Political Institutions and the Legislative Success of Brazilian Presidents: an Analysis of the Cardoso, Lula and Rousseff Governments^*

Abstract

It is known that Brazilian presidents are able to approve most of their own legislative bills. However, it is still unclear what factors influence and explain the variation in presidents' level of legislative success between different governments and even within an individual term. Seeking to understand this phenomenon, this article analyzes the legislative success of Brazilian presidents, based on the governments of Cardoso, Lula and Rousseff. More precisely, I examine the impact of prerogatives and exclusive policy matters on presidents and political contexts in determining legislative success. The results point to prerogatives and exclusive matters having a positive influence on success rates, such as provisional measures and administrative and budgetary matters. They also indicate that legislative success is enhanced during the honeymoon period, that is, the further away a government is from an electoral period, and there is significant positive variation as the size of the governing coalition increases. However, the results show that skill does not impact significantly on success, and popularity has a negative effect, thus not contributing to Brazilian presidents' legislative success.

Legislative success; president; approval; law

Brazilian presidentialism has long been an object of study in Political Science. Thanks to research already conducted, much is known about the mechanisms that presidents possess when seeking to garner support for implementing their policies, and about the roles of electoral rules and parties. However, several aspects of the Brazilian legislative process continue to lack detailed examination.

In Brazil, presidents are key actors in changing the legal status quo. If we look at the legislative record, that is projects that were analyzed and either approved or rejected/archived, the executive was responsible for 68.4% of the total, compared to 27% for the legislature and just 4.7% for the judiciary. Even leaving out budgetary projects (PLNs¹, which are exclusive to the Executive, it holds 55% of the total, a clear dominance over other powers in legislative production². That is to say, we can assume that the president of Brazil is the main actor within the legislative process to have measures that modify the legal status quo.

Furthermore, we can see that much of what is presented to Congress by presidents is approved. Around 70%³ of bills signed by presidents of the Republic were successful within the period analyzed. Some studies of legislative processes have analyzed the mechanisms that either facilitate or obstruct the passage of legislation, helping to explain this average percentage (FIGUEIREDO and LIMONGI, 1999; LIMONGI, 2006). However, descriptive and statistical analyses offering a more individual perspective – in this case, focusing on the presidents – have been rare. As such, there is limited understanding of the different forces that affect the legislative success of Brazilian governments. While questions regarding the formal authority of the Brazilian president are well documented (AMES, 2003; FIGUEIREDO and LIMONGI, 1999; MAINWARING, 1990; SANTOS and ALMEIDA, 2011), the extent to which these exclusive prerogatives, as well as contextual and specific variables related to legislative bills, affect the approval rate of bills originating from the Executive. Are exclusive prerogatives like provisional measures alongside budgetary and administrative matters really effective in influencing the final outcome of a bill? Furthermore, is the legislative performance of Brazilian presidents influenced by contextual issues such as their popularity, the electoral calendar, and their coalition's level of support in the legislature? Is their 'skill', as some authors have argued (ABRUCIO and LOUREIRO, 1999; AMES, 2003; STEPAN, 2000), the most important variable? These are the questions that this article aims to answer.

As such, based on the model presented by Alemán and Navia (2009), the objective is to analyze which factors weigh more heavily on the legislative success of Brazilian presidents and what explains variation in the approval rate of the bills under Brazilian presidentialism. This work contributes in four ways to studies of Brazilian presidentialism and legislative process.

First, because among the few studies on the determinants of legislative success outside the United States, a case regarded as a form of presidentialism with weak presidents⁴, this paper examines a case considered as having strong presidents, such as Chile as discussed by Alemán and Navia (2009).

Second, this work does not only observe the impact of institutions at a macro level, which is to say it does not consider the role of legislative organization in isolation⁵. This view helps us to understand legislative success in an aggregate way that considers the way in which executive-legislative relations are structured in Brazil. Contextual factors not previously analyzed in the Brazilian case are added to the institutional ones, allowing for greater micro-analysis and consideration of aspects influenced by political circumstances impacting on the passage of legislation.

