Jump to ContentJump to Main Navigation
Legislative Institutions and Lawmaking in Latin America$

Eduardo Alemán and George Tsebelis

Print publication date: 2016

Print ISBN-13: 9780198777861

Published to Oxford Scholarship Online: August 2016

DOI: 10.1093/acprof:oso/9780198777861.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2019. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 17 September 2019

Agenda Setting and Gridlock in a Multi-Party Coalitional Presidential System

Agenda Setting and Gridlock in a Multi-Party Coalitional Presidential System

The Case of Brazil

Chapter:
(p.61) 3 Agenda Setting and Gridlock in a Multi-Party Coalitional Presidential System
Source:
Legislative Institutions and Lawmaking in Latin America
Author(s):

Taeko Hiroi

Lucio R. Rennó

Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780198777861.003.0003

Abstract and Keywords

Does the concentration of agenda-setting powers in the president and the governing coalition in the legislature help to overcome institutional and political hindrances to lawmaking in Brazil? This chapter addresses this question by examining various legislative data, including legislative proposals considered by the Brazilian Congress, roll call votes, and obstructionist movements. It describes the various rules affecting legislative decision-making and the complexities involved in reaching coalition agreements. It also evaluates whether distinct configurations of the governing and opposition coalitions and their internal unity affect the processing of legislative bills. Hence, it provides a test for how agenda setting powers offset potential complexities of the legislative process that could stimulate gridlock and obstructionism.

Keywords:   Brazil, agenda setting, gridlock, obstructionism, coalitions, legislative rules, bill approval, amendments

Presidents in Latin America commonly have several agenda setting powers and are important lawmakers in the land. Yet, many of these presidents regularly face situations of minority governments or majority coalitions, with difficult intra-coalition bargaining and a significant potential for gridlock (Cheibub 2007; Shugart and Carey 1992; Tsebelis and Alemán 2005; Ames 2001). Although party systems in Latin America vary dramatically in their levels of institutionalization and fragmentation (Mainwaring and Scully 1995), overall the number of parties in the region surpasses the two-party logic of the US case. The dynamics of legislative conflict in such environments is quite distinct from that of divided government. Consequently, mapping agenda powers and the processes of negotiation within and between coalitions composed of heterogeneous parties is fundamental to understand lawmaking in the region.

Consider, for instance, that Brazilian presidents typically put together governing coalitions comprising various parties. With some twenty parties in the legislature, the president’s party alone never holds more than 20 percent of the seats in the Chamber of Deputies, generating the need and incentive for coalition government. The low level of Brazilian party institutionalization and high party fragmentation add to the governability problem (Mainwaring 1999). For these reasons, since 1994 successive presidents—Fernando Henrique Cardoso, Luiz Inácio Lula da Silva, especially during his second term in office, and Dilma Rousseff—have held numerically large coalitional majorities in the Chamber of Deputies and variously sized coalitions in the Senate. Despite the large coalitional majorities that Brazilian presidents have amassed, they have not always been successful in getting their legislative initiatives approved by Congress. In the situations where Congress approves the government’s proposals, the process is generally not swift or without major (p.62) alterations to the proposal. Why is this? A part of the answer to this question requires us to look closely inside the coalitions. Coalitions in Brazil do not function like a single party. Consequently, we need to examine the cohesiveness of coalitions and their consistency over time to understand the dynamics of legislative conflict in Brazil. The political landscape of Brazil also reinforces the argument of Alemán and Tsebelis in the introduction to this volume that underlines why, in the absence of a cohesive government majority, studying the workings of legislative institutions and the positions of the main legislative actors becomes crucial to understand policy outcomes.

The Brazilian case is also a great example to study institutional incentives and constraints in legislative decision-making. There are several institutional factors, such as various tools of legislative obstruction, sequential examination by legislative committees, and bicameral deliberation, that impose hurdles for the lawmaking process. The layers of complexity are increased when we consider the three-fifths majority and two-round votes in both the Senate and Chamber of Deputies that are required to pass constitutional amendment proposals. Clearly, the obstacles are many for policy-making inside the Brazilian Congress.

Yet, presidents hold powerful prerogatives that centralize the agenda setting power in their hands. Presidents possess special legislative privileges, such as the power to present proposals for constitutional amendments, issue decree-like provisional measures, and request urgency for the consideration of their ordinary law proposals in Congress. In addition to their proposal powers, presidents can also veto legislation either partially or fully. Presidents also hold the power of the purse, being responsible for the formulation and presentation of annual budgetary laws and, later, the disbursement of discretionary transfers (Ames 2001). Finally, presidents have discretion in the nomination of many political appointees, with estimates of over 20,000 nominations, some of which carry with them power over significant budgetary funds. Amorim Neto (2006) points out that presidents allocate cabinet positions to manage their coalitions, and other authors have shown that such strategy also affects much lower levels of the bureaucracy (Praça et al. 2011).1 Several studies have shown that the proportional distribution of cabinet positions among members of the coalition is an important factor to understand presidents’ legislative success (Amorim Neto 2002) and choices of policy-making instrument (Pereira et al. 2005).

Rules inside the Chamber of Deputies and Senate strongly centralize agenda setting powers in the hands of the leaders of the largest parties and the majority coalition. The majority coalition usually controls key positions in the hierarchy of the two houses, such as the presiding officer of the steering committee (called Mesa) and chairs of key committees, including the Constitution Committee (Comissão de Constituição e Justiça e de Cidadania) which oversees the constitutionality of all legislative proposals.

(p.63) Does the concentration of agenda setting powers in the president and the governing coalition in the legislature help overcome institutional and political hindrances to lawmaking? In this chapter, we address this question by examining various legislative data, including legislative proposals considered by the Brazilian Congress, roll call votes, obstructionist movements, and various configurations of the governing and opposition coalitions and their internal unity. We examine if Brazil is an example of how presidential agenda setting powers and a centralized decision-making process within Congress can offset the potential governability problems generated by complicated coalition building and management and the institutional complexity of the lawmaking process. In Chapter 1, Alemán and Tsebelis hypothesize that granting substantial agenda setting prerogatives to presidents should help overcome some of the problems generated by heterogeneous policy positions inside a governing coalition.

In the next section, we discuss the legislative processes stressing presidential prerogatives, actors, rules and procedures within Congress, and some current interpretations of lawmaking in Brazil. We then derive hypotheses from both the theoretical discussion and the implications of the Brazilian institutional and partisan framework. Finally, we test these hypotheses empirically using various legislative data.

Legislative Actors, Institutions, and Processes

This section discusses institutional prerogatives of the key legislative actors and formal legislative rules and procedures under Brazil’s 1988 constitution. Brazil has a presidential system with a bicameral Congress. Several actors can propose bills, including the judiciary, private citizens, congressional committees, federal deputies, senators, and the executive branch. There are many types of bills with different majority requirements. The legislative process is one of sequential decision-making, prone to nested games (Tsebelis 1990) and various occasions for the blocking of legislation by veto players (Tsebelis 2002).

Presidential Prerogatives

Unlike the US president, the Brazilian president is endowed with substantial legislative prerogatives. In the Shugart and Carey (1992) classification assessing presidential legislative powers, the Brazilian president ranks among the (p.64) most powerful presidents in the world. For example, the Brazilian president enjoys tremendous agenda setting power. The country’s constitution grants the president the power to initiate any type of bill,2 including exclusivity over the initiation of provisional measures, budgetary laws, and administrative changes. The president can also request urgency petitions and convoke extraordinary sessions. Executive proposals are automatically granted priority status for consideration in Congress, and with urgency motions (called pedido de urgência) the president can require Congress to expedite the examination of statutory bills. Should either house fail to conclude its deliberation of a bill under urgency regime within the deadline of forty-five days,3 the deliberation of any other proposals is suspended in that house until the bill is voted on.

