTB-PSD alliance back to government.⁸

Let us now turn to the PMDB's participation in the post-1988 period. In the first three governments, the nature of this participation is uncontroversial. It was the main party in opposition to the military regime and the main player in the transition to civilian rule through an indirect election. In the Sarney government, the PMDB allied with the PFL, a dissident wing of the party that had supported the military, to form the first civilian government since 1964. This is the only government coalition made up of just two parties, with the president's party nearly having a majority in the lower house of Congress. In the Collor government, the first to be directly elected, PMDB member Bernardo Cabral took over at the Ministry of Justice, but the party not only did not endorse this participation but also led the opposition to the government in Congress.⁹ The PMDB re-joins the government in all the coalitions formed by Franco, a former party leader of importance who had left the party to be Collor's running mate.

However, the PMDB's participation in the FHC and Lula governments was the object of considerable internal strife, often transformed into court battles that most of the time did not come to a close. As the PMDB is a party of national dimension, with leaders of relevance, one of the questions that divided it was the intention of running in the presidential elections with a candidate of its own. This issue became the focus of frictions between its internal factions whenever it came to deciding on its participation in the government. In Cardoso's first term, former president and Minas Gerais governor Itamar Franco led the defence of a candidacy, putting his own name forward, but was defeated by the decision of the party convention, held in March 1998, to support Cardoso's re-election. At the end of his second term, the support for the government was once more challenged, but the party convention once again sealed the electoral alliance with the PSDB, and the PMDB participated with the vice-presidential candidacy. In the Lula government, the issue emerged much sooner. Furthermore, the party failed to reach a majority position as to its participation in the government. In December 2004, less than a year after entering the government, the PMDB national executive decided by 9 votes to 8 to withdraw its ministers from the government. The party was divided. On the one hand, the national leadership and governors defended the ministers’ departure, while on the other, the parliamentarians — 20 out of 22 senators and 58 out of 76 federal deputies, according to the press — defended their permanence in office.¹⁰ The decision of the executive was ratified by the party convention held immediately afterwards, but remained sub judice due to the appeals lodged by the party's pro-government groups. In spite of its victory, the party leadership was not able to implement the convention's decision. The ministers remained in office without being meted out the appropriate punishments — expulsion or public censure — being meted out. For this reason, the PMDB remained as a member of the government. This seems to be a case in which maintaining an ambiguity was of interest to the party, as it might be used as a bargaining chip in future negotiations. And this was what in fact happened. In the first coalition of Lula's second term, the party, this time united, increased significantly its representation in the ministry.

A few remarks are necessary regarding the inclusion or exclusion of other parties in the coalitions identified in chart 2 above. In the Lula government, the PSB was kept in all the coalitions, despite not having taken up ministerial positions during the last coalition, because the minister appointed to the Ministry of Science and Technology, controlled by the PSB since the beginning of the government, was from the same state of the two previous ministers and, although not affiliated, he had links to PSB in the state of Pernambuco, where he had held positions in PSB's administrations. The PL (currently called PR) neither held a ministerial position after the vice-president, José Alencar, left the Ministry of Defence, but as the party of the vice-president, it in fact never ceased to be part of the government. With the other parties, the strict criterion prevailed. For instance, the PTB was not considered a member of the government during the first coalition of FHC's second term, in spite of the fact that the party was part of the electoral alliances that elected and re-elected Cardoso; that it held ministerial portfolios during the whole of the first term; that it held posts in state companies or in the lower ranks of the administration during this period; and, lastly, that it only broke formally with the government to support its own candidate in the presidential election of 2002.

Lastly, let me make a remark on the Itamar Franco government. This government began after a protracted political crisis that culminated in the Collor's impeachment. As it included various parties that had contributed to this outcome, it could be seen as a government of national unity. However, two of these parties explicitly refused to take part. One was the PDT, whose decision led senator Maurício Correa to break with the party and its president, Leonel Brizola, a personal friend, upon being appointed to the Ministry of Justice (Interview with Maurício Correia, Correio Braziliense, January 1, 2005). The PT also did not ratify the appointment of former São Paulo mayor Luiza Erundina to the Ministry of the Administration, even though, contrary to its history, it did not compel her to leave the party. Therefore, these two parties were not included in the coalitions formed during this government.

Let us now compare some characteristics of the coalitions formed during the two periods under consideration. In the 1946-64 period, there were 19 government coalitions (excluding the brief interim presidency of Mazzili in September 1961), while in the current period there were 17. These coalitions had an average duration of 13 and 11.5 months, respectively. Over a third of the changes in government (a little higher in the first period, 37%) were owed to the regular electoral calendar, not resulting, therefore, from conflicts within the coalition. Legislative elections in Brazil also lead to changes of ministers in the preceding months, as ministers must leave office to run as candidates.

The average time spent in office by ministers varies considerably, as shown by the tables 3 and 4, even considering the varying duration of governments, as some were interrupted before the end of the president's term. In the 1946-64 period, the time spent in office by ministers is equivalent to the average duration of the coalitions. One point to be explored is whether ministers’ greater or lesser stability is related to the policy areas of their portfolios.

TABLE 3
TIME SPENT BY MINISTERS IN OFFICE, BY GOVERNMENT – 1946-1964

GOVERNMENT			TIME SPENT IN OFFICE* (DAYS)
	AVERAGE	NUMBER	STANDARD DEVIATION	MINIMUM	MAXIMUM
DUTRA	565	22	427,6	174	1546
VARGAS	495	19	290,1	81	854
CAFÉ FILHO	232	15	140,1	30	436
NEREU RAMOS	79	8	0,0	79	79
KUBITSCHEK	599	24	422,9	30	1800
QUADROS	203	10	0,0	203	203
GOULART	197	51	111,3	21	553
TOTAL (for the period)	352	149	323,4	21	1800

• ^*
  Excluding military ministries.

TABLE 4
TIME SPENT BY MINISTERS IN OFFICE, BY GOVERNMENT – 1988-2007

GOVERNMENT			TIME SPENT IN OFFICE* (DAYS)
	AVERAGE	NUMBER	STANDARD DEVIATION	MINIMUM	MAXIMUM
SARNEY	305	29	179,7	96	518
COLLOR	346	44	245,2	18	915
FRANCO	322	53	219,3	11	810
CARDOSO I	731	43	486,7	36	1440
CARDOSO II	549	69	441,2	54	1440
LULA I	696	65	423,4	120	1470
TOTAL (for the period)	514	303	404,4	11	1470

• ^*
  Excluding military ministries; includes Ministry of Defence since 1999.