Third, although both Chile, as analyzed by Alemán and Navia (2009), and Brazil are cases of 'strong presidents', there are important institutional differences between the countries. In the year in which the above article was published, Chile had a binomial⁶ electoral system whereas in Brazil the system is proportional with open lists. Thus, the logic of coalition formation varies substantially between them. In Chile, because of the greater restriction on the formation of smaller parties due to the majoritarian character of elections, there was a tendency towards the formation of electoral coalitions composed of different parties such as the center-left 'Concertación' and the center-right 'Alianza'. These coalitions were reflected in the configuration of the Chilean Congress in two large, opposed political coalitions. On the other hand, in Brazil, the proportional system generates greater party fragmentation, meaning the president's party, because it cannot reach a sufficient majority of seats, needs to seek political support with other parties, including ideologically distant ones.

Thus, unlike Chile, where government coalitions are formed exogenously from victorious electoral coalitions, in Brazil, they are formed endogenously because of the president's party's need to seek support. Thus, understanding the success of presidents in a proportional open list system enriches our understanding of countries with 'strong presidents', because the different ways the congresses of the two countries are organized, despite both being multiparty systems, generate distinct strategies for the weaving together of political support. In Chile, presidents already knew the size of the coalition that would support them as soon as the election result was announced. In Brazil, on the other hand, negotiations over political support for presidents are much more intense due to party fragmentation that is not mitigated by the electoral result.

This leads us to the fourth and final point. Also unlike Chile – which, in addition to having a binomial electoral system, is a unitary state – as well as having proportional, open-list voting, Brazil is a federal state. Thus, some authors suggest such a combination might generate incentives for parliamentarians to support presidents (ABRUCIO and LOUREIRO, 1999; MAIWARING, 1990; STEPAN, 2000), demanding a lot of political 'skill' from the latter, while others argue that the endogenous institutions of congress decrease transaction costs between the executive and the legislative⁷ (FIGUEIREDO and LIMONGI, 1999). In this sense, this article adds to the literature by examining to what extent 'skill' and institutions contribute to legislative success in cases of presidentialism with 'strong' presidents, since this question was practically limited to the American context.

The governments analyzed in this article include the period spanning Cardoso's first term to Rousseff's first term, between 1995 and 2014. This includes the presidential terms Cardoso I and II, Lula I and II, and Rousseff I. The reason for the selection of these governments lies in the fact that they represent the most politically stable period in post-'88 Brazil, which may avoid some bias caused by periods of major political turmoil, such as the impeachments of Fernando Collor and of Dilma Rousseff in 2016, events not commonplace to all governments and which affected the legislative success of these presidents. Furthermore, in institutional terms, the only change during this period was the approval of the constitutional amendment EC 32/2001 which extended the period of assessment of provisional measures from forty-five to sixty days⁸.

Following the model of Figueiredo and Limongi (1999), Propostas de Emenda Constitucional (Proposals of Constitutional Amendments, PECs) and Projetos de Lei Complementar (Complementary Law Bills, PLPs) were not accounted for in the analysis because they require a separate quorum for approval. Therefore, in this paper, Medidas Provisórias (Provisional Measures, MPVs), Projetos de Lei Ordinários (Ordinary Law Bills, PLs) and types of matter (budgetary – whether in the form of PLNs, MPVs or PLs – administrative, economic, social, political-institutional and tributes) are analyzed according to institutional variables. In contextual terms, the electoral cycle, the size of the government coalition, and the president's popularity are analyzed. There is also the personal variable, since some authors use it as an argument (ABRUCIO and LOUREIRO, 1999; STEPAN, 2000), of the president's 'skill'.

The results show that exclusive presidential prerogatives such as a provisional measures as well as budgetary and administrative matters are more likely to succeed, in this regard matching the findings of Alemán and Navia (2009) for the Chilean case. Furthermore, economic matters also have a positive influence on the legislative success of presidents. As for the contextual variables, the electoral cycle, as also found by Alemán and Navia (2009), and the size of the coalition positively influence the approval rate of bills presented by the executive. Meanwhile, the popularity of the president did not have a positive effect, also reflecting the Chilean case. Finally, the 'skill' of the president does not significantly influence presidents' legislative success when compared to the importance of institutions.

Presidents and legislative success

As is shown by a significant part of the literature on Brazilian presidentialism, the Brazilian president is considered the main political actor dictating the national legislative agenda. Such centralization around this actor has led several authors (AMES, 2003; FIGUEIREDO and LIMONGI, 1999; MAINWARING, 1990; SANTOS and ALMEIDA, 2011) to argue that it is the measures of executive that generate the most significant legal changes in the country. This is because Brazilian presidents can introduce exclusive budgetary and administrative matters, and can shorten the processing time and make it more difficult for parliamentarians to assess a particular bill using provisional measures (MPVs), which, as soon as they are submitted immediately have the force of law⁹. However, there is still not much certainty regarding how efficient exclusive matters and MPVs are as a strategy for Brazilian presidents to pass legislation.