Presidents also have the prerogative to initiate budgetary laws and request additional budgetary credits during the year, tax collection permitting. The annual budget (called Annual Budgetary Law, or Lei Orçamentaria Anual) requires congressional approval to be effective, but its approval is not an automatic guarantee of its execution. In a given fiscal year, the Ministry of Finance monitors revenues and expenditures and exercises great discretion over the release of funds (Hiroi 2009). Known as the authorized budget (orçamento autorizativo) in Brazil, this prerogative grants presidents impressive leeway in deciding if and when budgetary funds will be transferred, especially the so-called voluntary (discretionary) transfers.4

Federal deputies and senators have the prerogative of presenting amendments to the budgetary law benefiting municipalities, states, and social organizations. Various studies have analyzed these transfers as forms of pork barrel politics (Ames 2001; Pereira and Renno 2003; Ames et al. 2011). After being approved as part of the annual budget law, the executive branch has the prerogative of deciding which and when budgetary amendments will be executed. Some authors argue that the execution of budgetary amendments is a bargaining tool of the executive branch, intermediated by party leaders, to increase support in floor votes (Ames 2001; Alston and Mueller 2005).

The most controversial among presidential prerogatives is probably the authority to issue executive decrees called provisional measures (medida provisória in Portuguese). Although the constitution allows the use of presidential decrees only in matters of “urgency and relevance” (Article 62), Brazilian presidents have used this prerogative in a wide range of areas—from purchasing an automobile for a vice president to outlawing bingo games, introducing new currencies, altering tax rates, and implementing economic stabilization plans. The range and frequency of policy changes enacted through presidential decrees have been such that many prominent members of Congress deplore that this constitutional authority is more authoritarian than the infamous institutional acts of the military regime.5 Frustrated with (p.65) the frequent use and coverage of presidential decrees, Congress attempted to regulate and restrict this powerful presidential power, which culminated in Constitutional Amendment No. 32 of 2001.

The 2001 constitutional amendment significantly curtailed the number of areas where decrees could be used. It also prohibits reissuing of decrees. Presidential decrees are now valid for sixty days, and are automatically renewed only once if Congress fails to vote on them expeditiously. The new rules also provide that the failure to vote on decrees within forty-five days triggers a suspension of deliberation of all other legislative activities in the respective house until final voting occurs. The modes of decree deliberation also changed. The constitutional amendment requires that a special joint committee of the National Congress (composed of senators and deputies) evaluate the urgency and need criteria before it could be voted on its merit. This joint committee, however, was never installed in practice until 2010, when deputies and senators reacted strongly against the abusive use of provisional measures (Renno 2010). Another major change introduced by the constitutional amendment is the sequential examination of decrees beginning in the Chamber of Deputies first, and then followed by the Senate, instead of requiring a joint session of the National Congress.

Some argue that the result of the 2001 reform was, ironically, an increase in the use of provisional measures (Pereira et al. 2008). The suspension of deliberation after forty-five days of the issuing of a provisional measure forces Congress to attend to the agenda of the executive branch. Therefore, presidents rely even more on provisional measures to further increase their control over the congressional agenda. In 2009, a new interpretation of the suspension clause emerged. Under the new interpretation, only ordinary law proposals would be blocked by a delay in a vote on a provisional measure, whereas complementary law proposals and constitutional amendments (along with all other types of legislative proposals) would not. Whether this change has had any effect in reducing presidential decrees is still being analyzed. Renno (2010) finds a reduction in decrees in the early aftermath of the reform, whereas Almeida (2011) has questioned these results with a slightly longer time frame.

The Brazilian president also has both partial and total veto powers. Although Congress can override the president’s vetoes with the vote of an absolute majority of each chamber in a joint session, it has been very rare that Congress overrides presidential vetoes. Between 1988 and 2005, Congress overturned presidential vetoes only thirteen times (Hiroi 2005). Furthermore, Congress rarely votes on presidential vetoes in flagrant disrespect of the constitution. Only recently (in 2014) has Congress been forced to position itself in relation to prior vetoes that were not voted on, after a ruling by the (p.66) Federal Supreme Court. With the judicial ruling, Congress voted on hundreds of vetoes en masse, with almost all being sustained.

To sum up, the list of positive agenda setting powers held by presidents includes the prerogatives to issue provisional measures, originate any type of bill, and request urgency motions. Negative agenda setting powers used for gatekeeping include the exclusive power of initiation for administrative and budgetary bills, partial and total veto powers, and control over the disbursement of budgetary funds.

Congress

The Brazilian Congress consists of the Chamber of Deputies and the Federal Senate. Any member of Congress can propose a statutory bill. The most influential actors within each house are the presiding officer (Presidente) of the steering committee (Mesa), who retains agenda setting and gatekeeping powers, and party leaders, who appoint and discharge committee members and make recommendations for votes. The president of the Senate is also the president of Congress, presiding over joint meetings of the Chamber of Deputies and the Senate convened to analyze presidential vetoes.

In each house of Congress, the presiding officer is responsible for organizing the legislative agenda every month. In the Chamber of the Deputies, the presiding officer consults with the College of Leaders (Colégio de Líderes) composed of party leaders, minority and majority leaders, and a deputy representing the government. In the Senate, there is no equivalent leadership caucus, and therefore the presiding officer has even more leeway. In both houses, the presiding officer also arranges the order of the day (Ordem do Dia), indicating which legislative proposals will be moved to floor discussions that week. In the Chamber of Deputies, this occurs during the weekly meetings of the College of Leaders. This prerogative gives the presiding officer the ability not only to select the materials to be discussed but also to decide when they are voted on.

The presiding officers of the Chamber of Deputies and the Senate can also call for extraordinary daily sessions as needed, and jointly can convoke an extraordinary legislative session during seasons of congressional recess to deliberate on the materials of legislative priorities. They can also install ad hoc and special committees to consider issues of particular importance. Special committees are also automatically implemented when a bill is referred by the steering committee to more than three permanent committees to examine merit or, in the case of the Chamber of Deputies, to consider proposals for constitutional amendment.6

(p.67) Party leadership also has significant control over the organization of legislative work in Congress. As discussed earlier, party leaders influence legislative priorities in the scheduling of the legislative agenda. Decisions in the College of Leaders are made by consensus whenever possible, and by an absolute majority of votes weighted by the size of each party when necessary. Party leaders also give vote recommendations to members of their parties and appoint (and discharge) their members to (or from) committees.

Party leaders and members of Congress also have many procedural prerogatives that may significantly affect the speed of decision-making within Congress. For instance, in the Chamber of Deputies, urgency motions (regime de urgêngia) may be presented (subject to approval by the plenary) by two-thirds of the members of the steering committee or one-third of the members of the house or party leaders representing this number. Urgência urgentíssima, a type of discharge petition requiring a bill to enter the order of the day immediately for discussion and votes, may be requested by an absolute majority of the chamber membership or party leaders representing this number and must be approved by an absolute majority in the plenary.