Considering the number of coalitions formed, the less stable governments were those of Goulart and Lula, the former with seven coalitions, the latter with five. In spite of this, the average time spent in office by ministers during the Lula government is long, practically the same as in Cardoso's first term, the one with the highest average. In Cardoso's second term, ministers’ average rate of permanence drops considerably, even though his first coalition was the longest-lasting one of the two periods.

The charts 1 and 2 show that only the first president in each period was a member of a large party and that most coalitions involve a large number of parties. During the 1946-64 period, of the nineteen coalitions formed, only one was a minority coalition. In the current period, six of the seventeen coalitions are minority coalitions. Only in the Collor government were all the coalitions minority coalitions. The formation of minority coalitions during this government is largely owed to the president's anti-party strategy, even though he tried to broaden his party base with the ministerial reform of April 1992, at the start of the crisis that was to lead to his impeachment. The minority coalition at the end of Cardoso's second term was owed to the PFL's departure to support its own candidate in the presidential election. President Lula, for his part, sought to keep in his first coalition only the parties that had allied with his for the election, in other words, all the parties of the left plus the PL, a rightwing party, to which the vice-presidential candidate belonged. Until shortly before the new government took office, José Dirceu, PT president and new Head of the , tried to negotiate the entry of the PMDB into the government, but then the president decided to begin his government with a ministry made up only of the parties of the electoral alliance. With this decision, Lula avoided including in the government the party that had supplied the vice-presidential candidate to the ticket of his main opponent, the PSDB's José Serra. Actually, the president did not need to pay the costs of forming a majority government. Bearing in mind the legislative agenda that flowed from the commitments made during the electoral process, he knew he would have the support of the opposition, i.e., that he would have a legislative majority (Laver and Schofield, 1990; Strom, 1990).

One aspect not mentioned here but deserving of further exploration, is the ideological composition of the parties in the coalition, as well as the ideological position of the president, or of her/his party. These factors may have an important impact both on the durability and the performance of government coalitions.

Degree of Executive-Legislative integration and proportionality of the coalitions

For Cox and Morgenstern (2002, 258), the Latin American presidential systems, in their modal case, occupy an intermediate position between “pure parliamentarism” and “pure presidentialism”. What defines this position is the degree of integration of the Executive in the legislative process, which is determined by two factors: participation of the Executive in the legislative process and control over the assembly's agenda. Here, I deal with the former.

With regards to the integration of the Executive in the legislative process, one of the dimensions taken into consideration by these authors is the possibility of ministers simultaneously holding cabinet and legislative posts, or, as in Brazil, of members of the legislature holding cabinet posts without losing their legislative terms (ibid., 258-59). In the case of Brazil, ministers must take leave from their legislative posts, but may take them up again very easily in order to, for instance, take part in discussions, negotiations and votes on matters of importance to the government. In some cases, when the position of the minister's substitute is contrary to the government's, or even just doubtful, the minister may leave her/ his post for only a day, take part in the vote and take up again her/his ministerial post.

This institutional mechanism really does approximate the Brazilian presidential system more to the parliamentary system than to the more rigid U.S. system of separation of powers. However, it does not fully explain the degree of integration between the Executive and Legislative branches observed in the coalitions formed in Brazil. The pattern observed allows us to go further and suggest the following hypothesis: the logic presiding over the process of coalition formation in Brazil does not differ substantially from that of countries with parliamentary systems. Let us look at the data.

The first three columns of the tables below show that party coalitions predominate. In other words, in most governments, a large share of the ministers who take part do so as representatives of their parties. The percentage of ministers that held parliamentary terms (as federal deputy or senator) when they were invited to join the government is significant.¹¹ Furthermore, the vast majority of ministers had already held legislative office at federal level at least once before.

When one compares the two periods, what is noticeable is that the indicators of integration between the Executive and Legislative branches are higher in the 1946-64 period than in the current one. The low levels of partisanship in the Collor government are largely responsible for the rather lower average in the current period. Collor was the winner of Brazil's first direct presidential elections after the re-establishment of democracy. He had adopted a campaign strategy of presenting himself as a political outsider. This, coupled with the enormous unpopularity of the Sarney government and his party, the PMDB, ensured Collor's victory. In forming his cabinet, Collor made good on the campaign promise to keep political parties away from government. But the strategy was altered in the middle of the crisis during his third year in office, when he sought to form a party base. At that moment, however, this was no longer possible. Even so, in his last coalition one notices an increase in the three integration indicators.

In the previous period, even Quadros, with a strategy similar to Collor's and also with a minority government, formed a cabinet in which 60% of the ministers had party support. The lowest rate of partisanship for the 1946-64 period is that of the coalition formed by prime minister Hermes Lima (32%), the last one formed during the period when Goulart ruled under a parliamentary regime instituted during the crisis brought about by Quadros’ resignation. During this government, one also sees the lowest proportion of ministers holding federal legislative terms, just 9%. This may be explained by the fact that there already existed an agreement between the parties to hold the referendum that was to re-establish presidentialism. The degree of partisanship of the coalitions formed under the mixed regime (Goulart's first four) is lower than that of the coalitions formed under the presidential system, while the first two formed under the new regime have the highest participation of parliamentarians in the cabinet.

An institutional mechanism (that which compels ministers to step down six months before the elections) may contribute to a fall in the number of ministers with an ongoing parliamentary term and even to the degree of partisanship in the coalitions formed towards the end of presidential terms. This may be seen in various coalitions of the two periods. In the second FHC term, for example, because of the 2002 elections, a ministerial reform was carried out to replace all the ministers who would stand in the elections. This led to a significant fall in these indicators, including in relation to the coalitions previously formed by the same president.

A more detailed presentation of the positions held by ministers at the time of their appointment, by government, may be found in the set of tables in ^{Appendix 2}. A few points regarding the tables are, however, worth highlighting. The presidential choice tends to fall on people in political posts, whether they are elected officers, public sector bureaucratic appointees or party bureaucrats. An expressive majority of ministers held these positions at the moment of their appointment: 83% in the 1946-64 period and 79% in the period since 1988. In the 1946-64 period, 56% of these ministers held elected office, while in the post-1988 period, 40% were in elected positions. However, most politicians in appointed positions in the administration or in party positions had occupied elected positions previously. In both periods, less than 20% of ministers, therefore, were technical staffers, career bureaucrats, businesspeople or unionists.