In the United States, on the other hand, there is a long tradition of analysing the drivers of legislative success. Authors who have addressed American presidentialism generally point to factors such as: 01. 'skill'¹⁰, also called the president's 'capacity for persuasion'; 02. the role of parties, mainly on issues such as partisan control of 'legislative gates' – where the more gates occupied by the president's party, the greater the chances for legislative success; 03. the degree of polarization – whereby the level of ideological polarization between the parties is fundamental to success, and the lesser the polarization the better it is for presidents, because there is greater willingness on the part of the opposition to collaborate; 04. the president's popularity vis-à-vis the electorate – high approval of the president being an incentive for deputies seeking re-election to support the president's legislative agenda; 05. and the electoral calendar – whereby the more distant the election, the greater the chances of presidents receiving political support because the concern with re-election is not a pressing issue for deputies (BOND and FLEISHER, 1990; COHEN et al., 2013; DAVIDSON, 1984; EDWARDS et al., 1997; NEUSTADT, 1960).

Given that in terms of both institutional structures and agenda-setting power, in addition to the different party compositions of the US House of Representatives and the Brazilian Chamber of Deputies (bipartisan and multiparty systems, respectively), there are major differences between the United States and Brazil, it is difficult to apply these results found by researchers in the US context to Brazilian presidentialism. On the other hand, some studies have adapted the above factors¹¹ to the contexts of countries such as Chile (ALEMÁN and NAVIA, 2009) and Argentina (ALEMÁN and CALVO, 2008), bringing analysis of legislative success to countries with multiparty systems and 'strong presidents'. Considering the presence of factors like exclusive prerogatives and government coalition size, the results found by these authors are likely to more be comparable to the Brazilian case.

In this article, like in Alemán's and Navia's (2009) discussion of the Chilean case, our focus is on analyzing the factors that best explain the legislative success rate of Brazilian presidents and thus changes in the country's legal status quo. That is to say, the focus is on bill's arising from the initiative of the executive and looks at the factors that affect their approval rates.

In the next session, we present a deeper analysis of the theoretical literature and offer a hypothesis drawn from this and observations of the Brazilian case during the governments of Cardoso, Lula and Rousseff.

The Brazilian case: institutions, presidents and contextual effects

Brazilian presidents can be considered as 'strong', since they possess powers that can speed up the passage of bills, and can present exclusive matters (both budgetary and administrative). Such arrangements are important in order to expedite the passage of bills in a busy legislative environment.

Provisional measures (MPVs), as specified in Article 62 of the Brazilian Federal Constitution, have a period of sixty days in which they must be debated, extendable once for a further sixty-day period. If a MPV is not considered within forty-five days of its publication, it blocks the passage of further legislation in the house in which it is located (Congress or Senate) until it is voted on. In such a case, Congress can only vote on certain types of proposal in extraordinary sessions. As for exclusivity in budgetary and administrative matters, this appears in Art. 61.

Mainwaring (1990) and Pessanha (1997) attribute these prerogatives of the president as an abdication by legislators to legislate. The problem with this argument is: since the constitution is designed by legislators, what is the rationale in giving away their main function? Along these lines arises the argument of 'delegation'. Its premise lies in the idea that the legislature delegates certain functions to the executive, such as MPVs and exclusive matters, without letting the interests of legislative majorities be represented (AMES, 2003; FIGUEIREDO and LIMONGI, 1999).

The abdication argument speaks to what authors such as Stepan (2000) and Abrucio and Loureiro (1999) have argued, that in an institutional scenario permeated by veto points – such as, robust federalism and conflicts of interests in the way electoral connections between executive and legislative are established, with the first representing national and the second particularist interests – there are few incentives for political actors with different interests to cooperate. Therefore, for presidents to be successful, they must be endowed with a good dose of virtù (ABRUCIO and LOUREIRO, 1999), that is great 'political skill' to persuade dispersed political actors to support their policies. So, from the perspective of 'abdication', what would the legislators' incentives for supporting something that did not suit their interests be? Hence, from this perspective, whether presidential initiatives, exclusive or non-exclusive matters, success would tend to be highly dependent on the ability of presidents to gain political support. Thus, the success rate would show significant variation from one government to another.