There are also procedures to delay discussion and votes. Party leaders, the rapporteur, and the author of a proposal may request postponing discussion and votes for up to ten days. A request for postponing the deliberation of materials examined under urgency regime and a request forcing a separate vote on distinct parts of a bill require a motion supported by one-tenth of the chamber members or party leaders representing this number.7

Committee membership in both houses is distributed to parties based on the proportionality principle. Party leaders appoint (and discharge) members to different committees. Each committee chooses its chairperson by a simple majority vote, and committee chairs designate rapporteurs for different proposals. Rapporteurs analyze bills and amendments submitted by committee members and make their recommendations about them. Each committee has forty legislative sessions to conclude its analysis, and rapporteurs must produce their reports within twenty days. It is up to the rapporteurs to decide, discretionarily, which amendments will be included in their reports. If rapporteurs fail to produce their reports within the deadline, they can ask for extensions. Committee chairs can accept the request or nominate another rapporteur, who will have five sessions to produce his or her report.8

The extent of changes on a bill proposed by rapporteurs varies. They can simply accept the initial bill without amendments, include amendments while maintaining the original text, or propose an altogether distinct bill, quite different from the original one, called a substitutive bill (substitutivo do relator). Rapporteurs and committee chairs who appoint them, therefore, are powerful actors in the legislative process. We can think of rapporteurs as (p.68) agenda holders, as suggested by Silva and Araujo (2012), because of their control over the bill for a predefined period of time, with great influence on its content and destiny.

The committee then votes on the report that defines the amendments and any change proposed by the rapporteur. Committees make decisions by a simple majority vote. This vote is symbolic, by a raise of hands, or roll call, if there is a specific request for it. If the committee is unable to come to a decision, the bill may die a slow death at the committee stage without receiving a vote, unless a discharge petition is approved to bring the bill to the floor.

Bills classified as ordinary laws may be approved by the assigned committee and sent to the reviewer house dispensing with floor discussion and vote. This fast-track procedure, called conclusive power (in the Chamber of Deputies) or terminative power (in the Senate), allows for approval or rejection of this type of bills after being examined only by the relevant committees.9 However, with the support of one-tenth of the members, deputies and senators can challenge the decisions made by the committees and bring the bill to a floor vote.

The most powerful committees in the Brazilian Congress are the Constitution, Justice, and Citizenship Committee (Comissão de Constituição e Justiça e de Cidadania), which verifies the legality of proposals in both houses, and Finance and Tax Committee (Comissão de Finanças e Tributação) in the Chamber of Deputies, which examines the financial viability of proposed projects. Both of these committees are gatekeepers, exercising negative agenda setting powers, with the prerogative to reject bills on the grounds of constitutionality and economic viability. After reforms in the early 1990s, the internal rules (Regimento Interno) of the Chamber of Deputies established that the evaluations of constitutionality and economic viability be issued after the debate on merit in the permanent committees with jurisdiction over a bill. This potentially weakened the gatekeeping capacity of the Constitution and Finance committees, as it may be more difficult to reject a bill that has already been approved by permanent committees.10

In a normal process (i.e., when the conclusive power does not apply), committees examine bills sequentially, and the ones approved by all committees then move to the plenary. If a committee rejects a bill, its examination by Congress typically ends. But committee decisions can be challenged with a request, supported by one-tenth of the house membership (or party leaders representing this number), to bring the bill to the plenary. In general, research on congressional committees is scarce in Brazil, and the actual powers of the committees are still relatively unknown.11

Referral of bills to multiple committees is common. For example, in the Chamber of Deputies,12 all bills in principle must be referred to the Constitution Committee, which analyzes their constitutionality. All bills that have (p.69) financial implications must also, in principle, be referred to the Finance and Tax Committee. In addition, bills are referred to other relevant committees to consider the merit of the proposals. As stated before, when the number of relevant committees to examine the merit of a proposal exceeds three, or when dealing with proposals for constitutional amendment, a special committee is formed to deliberate on the merit, rather than referring the bill to many standing committees. Committee deliberation is one area where many bills suffer a slow death, without ever reaching the plenary floor.

Decision Rules and Procedures

In Brazil there are three principal avenues of lawmaking through ordinary means: constitutional amendments and two types of statutory (ordinary and complementary) legislation. Complementary laws regulate provisions specifically referred to by the constitution. Ordinary laws are statutes that regulate areas not designated to complementary law. Article 60 of the constitution states that a constitutional amendment proposal may be submitted by: (1) at least one-third of the members of the Chamber of Deputies or of the Senate; (2) the President of the Republic; or (3) more than one-half of the Legislative Assemblies of the units of the Federation. In practice, all proposals for constitutional amendment have been submitted by the executive or Congress. The initial house of deliberation depends on the author of the proposal. The Chamber of Deputies is the first house to consider bills proposed by deputies and all executive proposals. The Senate is the initial house to consider bills proposed by senators. The initial house may approve the bill as it is, approve the bill with amendments, or reject the bill. An approval of a constitutional amendment proposal requires favorable votes by a three-fifths majority of its members taken by roll calls in two separate rounds. If the initial house approves the proposal, it moves to the second house for a review.

The reviewing house also has three options: approve the bill as it is, approve the bill with amendments, or reject the bill. If the reviewer house approves exactly the same text as the one approved by the first house, the steering committees of the Chamber of Deputies and the Senate promulgate the constitutional amendment. Constitutional amendments are not subject to presidential sanction or veto. If the reviewer house rejects the bill, it is sent to the archive. If the reviewer house approves the bill with amendments, the first house must consider the bill once again as a new proposal. It is worth re-emphasizing that the Brazilian constitution requires that an identical text of a constitutional amendment be approved by the two houses of Congress for enactment. Until (p.70) an identical text is approved, the bill could shuttle between the two houses indefinitely until it is rejected or terminated. In addition, urgency may not be requested for an examination of constitutional amendments.

In both houses, committees review constitutional amendment proposals. As stated in the previous section, in the Chamber of Deputies, the Constitution Committee considers, in principle, all bills and produces reports on their admissibility. In the case of inadmissibility, the author of the proposal, with the support of at least one-third of deputies, may appeal to the floor to consider the reversal of the Constitution Committee’s decision. If the Constitution Committee or the plenary floor votes in favor of admissibility, the presiding officer designates a special committee to examine the merit of the proposal. The special committee has forty sessions to produce its report. Amendments to the proposal must be submitted to the special committee with support by at least one-third of the chamber membership. After the examination by the special committee, the proposal moves to the plenary for discussion and vote.

In the Senate, the Committee on Constitution, Justice, and Citizenship analyzes the merit as well as the admissibility of constitutional amendment proposals. The Constitution Committee in the Senate has thirty days to submit a committee report. Following approval by the Constitution Committee, the plenary discusses the proposal for five sessions. Senators may submit amendments during these sessions with signatures of at least one-third of the members of the house. If no amendment is submitted, the bill is put for a vote at the fifth session. Should amendments be submitted, however, the bill returns to the Constitution Committee for their analysis. The proposal and amendments are voted on the floor after the Constitution Committee reports its analysis to the plenary.

Proposals for ordinary law and complementary law may be submitted by any of the following individuals or collectives within their competence: any member(s) or committee of the Chamber of Deputies, the Senate, or the National Congress; the President of the Republic; the Supreme Federal Court; the Superior Courts; the Prosecutor-General; and the citizens. An approval of a bill of ordinary law requires a simple majority whereas an approval of a bill of complementary law requires an absolute majority and roll call votes. Once a statutory bill is submitted, the presiding officer of the house distributes it to relevant committees, indicating whether the bill follows a normal or fast-tracking process. In the Chamber of Deputies, thematic committees and the Constitution Committee, in addition to the Finance and Tax Committee when it has financial or budgetary implications, examine statutory bills. In the Senate, only thematic committees examine these types of proposals. In the case of a fast-tracking process using conclusive or terminative power, (p.71) members of the respective house submit amendments to the committees and the reporting officers prepare their analyses of the bill and amendments and make recommendations. The committees then vote on the bill and amendments. With committee approval, the proposal is forwarded to the second house for a review. In the case of rejection, it goes to the archive. As discussed in the previous section, deputies and senators who do not agree with the committee decision may appeal to the plenary if they have the support of at least one-tenth of their house membership.