Going back to tables 5 and 6, let us examine the last column, which shows the degree of proportionality between the number of legislative seats of the parties that belong to the government coalition and the number of ministerial portfolios held by these parties, including the percentage of independent ministers, that is, those who do not hold legislative seats. Amorim Neto (2002) put together and has made good use of this indicator in the study of government coalitions in presidential systems in Latin America and Brazil. This leads me to discuss his work and the advantages and limits of this indicator in a bit more detail.¹²

TABLE 5
DEGREE OF EXECUTIVE-LEGISLATIVE INTEGRATION AND PROPORTIONALITY IN GOVERNMENT COALITIONS – 1946-1964

GOVERNMENT COALITIONS	DEGREE OF PARTISANSHIPa	PARLIAMENTARY TERMb	LEGISLATIVE EXPERIENCEc	PROPORTIONALITYd
Dutra I	85,7	28,6	57,1	0,84
Dutra II	72,7	50,0	63,6	0,68
Dutra III	66,7	50,0	75,0	0,67
Vargas I	87,5	75,0	87,5	0,88
Vargas II	77,8	55,5	77,8	0,78
Vargas III	75,0	25,0	50,0	0,75
Café Filho	75,0	37,5	50,0	0,83
Café Filho	46,2	38,5	58,3	0,56
N. Ramos	87,5	37,5	50,0	0,84
JK I	100,0	40,0	73,3	0,92
JK II	100,0	41,2	58,8	0,92
Goulart I (Neves I)	70,0	90,0	90,0	0,64
Goulart II (Neves II)	63,6	81,8	81,8	0,58
Goulart III (B. Rocha)	60,0	22,2	40,0	0,45
Goulart IV (H. Lima)	45,5	9,1	27,3	0,32
Goulart V	83,4	58,3	83,3	0,68
Goulart VI	68,7	62,6	68,8	0,64
Goulart VII	72,7	63,6	63,6	0,73
TOTAL (for the period)	75,0	48,8	65,3	0,71

• ^a
  Degree of Partisanship: Percentage of seats held by ministers with support from their party.
• ^b
  Parliamentary Term: Percentage of ministers who were serving terms as Federal Deputy or Senator or had served in the Congress immediately before.
• ^c
  Legislative Experience: Percentage of ministers who served legislative terms in their careers, excluding the ones they were serving when appointed.
• ^d
  Proportionality: Relation between the percentage of ministers and of seats per party, including non-partisan ministers. (See formula in Amorim Neto, 2000, 485.)

TABLE 6
DEGREE OF EXECUTIVE-LEGISLATIVE INTEGRATION AND PROPORTIONALITY IN GOVERNMENT COALITIONS – 1988-2007

GOVERNMENT COALITIONS	DEGREE OF PARTISANSHIPa	PARLIAMENTARY TERMb	LEGISLATIVE EXPERIENCEc	PROPORTIONALITYd
SARNEY 2	77,3	36,4	54,5	0,77
COLLOR 1	18,8	15,4	18,8	0,19
COLLOR2	25,0	21,4	18,8	0,25
COLLOR 3	18,8	14,3	12,5	0,19
COLLOR 4	42,1	35,3	36,8	0,42
FRANCO 1	57,1	52,4	57,1	0,48
FRANCO 2	52,4	42,9	52,4	0,44
FRANCO 3	23,8	19,0	23,8	0,20
FHC I 1	66,7	30,0	28,6	0,56
FHC I 2	65,2	38,1	34,8	0,59
FHC II 1	69,0	33,3	41,4	0,64
FHC II 2	46,2	20,0	28,0	0,46
LULA 1	87,1	41,9	54,8	0,59
LULA 2	87,1	45,2	54,8	0,53
LULA 3	83,9	45,2	54,8	0,54
LULA 4	83,9	46,7	58,1	0,53
LULA 5	80,0	35,7	36,7	0,53
TOTAL (FOR THE PERIOD)	62,7	35,7	41,6	0,49

• ^a
  Degree of Partisanship: Percentage of seats held by ministers with support from their party.
• ^b
  Parliamentary Term: Percentage of ministers who were serving terms as Federal Deputy or Senator or had served in the Congress immediately before.
• ^c
  Legislative Experience: Percentage of ministers who served legislative terms in their careers, excluding the ones they were serving when appointed.
• ^d
  Proportionality: Relation between the percentage of ministers and of seats per party, including non-partisan ministers. (See formula in Amorim Neto, 2000, 485.)

According to Amorim Neto, “(…) proportionality in cabinet shares is the equilibrium solution for the bargaining problems faced by presidents and parties regarding the division of the executive pie”; this being the case, he concludes: “the more cabinet shares deviate from proportionality, the less coalitional is the cabinet” (ibid., 52). The rationale that leads the author to this conclusion is the following. In coalition governments under presidentialism, it is not possible to identify a condition that Laver and Schofield (1990: 130) establish also for government coalitions under parliamentarism, i.e., the existence of a binding agreement. Under presidentialism, the agreements for the formation of coalitions do not work the same way as under parliamentarism, since there can be appointments of politicians not endorsed by their parties. Furthermore, given the secrecy that surrounds political negotiations, it would be difficult to observe the agreements reached. This being so, it would be necessary to make a distinction between coalition cabinet and coalition governance (Amorim Neto, 2002: 51-2). In the eyes of this author, the proportionality between ministries and parliamentary seats characterizes the first type: “A coalition cabinet is defined as featuring at least two parties and a high degree of proportionality between the parties’ cabinet shares and legislative weights” (ibid., 50).

However, the degree of proportionality may be a very limited indicator to define government type. This is so, firstly, because the proportion of independent ministers, an important component in the proportionality formula, is not peculiar to the presidential system. Even though under parliamentarism the correlation between participation in the ministry and legislative seats is very high, there certainly exist variations. For example, in Norway between 1945 and 1978, 52% of all cabinet members had no prior parliamentary experience and cabinet members cannot serve as representative as they hold cabinet office; in that country in ministries with well-organized clienteles, “it is much more desirable for cabinet members to have good interest-grou ties than to have parliamentary experience” (Strom, 1994, 43). In France, according to Thiébault, the participation of civil servants in ministerial posts, 35 percent between 1945 to 1985, “has led to the suggestion that French governments are dominated by a “mandarinate” in which the grandes ecoles (such as the Ecole nationale d'administration and the Ecole polytechnique) play a large part” (ibid., 143).