From perspective of 'delegation', there are two approaches. On one side we have Ames (2003) and on the other, Figueiredo and Limongi (1999). The former seeks to understand how electoral bases determine parliamentary behavior in Brazil, that is, the role of exogenous institutions. It is important to remember that Brazil has an open electoral proportional electoral system, which according to Ames (2003), encourages personalism to the detriment of the party. The author argues that centralizing the candidate generates incentives for deputies to focus on fiscal localism (or 'pork barrel') for their bases. Conversely, presidents because of the nature of their election, have an agenda that is national. Thus, there is a tendency for deputies and presidents to have very conflictual relationships in seeking to implement policy.

In that sense, there is a dilemma. How can presidents win support for their measures and deputies also continue to pursue their parochial policies? Through ad hoc coalitions and space made for amendments in the federal budget, since the concentration of these powers in the executive and the way in which the deputies are elected means there is little incentive for deputies to act outside of this sphere. According to Ames (2003), the agreements by which presidents gain political support for their bills occurs on a case-by-case basis, party discipline is not high, and the executive is dominant in relation to the legislative, though with difficulties. Thus, the tendency in this scenario, described by Ames (2003) as high cost bargaining, is for the agenda to become decentralized with commissions being the main institution where negotiations will occur due to conflicting interests between presidents and deputies and prerogatives in budgetary bills that restrict deputies' scope for action. Thus, it is expected that the legislative success rate of presidents will be very dependent on their ability to persuade deputies to swing considerably between presidents (NEUSTADT, 1960; STEPAN, 2000). Moreover, budgetary and administrative bills tend to be approved, although becoming far removed from the presidents' preferences because of deputies' distributivism.

It is noteworthy, therefore, that although the thesis of abdication appears to have become obsolete since the emergence of the delegation thesis, the idea of ​​the president's skill persisted even though it did not appear directly in the arguments associated with the distributivist view.

In contrast to Ames (2003), Figueiredo and Limongi (1999) emphasize the importance of endogenous congressional institutions in securing legislative support for the executive. These authors argue that parties, and more precisely party leaders, play a key role in coalescing political support within the Congress. Thus the party personalism and fragmentation that are products of electoral logic are countered by the powers that internal rules confer on party leaders. The latter, while disciplining their own benches, negotiate positions in the executive and support the projects of presidents, thus avoiding the need for bargaining on a case-by-case basis. Alongside the concentration of powers in party leadership in congress, the executive holds the power to appoint political offices. Thus, both powers have incentives to cooperate since presidents hold the prerogatives already mentioned, while party leaders concentrate a lot of power and deputies pursue political status with office-seeking strategies¹².

For these reasons, government coalitions will tend to be solid, predictable and as the cost of bargaining between the executive and legislative is low, alongside a range of mechanisms in the hands of the Executive to accelerate the legislative process, the agenda will tend to organize itself in a centralized form in the assembly and party discipline will be high since the leadership whip¹³ is strong. In this analysis, the executive will tend to be dominant and will approve bills with relative ease seeing as support is stable. Thus, given that institutions encourage cooperation, helping presidents to avoid difficulties in having their bills approved, the 'skill' factors tends to be mitigated, since executive-legislative agreements are more predictable and guaranteed. In this sense, there significant variations in success rates from president to president should not occur, and there should be no major distance between the preferences of presidents and congressmen in the approving budgetary and administrative bills¹⁴.

Tables 01 and 02 and Figure 01 below present data related to the roles of institutions and of skill, based on the different perspectives discussed of the organisation of the Brazilian legislative process:

Table 01
Legislative success15 by type of initiative (Cardoso, Lula and Rousseff)

TYPE OF INITIATIVE	PRESENTED	APPROVED	% APRV
MPVs	929	782	84.1
PLs	894	349	39
TOTAL	1823	1131	62

• Source: CEBRAP Legislative Database.

Table 02
Legislative success by type of bill (Cardoso, Lula and Rousseff)

TYPE OF BILL	PRESENTED	APPROVED	% APRV
Budgetary (PLN+MPV+PL)	1898	1651	87
Administrative	572	331	58
Economic	444	321	72.3
Tribute	17	04	23.5
Political-institutional	23	07	30.4
Social	562	292	52
TOTAL	3516	2606	74.1

• Source: CEBRAP Legislative Database.