Bills considered under the normal regime will move to the plenary for discussion and votes after committee examinations. The presiding officer of the house, in consultation with party leaders, determines when these bills enter the order of the day. Once included in the order of the day, deputies (or senators if the bill is initiated in the Senate) have five legislative sessions to present amendments, which are then sent to the relevant committees. Next, rapporteurs elaborate their reports on the floor amendments. The steering committee will then include the bill and accompanying reports again in the order of the day.

Inclusion in the order of the day is a very high hurdle, and it is at this point that the president of the steering committee very clearly exercises his agenda power. Inclusion in the order of the day then leads to debate on the floor and votes. Very few bills reach this stage of the process. Moreover, few bills that enter the order of the day are rejected; those bills are included precisely because they have the support of the party leaders for plenary deliberation. Hence, inclusion in the order of the day is a sign that there is enough agreement over content, at least among the majority, for the bill’s approval. Conversely, the wait period may represent a slow death for the bill. Because of this, legislators and party leaders may request a discharge petition as a strategy to stop or slow down the processing of a bill. Discharge petitions remove ordinary bills subject to deliberation by conclusive power from committees and move them to plenary deliberation. Those bills then enter a long waiting list for inclusion in the order of the day. Thus, discharge petitions can paradoxically lead to delays in the final evaluation of a legislative proposal, even though they dispense with the committee stage of the legislative process.13

In both houses, if there is a substitutive bill prepared by the bill’s rapporteur, the floor votes on it first. If it is approved, the floor subsequently votes on amendments and specific parts of the bill. If there is no substitutive bill or if it is rejected, then the original bill will be voted on, followed by votes on amendments and specific parts of the bill. Members of each house can contest the rapporteur’s decision to reject amendments on the floor. Once bills are approved on the floor, they proceed to the reviewer house.

(p.72) As with the case of constitutional amendment proposals, the reviewer house may approve statutory bills as they come from the initial house (i.e., without amendments), approve them with amendments, or reject them. Proposals received from the first house are examined by a committee or committees first, and, with favorable committee report, proceed to the plenary. If a bill is approved without amendments, the reviewer house sends the approved text to the president for enactment (or vetoing). If the reviewer house approves the bill but modifies the text, the bill is returned to the initial house for a review of the changes made by the second house. Unlike constitutional amendments, it is the prerogative of the initial house to accept or reject the modifications made by the reviewer house. In other words, the initial house may disregard all the amendments of the reviewer house and send its original text to the president. Hence, there is a substantial advantage to be the first house to consider statutory bills. Nevertheless, the reviewer house does have the power to veto bills sent by the other house.

Finally, Brazilian legislators have many instruments to obstruct or fast-track decisions. For example, legislators can ask for quorum and vote verifications, which change the votes from symbolic to roll call. They can also maneuver procedural rules by asking for votes on separate aspects of a bill, by requesting that discussion be postponed, or by requesting that a proposal be removed completely from the agenda. Such tactics may reconfigure the contexts of bargaining inside Congress. Indeed, the use of such instruments is on the rise in Brazil (Inácio 2009; Hiroi and Renno 2014).

Votes by which legislators can clearly delay legislative business include request for the removal of proposals from the order of the day, request to delay discussion and vote, and request to separately vote on different articles of a bill. Request for a removal of a proposal from the agenda may also be used as a defensive maneuver to avoid a defeat. In either case, it delays a final vote.

On the other hand, there are votes to expedite decision-making, such as urgency petitions or requests for termination of discussion, or agglutination of several amendments in a single one. A motion to invert the order in which a bill will be voted on can be used to delay a vote of interest in exchange for a less conflictual one, or to bring a more important vote to the top of the agenda.

There are also votes on substantive issues regarding the content of policy proposals. These can be in the form of a vote on an ordinary or complementary law proposal, constitutional amendment, or executive decree. However, some types of substantive votes may also result in prolonged debate. For instance, votes on a specific article that received a request for exclusive evaluation, a vote on an amendment, and votes on proposal reports (pareceres) all incur longer deliberation of single bills. Delays occur because they are split up and voted separately, generating a large number of votes on a single proposal.

(p.73) In summary, the legislative process is long and arduous. Approving a bill in the committee or on the floor is not an easy task and transforming it into law is even harder given the requirement for bicameral approval and presidential ratification.

Taking Actors’ Positions Into Account

In post-democratization Brazil, the ideological and policy positions of parties and their members have spanned from the left to the right. There are leftist parties, led by the Workers’ Party (PT); centrist parties, represented mainly by the Brazilian Democratic Movement Party (PMDB) and Brazilian Social Democracy Party (PSDB); and right-wing parties, such as the Democrats Party (DEM, former PFL) and the Progressive Party (PP), which are diminishing in size in recent years.

Heterogeneous coalition members mean more distance between their preferred policy positions. If members of the coalition are veto players (Tsebelis 2002), heterogeneity implies a diminished winset size for policy change (see Chapter 1 of this book), thus making effective coalition management even more imperative. The governing coalitions of Presidents Cardoso, Lula (after the first year), and Rousseff have all been oversized, including diverse mixes of political parties. The coalitions of Lula and Rousseff were particularly heterogeneous, including parties that are positioned on the far left and on the far right (see Table 3.1 for party positions). Although the composition of governing coalitions typically shifts during a presidential term, for example, during Lula’s two terms, the number of parties gaining cabinet positions was approximately nine. Between 2007 and 2009, ten parties from the left to the right (PT, PTB, PR, PMDB, PV, PDT, PSB, PP, PCdoB, and PRB) held cabinet posts. During Rousseff’s first term, her coalition was even larger in size than Lula’s and as heterogeneous as that of her predecessor. Such heterogeneous coalitions could exacerbate the problems of coalition unity. Even though a recent study shows an attitudinal convergence of Brazilian deputies on many fundamental questions (Power and Zucco 2012), the increase in governing coalition size and coverage in recent years appears to have shifted the major arenas of legislative conflict from that between government and opposition to that within the governing coalition.

Table 3.1 Brazilian parties from left to right, 1990–2009

Year

1990

PL

PDC

PCdoB

PT

PCB

PSB

PDT

PSDB

PMDB

PTB

PRN

PFL

PDS

1993

PT

PL

PCdoB

PSTU

PSB

PPS

PDT

PSDB

PMDB

PP

PTB

PFL

PPR

PRN

1997

PSB

PTB

PFL

PCdoB

PT

PPS

PDT

PMDB

PSDB

PR

PL

PPB

2001

PMDB

PL

PCdoB

PT

PSB

PDT

PSDB

PTB

PFL

PPB

2005

PT

PPS

PMDB

PTB

PFL

PCdoB

PSB

PDT

PSDB

PL

PP

2009

PT

PMDB

PTB

PSOL

PCdoB

PSB

PDT

PV

PPS

PSDB

PR

PP

DEM

Notes: Stacked parties, in any given year, indicate that the differences between their estimated positions are not statistically different at the 0.10 confidence level. All other adjacent parties are statistically different from one another.

Source: Based on data provided in an earlier version of Power and Zucco (2012).