Secondly, it is not possible to define what the point would be from which one could identify a coalition government. Amorim Neto warns us against a “loose application of the concept of coalition government to presidential systems” (2002, 48) and concludes that in Brazil only the Cardoso government was actually a coalition government (Amorim Neto, Cox and McCubbins, 2003). However, the proportionality rates of the second coalition of the FHC government — the longest-lasting one and the one that maintains the same party makeup throughout most of his second term — and of the first coalition of the Lula government are identical: 0.59. What explains this is the fact that in FHC's coalition the number of non-partisan ministers is high, while in Lula's, there is a disproportion of portfolios allocated to his party, a difference of 30% in relation to its legislative representation. However, the Lula government has a higher degree of partisanship, more ministers holding parliamentary terms and with parliamentary experience. What might be considered the most important factor in defining a coalition government?

As for the difficulties in identifying coalitions, it is worth remembering that not all coalition governments in the parliamentary system are formed on the basis of formal coalition agreements. An analysis of 238 coalition governments shows that only 63% were based on some kind of “identifiable coalition agreement”, and that the agreements reached vary as to how public and how general they are (Muller and Strom, 2000, 573-75). Lastly, it is worth noting that the condition imposed by Laver and Schofield, to which Amorim makes reference, aims to distinguish between legislative coalitions and executive coalitions. This distinction is made very precisely by the authors and is worth quoting:

A legislative ‘coalition’ is no more than a group of legislators who vote together on a particular issue. This concept of coalition carries no connotation of permanence, of institutional status, or of any executive role whatsoever. A government coalition carries a very heavy connotation of stability, of agreement over a wide range of issues, of formal institutional status that is only occasionally tested in the legislature, and of executive control, via cabinet portfolios, over all key policy areas (Laver and Schofield, 1990, 129).

The above definition of government coalition, like the one of legislative coalition, seems to me to be applicable also to presidential governments. For there to be a coalition government, all it takes is for the party forming the government to offer ministerial posts to other parties, which, by party decision rather than individual decisions, are accepted. Consequently, these parties join the Executive. Legislative support is expected of these parties, which leads one also to talk about minority coalition governments. Whether the latter are successful — indeed, whether single-party minority governments are successful — depends on the formation of legislative coalitions, which, in turn, depends largely on factors not related to the executive coalition. Hence the importance of the distinction, as emphasized by Laver and Schofield (ibid., 129-131). But for this to be possible, it is necessary to identify parties’ commitment to the governments they are part of, even if the risk of subjectivity is greater.

The comparison between the two periods strengthens the argument that the proportionality of cabinets cannot be considered a defining characteristic of coalition governments, even if it is an important indicator of the presidential strategy that might explain both the government's performance and parliamentary behaviour. The “co-option strategy”, as opposed to what Amorim Neto calls coalition strategy (2002, 49), might be demonstrated considering just the degree of partisanship of two governments, for example, 15% in the Collor government and 100% in Kubitschek's. But the rate of proportionality has the advantage of synthesizing in a single indicator, as well as the share of ministers called the president's “quota”, the distribution of portfolios according to the weight of parties in the legislature. Even though it is related to the former — and maybe more in the extreme cases, such as in the Collor and Kubitschek governments, with rates of proportionality of 0.19 and 0.92, respectively —, it supplies us with important additional information. For example, the first two coalitions of the Lula government have exactly the same degree of partisanship (87.1%), but the cabinet's rate of proportionality falls from 0.59 to 0.53. This decrease represents a less equitable distribution of portfolios in the second coalition, despite its enlargement with the incorporation of the PMDB.

This being the case, the rate of proportionality should be used only as a descriptive factor of the type of government and, thus, as something that explains relevant results or something that needs to be explained. However, it must not be used as a defining factor of a type of government or to differentiate the logic of coalition formation of different systems of government.

The regional dimension in the formation of coalitions

Both political journalism and academic studies have considered the regional dimension an important component of the president's choice of ministers (Abranches, 1988, 25). The tables 7 to 10 show the distribution of ministerial portfolios in the two periods broken down into the country's five regions and the states of the federation.

TABLE 7
REGIONAL COMPOSITION OF MINISTRIES, BY GOVERNMENT COALITION – 1946-1964

GOVERNMENTS	REGIONS (in %)
	CENTRE-WEST	NORTHEAST	NORTH	SOUTHEAST	SOUTH	TOTAL
Dutra I	…	14,3	…	71,4	14,3	100,0 (7)
Dutra II	…	9,1	…	72,7	18,2	100,0 (11)
Dutra III	…	25,0	…	75,0	…	100,0 (8)
Vargas I	…	25,0	…	62,5	12,5	100,0 (8)
Vargas II	…	37,5	…	37,5	25,0	100,0 (8)
Vargas III	…	42,9	…	42,9	14,3	100,0 (7)
Café Filho	…	12,5	…	75,0	12,5	100,0 (8)
Café Filho	…	15,4	…	69,2	15,4	100,0 (13)
N. Ramos	…	42,9	…	57,1	…	100,0 (7)
JK I	…	13,3	…	66,7	20,0	100,0 (15)
JK II	…	23,5	…	70,6	5,9	100,0 (17)
Quadros	…	30,0	10,0	50,0	10,0	100,0 (10)
Goulart I (Neves I)	…	40,0	…	60,0	…	100,0 (10)
Goulart II (Neves II)	9,1	45,5	…	45,5	…	100,0 (11)
Goulart III (B. Rocha)	…	44,4	…	44,4	11,1	100,0 (9)
Goulart IV (H. Lima)	…	36,4	…	54,5	9,1	100,0 (11)
Goulart V	…	33,3	8,3	58,3	…	100,0 (12)
Goulart VI	6,7	26,7	…	53,3	13,3	100,0 (15)
Goulart VII	10,0	50,0	…	10,0	30,0	100,0 (10)
TOTAL (FOR THE PERIOD) a	1,5 (3)	28,9 (57)	1,0 (2)	57,4 (113)	11,2 (22)	100,0 (197)

• ^a
  Refers to each appointment of a new minister and not only the first government coalition.