Figure 01
Rate of legislative success by government

First we should note that MPVs have a higher success rate than PLs and that budgetary and administrative matters have a higher success rate than other types, which show the use of accelerated procedures for bills and exclusive prerogatives (AMES, 2003; FIGUEIREDO and LIMONGI, 1999)¹⁶ have a positive impact on legislative success of the executive. It is also interesting to note the high percentage of economic matters approved. What perhaps helps to explain this finding is the greater informational advantage of the executive over the legislative, since it includes institutions such as the Ministries of Finance and the Central Bank (SANTOS and ALMEIDA, 2011), as well as the frequent approval of social matters initiated by deputies (AMORIM NETO and SANTOS, 2003), which ends up privileging the executive in other thematic areas, even when it does not have exclusivity¹⁷. Thus, from these tables, we have evidence that deputies regularly cooperate with presidents.

Figure 01 shows the variation in the success rate among presidents. Cardoso achieved the highest success rate during his mandates, followed by Lula and then Rousseff. However, despite the variation between them, all managed to approve more than 60% of their projects. So either all three were endowed with strong powers of persuasion or the institutions served to attract deputies to support presidents' projects.

Having highlighted the relative significance of institutional factors and 'political skill' for legislative success, I now turn to the question of contextual effects. Considering contextual effects rests on the assumption that specific political moments affect the legislative process. One such effect, which appears with particular frequency in the literature, is partisan support for the president in the legislature. In Latin American democracies, it is common for the president's party to be a minority, but this does not mean presidents' capacity to raise support through budgetary resources and other agenda-setting powers is restricted (CHEIBUB et al., 2004; ALEMÁN and NAVIA, 2009). In the Brazilian case, majority government coalitions regularly occur, showing that there are incentives for them to form and persist. This phenomenon may indicate that this is a strategy adopted by presidents in order to have greater confidence that they will able to approve their policies rather than having to rely on the uncertain support of other parties for particular bills¹⁸. Therefore, the offer of positions in the executive is a rational strategy for presidents to pursue when seeking political support from ideologically distant parties and deputies (CHEIBUB et al., 2004). Therefore, it is to be expected that presidents with majority coalitions are more likely to succeed than minority governments.

During the PT governments, this strategy became even more evident since the coalition that sustained both Lula and Rousseff was ideologically heterogeneous, formed of parties both of the left, like the PCdoB, and to right, such as the PP, among others. It seems that regardless of the ideological differences between them, due to the fact that they are constantly in the majority, Lula and Rousseff benefited from the margin of political support that these coalitions guaranteed in Congress. If Lula and Rousseff had resorted only to the support of parties ideologically close to the PT, it is likely to have had a very high cost in bargaining with other parties, reducing the chances of many of their bills being approved. The leftist parties were a minority in Congress vis-à-vis the right wing.

On the other hand, during Cardoso's two terms in office, the former president was able to count on the support of parties ideologically close to the PSDB within his coalitions. However, this was still not enough for the president to relinquish the strategy of offering positions in his executive in exchange for support. An emblematic example was the offer of a ministerial position to the PPB in 1996 so that the Pension Reform PEC would have a greater margin of support and guaranteed approval. Therefore, even with ideologically aligned parties, it is preferable for presidents to have a broad 'official'¹⁹ coalition so that there is greater assurance that their projects will be approved²⁰.

Another contextual variable that frequently appears in the literature is the electoral cycle. The idea here is that presidents benefit from a honeymoon period with congress at the beginning of their term. This is the case because, aside from the availability of institutional resources that can be distributed to supporters, recent electoral victory gives the president strong political capital, as popular approval encourages deputies to support the government's agenda. Furthermore, there is still a long time to go before the next elections, which in theory leads deputies focus on the legislative process, rather than strategies for re-election which become more significant during the final years of a legislative term. Alemán and Navia (2009) in the Chilean case and Cohen et al. (2013) in the US case find evidence that this variable has a positive effect on the approval of executive projects. However, Canes-Wrone and Marchi (2002) find no clear evidence of an electoral cycle effect.

In Brazil, the electoral cycle also appears to have a positive impact. The chart below shows the success rates of the legislative agendas of different governments between 1995 and 2014. A bill is considered successful if it is approved under a given president. For example, a project presented by Cardoso in 1995 is considered a success if it was approved by the year 2002, when he finished his second term.