Besides partisan differences, the literature on bicameralism indicates that bicameral incongruence is an important cause of legislative delay and gridlock (Tsebelis and Money 1997; Hiroi 2008a, 2008b). In Brazil, the upper and lower (p.74) chambers may well have different policy positions. The sources of bicameral incongruence include different electoral rules to select their membership. In Brazil, senators are elected by a majoritarian rule for an eight-year term and deputies by open-list proportional representation for a four-year term. The two houses also tend to differ on the political experiences of their members. Most of the senators have extensive political experiences; the Senate has had members who have served as presidents of the republic, vice presidents, ministers, governors, and federal deputies (Hiroi and Neiva 2013). Federal deputies’ political experiences pale in comparison, although there are several members with long and distinguished political careers. The political experiences of senators tend to make them more sympathetic to issues related to public finance and administration than deputies. As such, the Senate has been considered more “governista” than the Chamber of Deputies, rendering more cooperation with the executive branch in the amendment and passage of bills, coming scratched from the lower house. However, this relationship may have changed recently, especially during President Lula’s period. Research in Brazilian bicameralism is still scarce, but existing studies indicate that bicameral incongruence does affect both the speed and approval of legislative proposals (Hiroi 2008a).

(p.75) Hypotheses

The interaction between Brazil’s legislative institutions and procedures and policy positions of relevant actors generate interesting hypotheses. Our first hypotheses refer to presidents’ ability to advance their policy agenda in Congress. As discussed earlier, presidents possess various agenda setting prerogatives and tools to move forward their legislative proposals and block legislation they do not support. In addition, presidents’ allies in Congress usually occupy important positions endowed with additional positive and negative agenda setting powers, such as the leadership in the steering committee. We therefore expect high success rates of executive proposals relative to those of congressional proposals.

H1. Executive proposals are more likely to be approved, and approved faster, than congressional proposals.

H2. The government’s position will be on the winning side in most roll call votes.

Despite the agenda setting powers that favor the president, there are many hindrances to proposal approval due to numerous hurdles built into the legislative process, such as qualified majority rules, multiple veto points in the committee stage, many instruments for legislative obstruction, requirement for concurrent approval of the two distinct houses of the legislature, and other intricacies of the legislative process. These rules provide abundant opportunities for the members of Congress to stamp their marks on the executive proposals. Hence, even if many executive proposals pass Congress, we expect that those bills, particularly major bills that propose significant change, will be altered by Congress.

H3. Congress will amend many executive proposals and most major bills initiated by the executive.

In addition, as discussed in the introductory chapter of this volume, ideological distances between the key legislative actors have implications for policy change. Thus, we examine the ideological distances between legislative actors, not only the distance between the governing and opposition coalitions but also distances within legislative coalitions, in order to fully understand the legislative processes and outcomes. In Brazil, since coalitions are the major legislative forces, we need to examine the cohesiveness of legislative coalitions and coalition management. We define coalition cohesiveness as the ability of the members of the coalition to act in unity in legislative matters. Lack of cohesion within a governing coalition represents a smaller winset of the status quo (p.76) because the distance between political actors within the coalition is high. Conversely, a cohesive coalition represents actors that are closer together and hence a larger winset. We expect that an uncohesive government coalition will create more bargaining difficulties and more obstruction, generating delay in legislative approval.

H4. The less cohesive the governing coalition, the more bargaining and obstruction there will be in the legislative process, leading to a greater delay in the passage of bills.

On the other hand, lack of cohesion in opposition coalitions would result in the opposite effect: a heterogeneous group of opposition parties may not be able to function as a unified front to counterbalance the forces of the governing coalition. In contrast, oppositions’ ability to obstruct the legislative process and frustrate the governing coalition should be high when they are united. Given the massive institutional and resource advantage of the executive branch and its congressional allies that usually control the agenda setters’ positions and majority seats, we expect that the oppositions’ influence is observed mostly in obstructing the legislative process and delaying and amending proposals that will be approved.

H5. The more cohesive the opposition coalition, the greater its ability to engage in dilatory practices, leading to a greater delay in the passage of bills.

The distribution of patronage, especially in the allocation of cabinet posts among coalition partners, is presidents’ essential tool for coalition management. Presidents strategically allocate ministerial portfolios as a means to construct a legislative coalition (Amorim Neto 2006). The distribution of cabinet positions affects the division of power among coalition members. The better distributed these resources, the more satisfied coalition members, and hence the fewer disputes there are internally. If parties participating in the coalition do not share these positions in a seemingly “fair” manner, dissatisfied coalition partners may impede or threaten to impede proposed legislation until satisfactory reallocation of these posts is undertaken (Amorim Neto 2002).

H6. The less proportional the distribution of cabinet posts among coalition partners relative to their legislative weights, the more bargaining and obstruction there will be in the legislative process, leading to a greater delay in the passage of bills.

The traditional literature on legislative conflict in presidential systems has emphasized government–opposition conflict. If a line separating the government and opposition exists in a meaningful way, such as along the line of ideological or policy disputes, and if there is an opposition large enough to (p.77) override or attenuate the institutional advantages of the president, then the larger the distance between governing coalition and opposition, the less likely legislative proposals would be approved. The following hypothesis suggests how the ideological distance between the two coalitions affects legislative expediency.

H7. The greater the ideological divide between the opposition and the government coalition, the more obstruction there will be in the legislative process and hence the greater the delay in the passage of bills.

Brazil’s legislature has a symmetric bicameral structure.14 Legislative approval requires sequential examination by the two houses. There is no practice of a conference committee to resolve bicameral disputes, although bicameral joint committees are occasionally created.15 Incongruence between the two houses is likely to lead to disagreements over policy change, which in turn should generate a greater delay in approval.

H8. The greater the incongruence between the upper and lower houses of Congress, the more time it will take for a bill to be approved.

Data

We test the hypotheses using data on individual bills submitted to the Brazilian Congress between 1995 and 2003 and roll call votes that occurred in the Chamber of Deputies between 2003 and 2011.16 The roll call data include all roll call votes that occurred during this period, covering the two Lula da Silva terms and the beginning of Dilma Rousseff’s term. The proposal dataset covers the two terms of Fernando Henrique Cardoso (1995–8 and 1999–2002) and the initial year of Lula da Silva’s first term (2003–6).

The proposal data track the fate of individual bills submitted to Congress between 1995 and 2003. The dataset consists of proposals for constitutional amendment and two types of statutory bills—ordinary and complementary. It does not include presidential decrees (provisional measures) because they are not bills. The dataset includes all executive and judicial proposals that were introduced in Congress. With respect to congressional proposals, the dataset includes all bills submitted and subsequently approved at least by the house of origin.17 By this method, we are in practice examining institutional bills that passed the initial internal deliberation process. We thus call these bills “institutional” bills, and when referring to “all” bills in the subsequent discussion, we are referring to all of these institutional bills in our sample.

(p.78) We are interested in the approval of bills and how long it takes for a bill to be approved. We converted the proposal data into monthly data and traced, in days, the histories of these bills from their introduction to Congress by assigning a series of 0s for each observational period until their approval, at which time a value of 1 is assigned. Bills that were terminated for reasons other than their passage (e.g., rejection, withdrawal by the author, or simply termination of deliberation at the end of a legislative period) are treated as censored at the time of the decision.18 In the case of pending bills, their histories are traced until July 31, 2004, on which date they “exit” the dataset. This cut-off date was determined by the availability of data in the dataset.

Analysis

To probe the hypotheses outlined in the previous section, we examine descriptive statistics and then statistically analyze the passage of time in Congress until a legislative proposal is approved.