TABLE 8
REGIONAL COMPOSITION OF MINISTRIES, BY GOVERNMENT COALITION – 1988-2007

GOVERNMENTS	REGIONS (in %)
	CENTRE-WEST	NORTHEAST	NORTH	SOUTHEAST	SOUTH	TOTAL FOR THE PERIOD
SARNEY 2	4,0	36,0	4,0	44,0	12,0	100,0 (25)
COLLOR 1	20,0	20,0	10,0	30,0	20,0	100,0 (10)
COLLOR2	20,0	20,0	10,0	30,0	20,0	100,0 (10)
COLLOR 3	5,9	17,6	5,9	52,9	17,6	100,0 (17)
COLLOR 4	6,3	25,0	…	50,0	18,8	100,0 (16)
FRANCO 1	7,4	14,8	3,7	63,0	11,1	100,0 (27)
FRANCO 2	5,6	16,7	5,6	61,1	11,1	100,0 (18)
FRANCO 3	11,1	16,7	…	66,7	5,6	100,0 (18)
FHC I 1	…	16,7	…	61,1	22,2	100,0 (18)
FHC I 2	2,9	20,6	…	55,9	20,6	100,0 (34)
FHC II 1	14,9	19,1	2,1	55,3	8,5	100,0 (47)
FHC II 2	8,7	13,0	8,7	65,2	4,3	100,0 (23)
LULA 1	6,9	17,2	3,4	55,2	17,2	100,0 (29)
LULA 2	3,1	18,8	9,4	53,1	15,6	100,0 (32)
LULA 3	3,1	21,9	9,4	46,9	18,8	100,0 (32)
LULA 4	3,0	24,2	6,1	45,5	21,2	100,0 (33)
LULA 5	5,4	24,3	5,4	45,9	18,9	100,0 (37)
TOTALa	6,8	20,4	4,7	52,8	15,3	100,00
(FOR THE PERIOD)	(29)	(97)	(20)	(225)	(65)	(426)

• ^a
  Refers to each appointment of a new minister and not only the first government coalition.

TABLE 9
STATE OF ORIGIN OF MINISTERS, BY GOVERNMENT – 1946-1964

PERCENTAGE OF MINISTERS
STATE	DUTRA	VARGAS	CAFÉ FILHO	NEREU RAMOS	KUBITSCHECK	QUADROS	GOULART	TOTAL
AM	…	…	…	…	…	…	1,3	0,5 (1)
BA	11,5	13,0	…	14,3	…	10,0	15,4	10,2 (20)
CE	…	…	…	14,3	9,4	…	5,1	4,1 (8)
DF	…	4,3	4,8	…	…	…	…	1,0 (2)
ES	…	…	…	…	3,1	…	…	0,5 (1)
GB	…	…	…	…	…	10,0	5,1	2,5 (5)
GO	…	…	…	…	…	…	1,3	0,5 (1)
MG	15,4	13,0	4,8	14,3	28,1	10,0	19,2	17,3 (34)
MT	…	…	…	…	…	…	2,6	1,0 (2)
PA	…	…	…	…	…	10,0	…	0,5 (1)
PB	…	8,7	…	…	6,3	10,0	2,6	3,6 (7)
PE	7,7	13,0	9,5	…	3,1	10,0	14,1	10,2 (20)
PR	…	…	14,3	…	…	…	3,8	3,0 (6)
RJ	23,1	4,3	38,1	…	9,4	10,0	11,5	14,2 (28)
RN	…	…	4,8	14,3	…	…	1,3	1,5 (3)
RS	11,5	17,4	…	…	9,4	10,0	5,1	7,6 (15)
SC	…	…	…	…	3,1	…	…	0,5 (1)
SP	30,8	26,1	23,8	42,9	28,1	20,0	11,5	21,3 (42)
TOTALa (for the period)	100,0	100,0	100,0	100,0	100,0	100,0	100,0	100,0
	(26)	(23)	(21)	(7)	(32)	(10)	(78)	(197)

• ^a
  Refers to each appointment of a new minister and not only the first government coalition.

TABLE 10
STATE OF ORIGIN OF MINISTERS, BY GOVERNMENT – 1988-2007

PERCENTAGE OF MINISTERS
STATE	SARNEY	COLLOR	FRANCO	CARDOSO I	CARDOSO II	LULA I	TOTAL
AC	…	…	…	…	…	3.1	1.2 (5)
AL	…	6.0	…	1.9	2.9	1.8	2.1 (9)
AM	…	2.0	…	…	4.3	2.5	1.9 (8)
BA	12.0	4.0	3.2	5.8	2.9	6.1	5.2 (22)
CE	8.0	6.0	1.6	3.8	…	5.5	4.0 (17)
DF	…	…	4.8	…	7.1	4.3	3.5 (15)
ES	…	…	1.6	…	2.9	…	0.7 (3)
GO	4.0	12.0	1.6	1.9	4.3	…	2.8 (12)
MA	4.0	…	4.8	…	1.4	1.2	1.7 (7)
MG	28.0	…	42.7	15.4	7.1	20.2	18.9 (80)
MS	…	…	1.6	…	1.4	…	0.5 (2)
PA	4.0	4.0	3.2	…	…	…	1.2 (5)
PB	…	…	…	…	2.9	…	0.4 (2)
PE	…	…	1.6	5.8	4.3	6.7	4.3 (18)
PI	4.0	…	1.6	1.9	1.4	…	0.9 (4)
PR	4.0	10.0	3.2	5.8	1.4	1.8	3.5 (15)
RJ	…	14.0	1.6	5.8	4.3	3.7	4.7 (20)
RN	4.0	…	1.6	…	1.4	…	0.7 (3)
RO	…	…	…	…	…	1.2	0.5 (2)
RS	8.0	8.0	6.3	15.4	7.1	12.9	10.4 (44)
SC	…	2.0	…	…	…	3.7	1.7 (7)
SE	4.0	…	3.2	…	…	…	0.7 (3)
SP	16.0	32.0	15.8	36.5	42.9	25.3	28.4 (120)
TOTALa	100,00	100,00	100,00	100,00	100,00	100,00	100,00
(for the period)	(25)	(50)	(63)	(52)	(70)	(163)	(423)

• ^a
  Refers to each appointment of a new minister and not only the first government coalition.

The main point to underline regarding this set of tables is the similarity of the pattern observed in the two periods. Firstly, there is a major concentration of ministers from the Southeast region, more precisely form the states of São Paulo, Minas Gerais and Rio de Janeiro. (The participation of Espírito Santo state is minimal.) In the period since 1988, the participation of Rio de Janeiro falls sharply, while São Paulo's increases. The Northeast appears in the two periods in second place, well behind the Southeast, but in the 1946-64 period with almost three times the representation of the South region, a difference that has become much smaller in the current period: 19.9% vs. 15.4% of the total. In the two periods, therefore, the cabinet representation of states of the North and Centre-West regions is very low.