Note that the closer elections get, the more the presidential success rate falls. There is a decrease in almost every electoral year (1998, 2002, 2006, 2010 and 2014), the most marked being 2010 and 2014 when success rates were below 40%. It should be noted that the latter elections were intense and troubled, especially that of 2014, when Rousseff was re-elected with just 51% of the vote. These events demanded great attention on the part of deputies seeking re-election supporting a particular presidential candidate. The low success rates of these years, especially 2014, may reflect these issues.

The final contextual variable is the president's popularity. It is expected that the greater public approval is of the president, the greater the probability that his or her bills will be approved. The reason is simple, in that deputies will seek to associate itself with a well-evaluated government because it increases their chances of re-election. In other words, the incentives to support the executive's projects are strengthened during periods of high government popularity. The literature in general does not find significant effects of this variable on presidential legislative success (ALEMÁN and NAVIA, 2009; BOND et al., 2003; CANES-WRONE and MARCHI, 2002). Only Calvo (2007) finds a positive effect of the president's popularity on his legislative success for the Argentine case.

In Brazil, Lula and Rousseff obtained popular approval rates above 50% at certain times in their governments, with Lula achieving greater than 80% in the final years of his second term²¹. However, as shown in Figure 02, it appears that these figures were not sufficient to mitigate the effect of the electoral cycle on success rates.

Figure 02
Annual variation in legislative success of the Executive (1995-2014)

Hypotheses, data and model

The points already raised, both theoretical and empirical, lead to the following hypotheses which I separate into three categories:

01. Institutional: H1. Because of the powers centralized in the Executive and the attraction that the latter exerts on deputies, MPVs are more likely to succeed than PLs; and H2. Budgetary and administrative matters are more likely to be approved than others.

02. Personal characteristics: H3. The 'skill' of the president impacts less on legislative success than institutions.

03. Contextual: H4. Because it produces a wider margin for guaranteeing support of the president, the greater the number of seats in the governing coalition, the greater the likelihood of legislative success; H5. Due to the electoral cycle and the concerns of other political actors, bills sent by presidents in their first year of office have a greater chance of success; and H6. As can be observed empirically in Brazil, the president's popularity does not impact positively upon legislative success.

For the empirical test, information was collected for each bill presented by presidents between, January 01, 1995, when FHC's first term of office began, until December 31, 2014, when Rousseff's first term ended. This period covers the governments FHCI (1995-1998), FHC II (1999-2002), Lula I (2003-2006), Lula II (2007-2010) and Dilma I (2011-2014). During the entire period, 3,516 bills were presented by these presidents, among the different types analyzed. Within this total, 929 are MPVs and 894 are PLs. Among these, 572 administrative, 444 economic, 17 tributes, 23 political-institutional, 562 social policy and 1,898 budgetary matters are recorded²². The source of this data is the Legislative Database of CEBRAP.

In order to structure the independent variables, I included: 01. the size of the government coalition, adding together the percentages of seats held by parties forming the government and considering the duration of each coalition at the time of the final outcome of a bill; 02. a dummy variable with value=01, if the president is in his first year of office, to measure the effect of the electoral cycle; 03. data on the president's popularity, based on data collected by Instituto Datafolha between 1995 to 2014, measured as the combined percentage of the categories 'Great' and 'Good'; 04. a numerical classification for the type of legislative initiative, where 01=provisional measures, 02=PLNs, and 03=draft ordinary laws; 05. dummies indicating the type of matter, using categorical variables: 'Adm' for administrative; 'Eco' for economic matters; 'Hom' for tribute; 'Pol' for political-institutional; 'Orc' for budget and 'Soc' for social; and 06. a dummy indicating each presidential term (Dilma I, FHC I, FHC II, Lula I and Lula II) to measure the influence of the personal variable through categorical variables. This dummy captures everything the other variables do not capture for each president. That is, they concern not only personal characteristics, but also the specific characteristics of each term and four-year period.

It is important to emphasize that in the case of MPVs, due to their timeframe for assessment and the fact that several are presented towards the end of the year, many are voted on only in the following year. In such cases, when there was a change of president the final result was credited to the outgoing president, because the period for deputies to debate the bill was set by the president who presented it.