Descriptive Statistics

Table 3.2 Number of proposals mentioned by the media, 1995, 1999, 2003

Media mention

Cardoso—1994

Cardoso—1998

Lula—2002

Proposals not mentioned by the media

628

862

82

Proposals mentioned by the media

16 (8)

7 (1)

2 (2)

Note: The figures in parentheses represent executive proposals.

Source: Compiled by the authors.

Bills’ fates in Congress likely depend on their salience. To measure issue salience, we first identified legislative issues that appeared on the first page and in the politics section of Folha de São Paulo, a leading newspaper in Brazil, during the first years of the first and second terms of President Cardoso and the first term of President Lula. We coded not only the explicit mention of a law proposal, but also more general issues that would be on Congress’ agenda, such as “tax reform,” and “social security reform.” We then identified these issues with the specific bills dealing with the topics.

Table 3.2 presents the number of major bills mentioned by the Folha during the first years of the Cardoso and Lula administrations. First, it is interesting to (p.79) notice the small number of such proposals, totaling only twenty-five cases of all institutional bills that were submitted during this period. This is partly due to the fact that Brazilian presidents often rely on issuing decrees, which effects instant change, rather than proposing bills, which will be subject to long congressional deliberation. Some of these decrees received intense media coverage, but they are not included in the table because they are not legislative proposals.

We also note that the majority of legislative action regarding major bills occurred during the Cardoso administration. This is not surprising given that Cardoso’s legislative coalition during his first term was assembled based on a pact of economic modernization and reform. During the first year, a number of major constitutional amendments dealing with economic issues were proposed and subsequently approved.

In contrast, only two major bills were proposed during the Lula period in our dataset. This is partly because we only have one year of proposal initiation during the Lula government in our dataset. But it is also because the Lula government considered two constitutional amendment proposals (pension reform and tax reform) as its hallmark legislation and focused on the passage of these proposals within the first year.19 Still, the number of mentions at the beginning of Lula’s first term, in comparison to Cardoso’s first term, is much lower.

Of the sixteen major bills considered during Cardoso’s first term, the executive branch presented eight proposals. This is the highest level of activity by the executive branch during the period under investigation. During Cardoso’s second term, of the seven major bills, the executive initiated only one. During Lula’s first term, the executive branch originated both major bills. Thus, fourteen major bills originated in the legislative branch and eleven in the executive branch.

Agenda Setting and Gridlock in a Multi-Party Coalitional Presidential SystemThe Case of Brazil

Figure 3.1 Proportion of enacted bills by origin

The passage rates of major bills indicate greater success rates of the executive branch compared to those of Congress. During the Cardoso periods, all but one major bill proposed by the executive was approved, whereas only one-half of the major bills initiated by the legislative branch were approved. Both of Lula’s bills were approved. Hence, if we focus on the passage rates, the executive branch is twice as successful as the legislative branch in approving major bills.

In general, the executive branch is more successful than Congress in getting its legislation enacted. Figure 3.1 presents the proportion of institutional bills that were enacted by administration and origin.20 During the first Cardoso term, 64 percent of all the executive proposals were enacted into law, compared to 29 percent of the congressional proposals. During Cardoso’s second term, the enactment rate dropped for both branches; only 40 percent of the executive proposals and 20 percent of the congressional proposals were (p.80) enacted. During the first nineteen months of Lula’s presidency (January 2003 through July 2004) in the dataset, the enactment rates slightly improved: 49 percent for the executive proposals, and 24 percent for congressional proposals. Even though the enactment rates of the executive proposals pale compared to those of the Brazilian counterparts in Latin America (see other chapters in this volume), in all three periods, executive proposals fared much better than congressional proposals. Holding agenda setting powers pays off.

Agenda Setting and Gridlock in a Multi-Party Coalitional Presidential SystemThe Case of Brazil

Figure 3.2 Time for passage of legislative proposals by origin

However, a simple passage rate may mask many challenges to bill approval. Figure 3.2 illustrates how long it took for bills to pass in Congress.21 It shows that executive proposals tend to be approved much faster than congressional proposals. Still, only 15 percent of the executive bills are approved by Congress within the first three months after their introduction, and about a half stay in Congress for 840 days or longer.

Table 3.3 Percentage of roll call votes in which government position is on the winning side, 2003–2011

Lula I

Lula II

Rousseff*

Total votes (including those in which government is on the losing side)

Approval and government yes vote

Obstruction

85 (6)

50 (4)

15

Fast-track

82 (78)

76 (94)

100 (8)

227

Substantive

75 (83)

78 (154)

66 (6)

315

Rejection and government no vote

Obstruction

83 (122)

89 (149)

60(9)

347

Fast-track

75 (9)

57 (12)

33

Substantive

84 (53)

70 (54)

83(5)

146

Note: The figures in parentheses represent the number of times the government was on the winning side. The total votes column contains the total number of roll call votes in each situation for the entire period, covering all three administrations. It also includes roll call votes in which the government was on the losing side. For this reason, percentages and absolute values do not add up.

(*) It covers only one year (2011) of the Rousseff administration.

Source: Compiled by the authors based on data from the FSB Institute.

We analyze the results of roll call votes in the Chamber of Deputies to evaluate whether the government is on the winning side of most votes. Table 3.3 presents the percentage of roll calls in which the position of the government leader is on the winning side when the roll call results in the approval or rejection of the proposal or procedural request. The column “total votes” indicates the absolute number of roll call votes of each type (obstruction, fast-track, and substantive) that were approved (in the first part of the (p.81) (p.82) table) and rejected (in the second part of the table). This total includes roll calls in which the government was also on the losing side. The percentages include only values for the cases in which the government was on the winning side. These are calculated in relation to all votes of each type by administration. For example, during Lula’s first term, there were seven roll call votes on obstructionist procedures approved by the floor, and the government was on the winning side for six of them, or 85 percent of the time. That is the value of the first cell in the table. Therefore, the percentages of the columns do not add up by row. The column with total values is included to give an idea of the number of votes of each type during each administration. Hence, Table 3.3 contains information that allows us to explore the use of obstructionist moves and the government’s incidence of victory in roll call votes.

First, it is interesting to notice that very few obstructionist votes (only 15) were approved in the Chamber of Deputies. Most of these requests (347) were rejected. Nonetheless, obstructionist votes represent a third of all roll call votes (362 votes) in the entire period. The government is predominantly on the winning side when an obstructionist vote is rejected: 83 percent of the time in Lula’s first term, 89 percent of the time in Lula’s second term, and 60 percent of the time in Rousseff’s first year. Interestingly, during Lula’s first term, the government was on the winning side of 85 percent of the approved obstructionist moves, but this tendency decreases steeply in Lula’s second term and Rousseff’s first year. Still, absolute values of these votes for these years are very small. On the other hand, the government usually gets its way on fast-track moves and on substantive votes, both in approving bills and requests it favors and in rejecting those it opposes. The table thus shows that the government is usually on the winning side in all of these measures. Yet, it is also important to underline that the government’s position is on the losing side in about one in five votes despite the tremendous institutional and resource advantages discussed previously.

Agenda Setting and Gridlock in a Multi-Party Coalitional Presidential SystemThe Case of Brazil

Figure 3.3 Proportion of approved bills that received substitutive bills or floor amendments

Next, we examine our hypotheses on the role of Congress in influencing the lawmaking process. The dark-shaded bars in Figure 3.3 indicate the proportion of approved executive or congressional bills that received substitutive bills and/or floor amendments to all institutional proposals with executive or legislative origins.22 It shows high levels of congressional activity in altering legislative proposals, with 60 percent of executive proposals and 46 percent of congressional proposals receiving substitutive bills and/or amendments in the plenary. Moreover, the light-shaded bars show that over 30 percent of approved executive and congressional proposals are amended through these means.