The third aspect to be highlighted is that the predominance of the Southeast in the cabinet makeup does not undergo significant alterations in the various governments of the two periods. In the first period, the Southeast's participation only goes below 50% in three coalitions of the Goulart government, but even then it remains around 45%, with the exception of the last coalition, when there is a complete inversion: the percentage of ministers from the Southeast falls to 10%, and of those from the Northeast rises to 50%. In the current period, the Southeast's participation gets reduced to a third in the first two coalitions of the Collor government, but surpasses 50% in the other two. Lastly, in two coalitions of the Lula government the figure stays below, but very close to 50%.

This pattern of regional distribution of ministerial portfolios does not seem to me to conform well to the importance attributed to the regional dimension in the formation of government coalitions in Brazil. It is obvious that in a country of continental dimensions and with a federal structure, where some parties have clear regional features, regional criteria must be taken into consideration. It is not about demanding parity between the different regions in the division of ministerial portfolios, such as in a country like Belgium, an extreme example of two-dimensional cabinet formation, where the criterion of parity in regional distribution was transformed into a constitutional norm (Müller and Strom, 2000, 333). However, if in fact the regional dimension were to determine the distribution of ministries in Brazil, one would expect it to be less skewed in favour of the Southeast and Northeast, where around 70% of the electorate reside and the larger parties have their greatest political influence. This is all the more surprising if one considers the fact that the North and Centre-West regions are over-represented in the legislature. In principle, their weight in number of seats would place them in a privileged position to demand greater representation in government. To give an extreme example in another direction: in a country where political bargaining could hardly be called two-dimensional, the United Kingdom, even so, the regional issue is taken into account in the distribution of ministerial portfolios (Norton, 1991, 201). These considerations indicate that a deeper analysis of the two-dimensional nature of political bargaining remains necessary for one to be able to affirm the importance of the regional issue in the formation of government coalitions in Brazil.

Notes

• 1
  See Cheibub, Przeworski and Saiegh (2004) and Cheibub (2007)
• 2
  See particularly Laver and Shepsle (1994) and Mueller and Strom (2000)
• 3
  Furthermore, there are parliamentary systems, such as Norway's, in which the parliamentary term is fixed, i.e., independent from the Executive
• 4
  Amorim Neto (2000) considers one out of the three military ministers before the creation of the Ministry of Defence in all the calculations referring to independent ministers
• 5
  Interim ministers were considered only for the 1988-2007 period in the following situations: when they remained in office for more than one month and the minister had left the position definitively. There were only 18 interim ministers in the whole period.
• 6
  The use of the expression may be owed to the importance accorded by newspapers at the time to international news, which in general made the front-page headlines. News items of cabinet changes in European countries were frequent
• 7
  The PTB forbade its members from participating in the new administration. The minister appointed for the Labour portfolio, Alencastro Guimarães, became a party dissident (Skidmore, 1988, 182)
• 8
  See details in Skidmore (1988, 185-198)
• 9
  For the PMDB's actions in relation to the Collor Plan, see Figueiredo and Limongi (1999, 184-191)
• 10
  O Estado de São Paulo, 12-9-2004
• 11
  See Tables 1 and 2, in ^{Appendix 2}, for the low percentages of state deputies and city/town councillors
• 12
  Amorim Neto initially called this indicator Index of Cabinet-party Congruence (1998, 57), later renaming it Degree of Coalescence (2000, 2002, 2006).
• *
  I thank José Antônio Cheibub who helped me with many suggestions and much advice; Marcus Figueiredo and Fernando Limongi for their last-minute reading and comments; Sérgio Abranches for his reading of a previous version and the information on some ministers of the 1946-64 period; Daniel Heller and Andrezza Ribeiro for their participation in gathering and organizing data. The research that gave rise to this text had the support of CNPq, Fapesp and Faperj.
• *
  PDS was the success or of ARENA, the government party during military rule; PPR was created in 1993, incorporating the PDC – Partido Democrata Cristão; in 1995, PPR changes the name to PPB and in 2005 to PP.
• **
  In 2005, PL is incorporated to the newly created PRB - Partido Republicano Brasileiro.
• ***
  Created in 1993 from the union of PTR - Partido Trabalhista Renovador - and PST - Partido Social Trabalhista; in 1995 is incorporated to PPR which changes the name to PPB.

Appendix Data sources

Websites

PRESIDêNCIA DA REPÚBLICA. Brasil. www.presidencia.gov.br/info_historicas/galeria_pres/.

CâMARA DOS DEPUTADOS. Brasil. www.camara.gov.br.

SENADO FEDERAL. Brasil. www.senado.gov.br.

TRIBUNAL SUPERIOR ELEITORAL (TSE). Brasil. www.tse.gov.br.

WIKIPEDIA. www.wikipedia.org.

GOOGLE. www.google.com.br.

Official Publications

Brasil. Congresso. Câmara dos Deputados. 1981. Deputados brasileiros: Repertório biográfico dos Senhores Deputados, abrangendo o período de 1946-1967. Org. David V. Fleischer. Brasília: Câmara dos Deputados.

Brasil. Congresso. Câmara dos Deputados. 1983. Deputados Brasileiros: 47ᵃ. Legislatura, 1983-1987: Repertório Biográfico. Brasília: Centro de Documentação e Informação.

Brasil. Congresso. Câmara dos Deputados. 1989. Deputados Brasileiros: Constituinte, 1987-1988: Repertório Biográfico. Brasília: Centro de Documentação e Informação.

Brasil. Congresso. Câmara dos Deputados. 1995. Deputados Brasileiros: 50ᵃ. Legislatura, 1995-1999: Repertório Biográfico. Brasília: Centro de Documentação e Informação.

Brasil. Congresso. Câmara dos Deputados. 1999. Deputados Brasileiros: 51ᵃ.. Legislatura, 1999-2003: Repertório Biográfico. Brasília: Centro de Documentação e Informação.

Alzira Alves Abreu et al., eds. 2001. Dicionário Histórico-Biográfico Brasileiro Pós-1930. 5 vols. Rio de Janeiro: Editora FGV / CPDOC.

Brasil. Tribunal Superior Eleitoral. 1964. Dados Estatísticos Eleitorais TSE: Eleições Federal, Estadual e Municipal realizadas no Brasil a partir de 1945. Rio de Janeiro: Departamento de Imprensa Nacional.

Brasil. Tribunal Superior Eleitoral. 1955. Dados Estatísticos TSE: Eleições Federais, Estaduais realizadas no Brasil em 1952, 1954 e 1955 e em confronto com os anteriores, vol. 3, parte I. Rio de Janeiro: Departamento de Imprensa Nacional.

Brasil. Tribunal Superior Eleitoral. 1966. Dados Estatísticos TSE: Eleições Federais, Estaduais realizadas no Brasil em 1958 e em confronto com anteriores, referendum 6.163, vol. 4. Rio de Janeiro: Departamento de Imprensa Nacional.

Brasil. Tribunal Superior Eleitoral. 1963. Dados Estatísticos TSE: Eleições Federais, Estaduais realizadas no Brasil em 1960 e em confronto com anteriores, vol. 5. Rio de Janeiro: Departamento de Imprensa Nacional.

Brasil. Tribunal Superior Eleitoral. 1964. Dados Estatísticos TSE: Eleições Federais, Estaduais realizadas no Brasil em 1962 e em confronto com anteriores, referendum 6.1.63, vol. 6. Rio de Janeiro: Departamento de Imprensa Nacional.

Brasil. Tribunal Superior Eleitoral. 1973. Dados Estatísticos TSE: Eleições Federais e Estaduais – Quadros Comparativos dos Pleitos entre 1945 e 1963. Rio de Janeiro: Departamento de Imprensa Nacional.

Newspapers

O Correio da Manhã. Selected issues from January 1946 to March 1964.

Jornal do Brasil. Selected issues from January 1946 to March 1964.

O Globo. Selected issues from October 1988 to January 2007.

Folha de São Paulo. Selected issues from October 1988 to January 2007.

Secondary Sources

Hippolito, Lucia. 1985. De raposas e reformistas: o PSD e a experiência democrática brasileira (1945-1964). Rio de Janeiro: Paz e Terra.

Meneguello, Rachel. 1998. Partidos e governos no Brasil contemporâneo (1985-1997). São Paulo: Paz e Terra.

Nicolau, Jairo, ed. 1998. Dados Eleitorais do Brasil (1982-1996). Rio de Janeiro: Revan / IuperjUCM.

Santos, Wanderley G., ed. 1990. Que Brasil é este?: Manual de indicadores políticos e sociais. Rio de Janeiro: Vértice.

NOTE: The basic source for the data set for the 1946-64 period was the Appendix of Hippolito (1985), in which the author lists all the ministries with respective ministers and their party affiliations. Some changes, however, were made when new information was found in official sources. The basic source for the data set for the 1988-2007 period was the information collected in the presidency's website, but the Appendix presented by Meneguello (1998) was also a valuable source of information.

The data set on both periods is available from the author upon request.

APPENDIX 2

TABLE 1
POSITION HELD AT THE TIME OF APPOINTMENT, BY GOVERNMENT: 1946 - 1964

GOVERNMENT
POSITION	DUTRA	VARGAS	CAFÉ FILHO	NEREU RAMOS	KUBITSCHEK	QUADROS	GOULART	TOTAL (for the period)
SENATOR	4,0	8,0	14,3	…	6,3	10,0	6,3	7,0 (14)
FEDERAL DEPUTY	40,0	44,0	23,8	37,5	34,4	40	49,9	41,8 (84)
STATE DEPUTY	…	…	…	…	…	…	…	…
CITY / TOWN COUNCILLOR	…	…	…	…	…	…	…	…
EXECUTIVE	8,0	8,0	4,8	…	18,8	20,0	1,3	7,0 (14)
APPOINTED FEDERAL BUREAUCRAT	8,0	24,0	14,3	37,5	37,4	10,0	23,7	22,9 (46)
CAREER FEDERAL BUREAUCRAT	…	8,0	…	…	…	…	5,0	3,0 (6)
BUREAUCRAT (STATES OR MUNICIPALITIES)	4,0	4,0	…	…	…	…	…	1,0 (2)
TECHNICAL STAFFER	20,0	4,0	19,0	25,0	3,1	10,0	12,5	11,9 (24)
BUSINESSPERSON OR UNION LEADER	…	…	4,8	…	…	…	…	0,5 (1)
PARTY LEADER OR MILITANT	16,0	…	19,0	…	…	10,0	1,3	5,0 (10)
TOTAL*	100	100	100	100	100	100	100	100
(for the period)	(25)	(25)	(21)	(8)	(32)	(10)	(80)	(201)

TABLE 2
POSITION HELD AT THE TIME OF APPOINTMENT, BY GOVERNMENT : 1988-2007

GOVERNMENT
POST	SARNEY	COLLOR	FRANCO	CARDOSO I	CARDOSO II	LULA	TOTAL
SENATOR	3,4	10,8	15,4	15,0	9,1	10,8	11,1 (32)
FEDERAL DEPUTY	31,0	21,7	15,4	25,0	22,8	30,7	24,2 (70)
STATE DEPUTY	…	…	…	2,5	1,5	…	0,7 (2)
CITY / TOWN COUNCILLOR	…	…	…	…	…	1,5	0,3 (1)
EXECUTIVE	10,3	…	3,8	…	…	7,7	3,5 (10)
APPOINTED FEDERAL BUREAUCRAT	24,1	16,2	27,0	22,5	24,2	10,8	20,4 (59)
CAREER FEDERAL BUREAUCRAT	3,4	16,2	17,3	2,5	7,6	3,1	8,3 (24)
BUREAUCRAT (STATES OR MUNICIPALITIES)	…	5,4	5,8	2,5	1,5	3,1	3,1 (9)
TECHNICAL STAFFER	13,9	18,9	11,5	10,0	7,6	1,5	9,3 (27)
BUSINESSPERSON OR UNION LEADER	…	10,8	1,9	…	1,5	4,6	3,1 (9)
PARTY LEADER OR MILITANT	13,9	…	1,9	20,0	24,2	26,2	15,9 (46)
TOTAL* (for the period)	100 (29)	100 (37)	100 (52)	100 (40)	100 (66)	100 (65)	100 (289)

• ^*
  Refers only to appointments in each of the new government coalitions (excluding military ministries; includes Ministry of Defence since 1999).

TABLE 3
TYPE OF POSITION HELD AT THE TIME OF APPOINTMENT, BY GOVERNMENT: 1946-1964

GOVERNMENT	TYPE OF POST HELD UPON APPOINTMENT
	ELECTED	APPOINTED OR PARTY BUREAUCRACY	CAREER BUREAUCRACY OR TECHNICAL	COMPANY/UNION	TOTAL MINISTERS*
DUTRA	69,3	3,8	19,2	7,7	100,0 (26)
VARGAS	72,0	16,0	12,0	…	100,0 (25)
CAFÉ FILHO	60,0	15,0	25,0	…	100,0 (20)
NEREU RAMOS	75,0	12,5	12,5	…	100,0 (8)
KUBITSCHEK	71,9	25,0		3,1	100,0 (32)
QUADROS	80,0	…	10,0	10,0	100,0 (10)
GOULART	67,9	12,3	19,8	…	100,0 (81)
TOTAL*	69,3	13,4	15,3	2,0	100,00
(for the period)	(140)	(27)	(31)	(4)	(202)

• ^*
  Refers only to appointments in each of the new government coalitions (excluding military ministries).