Agenda Setting and Gridlock in a Multi-Party Coalitional Presidential SystemThe Case of Brazil

Figure 3.4 Percentage of obstructionist movements by coalition

Note: This covers only one year (2011) of the Dilma Rousseff administration.

Among the major bills (which are not shown in the figure), 43 percent of congressional proposals and nearly all (90 percent) of executive proposals (p.83) were amended. This provides strong support for H3. These data indicate that although Brazilian presidents may be quite successful in getting their proposals approved by Congress, a point demonstrated by the relative proposal approval rates, as we argued, they are far from completely dominating the legislative process. Congress amends many executive proposals and most major bills initiated by the executive. Therefore, even though Brazilian presidents may enjoy a high passage rate, they still must work with an active Congress that exerts significant influence on the content of legislation. Since many amendments occur in committees and since this measure does not capture committee amendments that were not incorporated in substitutive bills, the actual proportion of amended bills is even greater. The Brazilian Congress certainly is not a rubber-stamping institution.

Moreover, legislative obstruction is increasingly common in Brazil. Figure 3.4 indicates that the opposition is the main proposer of obstructionist moves, albeit the difference is not as substantive as one would expect, especially during the Lula administration. It is also interesting that many obstructionist moves are also born from within the governing coalition, which indicates potential trouble in its coalition management during the PT presidential administrations. Although legislative obstructionism from within the governing coalition diminished during the first year of Dilma Rousseff’s term (p.84) (2011), available evidence indicates that the president faced a series of obstructions and rebellions by members of the governing coalition in subsequent years. Thus, the analysis of obstructionist movements suggests that oppositions are not the only challenges to Brazilian presidents. In fact, many challenges also spring from the legislative coalitions they put together with the intention to facilitate governing (Hiroi and Renno 2014).

Event History Analysis of Legislative Approval

As we have discussed, how long it takes for a bill to be approved is a great concern for many. Comparatively speaking, the legislative process in Brazil is slow. Figure 3.5 shows the Kaplan-Meier estimates of survivor functions for approved bills, rejected bills, and all bills in the sample. It indicates that most approved bills have their decisions made relatively “quickly” (given the Brazilian norm); approximately 80 percent of the approved bills have their passage within the first 1,000 days. In contrast, decisions to reject involve a rather prolonged process, with less than half of the rejected bills having their (p.85) deliberations concluded within 1,000 days of their initial introduction. However, some of the approved bills also suffered protracted deliberations in Congress: 4.5 percent of approved bills passed Congress after more than 2,000 days of their journey in the institution. These data demonstrate that there is a significant variation in the timing of their passage.

Agenda Setting and Gridlock in a Multi-Party Coalitional Presidential SystemThe Case of Brazil

Figure 3.5 Kaplan-Meier survival estimates of bills in the Brazilian Congress, 1995–2004

In this section, we test hypotheses 1 (executive proposals) and 4 through 8. Cohesiveness of the governing coalition (H4) and opposition (H5) is measured by the Rice Index of Cohesion, calculated by an absolute difference in the percentages of those voting yes and those voting no in a coalition. We computed a monthly index for each of these coalitions using every roll call vote recorded in Cebrap’s legislative database for the Chamber of Deputies. We use the mean of the Rice Index during three preceding months because any particular month’s index may reflect the nature of the specific vote taken rather than the degree of cohesion in general. We call these variables government coalition unity and opposition unity, respectively.

We use cabinet coalescence rates as a proxy for coalition management by the government (H6). Cabinet coalescence rates measure the proportionality in percentages between the shares of congressional seats in the Chamber of Deputies held by the parties in government and their ministerial shares in the (p.86) cabinet (Amorim Neto 2002). When there is no correspondence between cabinet shares and legislative seats, the index takes a value of 0, and perfect correspondence between cabinet shares and legislative weights takes a value of 100. Cabinet coalescence rate is lagged by one month.

The government–opposition ideological divide is our measure of inter-coalition conflict (H7). It uses composite indices of ideological positions based on survey items generated by Zucco and Lauderdale (2011) and Zucco and Power (2009). In the original dataset, these indices are first calculated at the level of deputies and then aggregated at the party level. The ideological divide variable is the difference between the mean ideological positions of the parties that compose the governing or opposition coalition, lagged by one month. Finally, bicameral incongruence is our measure of the conflict between upper and lower houses (H8). It is measured by an absolute difference in the percentage share of seats held by the governing coalition in the upper and lower houses.

Coalition unity, cabinet coalescence rates, government–opposition ideological divide, and bicameral incongruence are all time-varying covariates. Thus, it is possible that the values of these variables change from the time of proposal initiation during the course of the deliberative process.

The analysis also includes the following control variables: judicial proposal, different majority requirements to approve a bill, whether or not bills were considered under the regime of urgency, whether substitutive bills were proposed, number of floor amendments submitted, issue salience of the bill, provisional legislation, and time to election. The number of floor amendments is logged. Issue salience is indicated by appearances in the Folha de São Paulo (see previous discussion), lagged by one month. Time to election counts the number of months until general elections. We estimate the timing of approval with Cox regression.

Estimation Results

The results of Cox regression analysis of the passage of bills are shown in Table 3.4. Variables that violated the proportional hazards assumption based on the analysis of Schoenfeld residuals were interacted with a natural logarithm of time. The results are interesting and revealing, but at the same time, we must be careful about generalizing these findings because the analysis is based on only nine years of record, mostly during the period of President Cardoso. Nonetheless, the results of the analysis support many of our hypotheses.

(p.87) (p.88) As expected, how well coalitions are managed influences the speed of legislation. The coefficient of cabinet coalescence is positive and significant, indicating that greater proportionality in the allocation of cabinet positions among coalition partners contributes to swifter bill approval. However, the negative coefficient of this variable interacted with time suggests that this effect wanes with the passage of time. In other words, proportional allocation of cabinet posts has greater influence initially, but as bills linger in Congress without decision, its effect diminishes.

Table 3.4 Event history analysis of bills’ passage in the Brazilian Congress, 1995–2004

Executive proposal

5.51***

(0.87)

Executive proposal*ln(t)

−0.78***

(0.14)

Government coalition unity

0.01

(0.01)

Opposition unity

−0.01***

(0.003)

Cabinet coalescence

0.27***

(0.06)

Cabinet coalescence*ln(t)

−0.04***

(0.01)

Ideological distance

0.16

(0.27)

Bicameral incongruence

3.66***

(1.36)

Control variables

Judicial proposal

−0.09

(2.41)

Judicial proposal*ln(t)

0.22

(0.38)

Supermajority

0.47**

(0.23)

Absolute majority

−0.16

(0.21)

Urgency

9.48***

(1.14)

Urgency*ln(t)

−1.23***

(0.18)

Substitutive bill

−1.99***

(0.68)

Substitutive bill*ln(t)

0.26**

(0.11)

Floor amendments (logged)

−0.69**

(0.28)

Floor amendments*ln(t)

0.17***

(0.05)

Issue salience

6.34***

(1.25)

Issue salience*ln(t)

−0.78***

(0.22)

Provisional legislation

1.43***

(0.23)

Month to election

−0.01*

(0.004)

N

66,134

Wald Chi-Square

708.71

Note: The estimation used Cox regression with the Efron method for ties. Entries are coefficients. Robust standard errors clustered on individual bills are in parentheses. * p<0.10, ** p<0.05, and *** p<0.01, two-tailed tests.