TABLE 4
TYPE OF POSITION HELD AT THE TIME OF APPOINTMENT, BY GOVERNMENT: 1988-2007

GOVERNMENT	TYPE OF POST HELD UPON APPOINTMENT
	ELECTED	APPOINTED OR PARTY BUREAUCRACY	CAREER BUREAUCRACY OR TECHNICAL	COMPANY/UNION	TOTAL MINISTERS*
SARNEY	44,9	37,9	17,2	…	100,0 (29)
COLLOR	32,4	21,6	35,2	10,8	100,0 (37)
FRANCO	34,6	34,6	28,9	1,9	100,0 (52)
CARDOSO I	42,5	45,0	12,5	…	100,0 (40)
CARDOSO II	33,3	50,0	15,2	1,5	100,0 (66)
LULA I	50,8	40,0	4,6	4,6	100,0 (65)
TOTAL (for the period)	39,8	39,4	17,6	3,1	100,0 (289)

• ^*
  Refers only to appointments in each of the new government coalitions (excluding military ministries; includes Ministry of Defence since 1999).

Political parties	Period: 1946-1964
PDC		Partido Democrata Cristão
PR		Partido Republicano
PSB		Partido Socialista Brasileiro
PSD		Partido Social Democrático
PSP		Partido Social Progressista
PTB		Partido Trabalhista Brasileiro
UDN		União Democrática Nacional

Period: 1988-2007
• PCBdoB  Partido Comunista do Brasil
• ^*PDS/PPR  Partido Democrático Social/ Partido Progressista Reformador
• PPB/PP  Partido Progressista Brasileiro /Partido Progressista
• PDT  Partido Democrático Trabalhista
• ^**PL  Partido Liberal
• PMDB  Partido do Movimento Democrático Brasileiro
• ^***PP  Partido Progressista
• PPS  Partido Popular Socialista
• PRN  Partido da Reconstrução Nacional
• PSB  Partido Socialista Brasileiro
• PSDB  Partido da Social Democracia Brasileira
• PT  Partido dos Trabalhadores
• PTB  Partido Trabalhista Brasileiro
• PV  Partido Verde

STATES
• AC  Acre
• AL  Alagoas
• AM  Amazonas
• AP  Amapá
• BA  Bahia
• CE  Ceará
• DF  Distrito Federal
• ES  Espírito Santo
• GO  Goiás
• MA  Maranhão
• MG  Minas Gerais
• MS  Mato Grosso do Sul
• MT  Mato Grosso
• PA  Pará
• PB  Paraíba
• PE  Pernambuco
• PI  Piauí
• PR  Paraná
• RJ  Rio de Janeiro
• RN  Rio Grande do Norte
• RO  Rondônia
• RR  Roraima
• RS  Rio Grande do Sul
• SC  Santa Catarina
• SE  Sergipe
• SP  São Paulo
• TO  Tocantins

Bibliography

• Abranches, Sérgio Henrique. 1988. Presidencialismo de coalizão: o dilema institucional brasileiro. Dados – Revista de Ciências Sociais 31 (1): 5-34.
• Amorim Neto, Octavio. 1998. Of presidents, parties and ministers: cabinet formation and legislative decision-making under separation of power. Ph.D. diss., University of California, San Diego.
• Amorim Neto, Octavio. 2000. Gabinetes presidenciais, ciclos eleitorais e disciplina legislativa no Brasil. Dados – Revista de Ciências Sociais 43 (3): 479-519.
• Amorim Neto, Octavio. 2002. “Presidential cabinets, electoral cycles, and coalition discipline in Brazil”. In Legislative politics in Latin America, ed. Scott Morgenstern and Benito Nacif, 446- 468. Cambridge: Cambridge University Press.
• Amorim Neto, Octavio. 2006. Presidencialismo e governabilidade nas Américas Rio de Janeiro: Editora FGV / Konrad Adenauer Stiftung.
• Amorim Neto, Octavio, Gary Cox and Mathew Mccubbins. 2003. Agenda power in Brazil's Câmara dos Deputados, 1989-98. World Politics 55 (4): 550-578.
• Cheibub, José Antônio, Adam Przeworski and Sebastian Saiegh. 2004. Government coalitions and legislative success under parliamentarism and presidentialism. British Journal of Political Science 34:565-87.
• Cheibub, José Antônio, Adam Przeworski and Sebastian Saiegh. 2007. Presidentialism, parliamentarism, and democracy Cambridge: Cambridge University Press.
• Cox, Gary and Scott Morgenstern. 2002. “Epilogue”: Latin America's reactive assemblies and proactive presidents. In Legislative Politics in Latin America, comps. Scott Morgenstern and Benito Nacif, 446-468. Cambridge: Cambridge University Press.
• Figueiredo, Argelina and Fernando Limongi. 1999. Executivo e legislativo na nova ordem constitucional Rio de Janeiro: FGV.
• Laver, Michael and Norman Schofield. 1990. Multiparty government: The politics of coalition in Europe Ann Arbor: The University of Michigan Press.
• Laver, Michael and Norman Schofield, and Kenneth A. Shepsle. 1994. Cabinet ministers and parliamentary government, Cambridge, Cambridge University Press.
• Müeller, Wofgang and Kaare Strom, eds. 2000. Coalition governments in Western Europe London: Oxford University Press.
• Norton, Philip. 1991. The British polity Longman: New York.
• Skidmore, Thomas. 1988. Brasil: de Getúlio Vargas a Castelo Branco (1930-1964) Rio de Janeiro, Paz e Terra.
• Strom, Kaare. 1994. “The Political Role of Norwegian Cabinet Ministers”. In Cabinet Ministers and Parliamentary Government, ed. Michael Laver and and Kenneth A. Shepsle, 35-55. Cambridge: Cambridge University Press.
• Thiébault, Jean-Louis. 1994. “The Political Autonomy of Cabinet Ministers in the French Fifth Republic.” In Cabinet ministers and parliamentary government, ed. Michael Laver and Kenneth A. Shepsle, 139-149. Cambridge: Cambridge University Press.

Publication Dates

• Publication in this collection
  15 Sept 2023
• Date of issue
  2007

History

• Received
  Sept 2007
• Accepted
  Dec 2007