The analysis also reveals that a united opposition is a major hindrance to legislative approval and significantly slows the deliberation process. On the other hand, neither the cohesiveness of the governing coalition nor the ideological distance between the governing coalition and opposition has a statistically significant impact on the timing of passage. In addition, the analysis shows effects of bicameral incongruence that are contrary to our hypothesis and findings by prior research. We find that during the period under investigation, bicameral differences in the governing coalition had a positive, not negative, effect on proposal passage. One possible explanation is related to the peculiarities of the time being analyzed. The variable’s positive coefficient may be an artifact of large majorities President Cardoso enjoyed throughout his two terms. That is, the variable may be reflecting the degree of an oversized coalition in one of the houses, which is likely to help bills’ passage, rather than the effect of a bicameral divergence.

Finally, as we already know from the analysis of descriptive statistics, the Cox regression analysis confirms that executive proposals are more likely to have (p.89) faster approval than congressional proposals, demonstrating the importance of agenda setting powers. Interestingly, the analysis also shows the diminishing effect of executive proposals over time; with the passage of time, executive proposals tend to lose the advantage of being executive proposed.

To summarize, our data analysis found support for hypotheses on executive proposals (H1), coalition management (H6), and opposition unity (H5). It points to the importance of analyzing agenda setting powers and politics within coalitions.

Conclusion

In this chapter we explore how various institutional and partisan factors influence lawmaking in Brazil, a paradigmatic case of coalitional presidentialism. Using data on legislative proposals and roll call votes, we test implications of the analysis of institutional prerogatives, policy distances between major actors, and the interactions between these factors.

Our empirical analysis confirms a significant agenda setting advantage of the executive branch in advancing its legislative agenda. However, executive proposals do not pass unscarred by Congress or without the long and arduous process of bargaining and legislative obstruction. The evidence clearly indicates that the legislative branch has an appetite for amending legislative proposals, and obstruction and delay are more common than previously acknowledged.

Comparison of Brazilian presidents with other Latin American presidents indicates that the approval rate of executive proposals in Brazil is the lowest of any Latin American president examined in this volume. The difficulties of dealing with wide coalitions with actors that have very different goals and preferences severely limit what the president can get passed in the normal legislative process. It is through the use of agenda setting powers, particularly provisional presidential decrees, that the record of the Brazilian executive comes closer to those of other presidents in Latin America.

All in all, however, the Brazilian case seems as one in which the institutional mechanisms for agenda setting allow for legislative bills to be approved, permitting a certain degree of governability. Yet, this does not happen, as once thought, based exclusively on the wishes and powers of the executive branch. In fact, the institutional design of the legislative process creates opportunities for shared responsibility over lawmaking between the executive and legislative branches.

Notes:

(1.) Some scholars even claim that the allocation of positions in the cabinet indicates a shared policy platform and agenda between all members of the coalition, instead of simple patronage (Figueiredo and Limongi 1999). However, evidence of this shared agenda among coalition members has not been sufficiently demonstrated.

(2.) This excludes legislative decrees called decretos legislativos.

(3.) Strictly speaking, both the constitution and the internal rules of Congress refer to sessions. When considering deadlines, we use sessions and days interchangeably.

(4.) The government has less flexibility with constitutionally defined transfers.

(5.) See, for example, “Uso abusivo de MPs contraria Congresso,” Jornal do Brasil, January 29, 1995, p. 3; and “Congresso limitará edição de MP,” Jornal do Brasil, October 23, 1995.

(6.) During the 54th legislature, the Chamber of Deputies had twenty permanent committees with predefined jurisdictions, thirty-six special committees to deal with specific topics and constitutional amendments, and eight external committees, which followed some external event. Congress also has the prerogative to install a Parliamentary Committee of Investigation (Comissão Parlamentar de Inquérito—CPI), with a limitation of five functioning CPIs in any given year. Congress can also create a Joint Parliamentary Committee of Investigation (Comissão Parlamentar Mista de Inquérito), which congregates senators and federal deputies.

(7.) Six percent of the chamber members or leaders representing this number may request a roll call vote to the steering committee.

(8.) Rapporteurs have shorter deadlines to produce materials considered under priority or urgency regime. No extension is allowed for the materials under urgency regime. Extensions may be granted to materials considered under ordinary or priority regime for up to one-half of the original term.

(9.) In the Chamber of Deputies, conclusive power may not be used for bills of complementary law, constitutional amendments, codes (including electoral and criminal codes), bills of popular initiative, committee proposals, Senate proposals, bills amended by the Senate, bills that have divergent committee reports, and bills under urgent consideration. In the Senate, terminative power may be used in international treaties, Chamber bills that were approved by conclusive power, and other types of proposals except constitutional amendments.

(10.) Testing this hypothesis requires distinct data that we do not have at this point. We thus put this on the agenda for future research.

(11.) For exceptions, see Pereira and Mueller (2000), Ricci and Lemos (2004), Santos (2002), Santos and Rennó (2004), and Santos and Almeida (2005).

(12.) The Senate has more simplified procedures.

(13.) It would also be interesting to verify the success rate of bills with discharge petitions in comparison to bills that follow the normal course.

(p.91) (14.) Even though Brazil’s bicameral congress is in principle symmetric in the distribution of prerogatives, the bicameral relationship becomes asymmetric with respect to the processing of statutory law proposals. In such cases, the first house where the proposal is originally introduced has an overriding power over the reviewer house’s amendments (Hiroi 2008a).

(15.) Joint committees produce reports, and often bills, but these committee proposals still need to be examined separately and sequentially and are open to amendments.

(16.) The proposal data come from Hiroi (2008a) and roll call votes from Cebrap (2011). The periods of data coverage are different because these data were coded at different times. Although they cover different years, each dataset has unique information that is useful for our analysis. Moreover, our hypotheses do not require testing using data from the same period.

(17.) In other words, this selection method identifies and considers institutional proposals. Executive and judicial proposals introduced to Congress are institutional proposals because they cleared internal deliberation processes within these branches prior to their introduction to Congress. We define Chamber and Senate bills as institutional proposals when they passed the first house where they were introduced. Essentially this method establishes equivalency between executive, judicial, and congressional proposals by helping to remove the bills that were proposed simply for the sake of proposing (which many members of Congress do). See Hiroi (2008a) for detailed information on the dataset and selection procedure.

(18.) Censoring bills that are terminated for reasons other than approval basically uses the same approach as the latent survivor time approach to estimate competing risks. In the latter, the event of interest is assumed to occur eventually if time went on long enough without the “failure” from other types of event. Based on this assumption, observations are censored when cases experience other types of event. This is the approach we use in estimating the approval duration model. We also attempted to estimate a rejection model. However, due to the fewer cases of rejection, the model did not converge.

(19.) The Lula government indeed concentrated its effort on the passage of these two proposals to the extent that it led to an illegal legislative support-buying scheme called mensalão (big monthly stipend), arguably the biggest scandal involving the executive and legislative branches.

(20.) Readers are reminded that the denominator includes, respectively, all executive or congressional institutional proposals introduced to Congress during each specified period. Thus, in addition to enacted bills, it includes bills that are still pending at the end of the period as well as bills that were withdrawn, rejected, appended to other bills, totally vetoed, or terminated for other reasons.

(21.) As discussed in the previous section, the figure is based on the data coding passage as the event of interest and all other outcomes as censored.

(22.) The data include only amendments submitted on the floor. Amendments proposed in committees are reflected if they are adopted by rapporteurs in their substitutive bills.