To test the hypotheses, a logistic regression model with random effects was used. With this approach, I can observe variables that are specific to each bill as well as a group of variables that vary only across the different legislatures. Since the dependent variable (final result of the bill) is a binary model (0,1), I estimate the probability of a project being approved (1) or not (0) by logistic function logit (x) = 1 / 1 + e-bx, where x is a vector with the independent variables used in the model, and b its coefficients. The inverse of the logistic function (logit-1 (x) = ex / 1 + ex) allows us to express the results probabilistically.

Results

Table 03 presents the results for 04 models tested. Model 01 measures probabilities for all listed variables. Model 02 removes the PLN and PL types to compare the bills while only considering the matter in question. In Model 03, PLNs are treated the same as MPVs, so as to compare both exclusive initiatives in relation to the PLs. In Model 04, PLNs continue to be treated the same as MPVs, however, budgetary matters are removed. Independent variables appear in the first column of the table, with the regression coefficients and standard errors in parentheses on the same row.

Table 03
Legislative success of the executive in Brazil: 1995-2014

	Model 01	Model 02	Model 03	Model 04
Constant	- 1.2397* (0.5378)	- 1.6731* (0.4997)	- 1.2074* (0.5374)	- 1.4347** (0.5388)
FHC I	0.4287* (0.1908)	0.1105 (0.1773)	0.4248* (0.1907)	0.4271* (0.1908)
FHC II	0.4428* (0.2077)	- 0.0435 (0.1943)	0.4371* (0.2074)	0.4433* (0.2077)
Lula I	0.7977*** (0.1647)	0.5884*** (0.1555)	0.8001*** (0.1647)	0.7967*** (0.1647)
Lula II	1.3056*** (0.2208)	1.3102*** (0.2077)	1.3039*** (0.2206)	1.3067*** (0.2209)
Budgetary law (PLN)	- 0.3126 (0.2289)	~	~	~
PL	- 2.1918*** (0.1358)	~	- 2.1867*** (0.1355)	- 2.1948*** (0.1356)
Coalition	0.0643*** (0.0069)	0.0559*** (0.0064)	0.0638*** (0.0069)	0.0644*** (0.0069)
Popularity	- 0.0383*** (0.0061)	- 0.0463*** (0.0057)	- 0.0383*** (0.0061)	- 0.0383*** (0.0061)
Electoral cycle	2.7945*** (0.2911)	2.6335*** (0.2792)	2.7769*** (0.2909)	2.7976*** (0.2912)
Administrative	~	~	~	0.1934 (0.1634)
Economic	- 0.1895 (0.1647)	0.5481*** (0.1439)	- 0.1898 (0.1648)	0.0005 (0.1554)
Tribute	- 1.3017* (0.6132)	- 1.8624** (0.6066)	- 1.3007* (0.6128)	- 1.1101 (0.6227)
Budgetary	0.1265 (0.2561)	1.3788*** (0.1152)	- 0.1481 (0.1606)	0.3267 (0.2303)
Political-institucional	- 0.5117 (0.5329)	- 0.8535 (0.4769)	- 0.5146 (0.5328)	- 0.3202 (0.5426)
Social	- 0.6037*** (0.1424)	- 0.4593*** (0.1285)	- 0.6050*** (0.1424)	- 0.4136** (0.1561)
AIC	3008.8	3312.7	3008.7	3008.6
Residual deviance	2978.8	3286.7	2980.7	2978.6
N de observações	3516	3516	3516	3516

• Source: Elaborated by the author.
• Note: All models were constructed using logit random effects technique.
• * Codes of significance: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.01 "1.

It can be observed that, in general, the results are consistent with the hypotheses presented. When we look at the personal variables, we can see that the impact of political capacity is greater under Lula, in both the first and second mandates in all models. As for Cardoso, the coefficients indicate relative significance in relation to the base category (Rousseff), except for Model 02 in both mandates. Both Lula and Cardoso have positive regression coefficients against Rousseff. However, Cardoso has low statistical significance compared to Rousseff. Therefore, the results indicate that among the presidents analyzed, the one with greatest probability of success is Lula, and that Cardoso and Rousseff's skill are similar with a small advantage for Cardoso.

Regarding the contextual variables, it is observed that the coefficients found for the coalition are positive and statistically significant in all models. Therefore, in support of H4, there is strong evidence that the greater the number of seats occupied by the governing coalition, the greater the probability of the president's success. The electoral cycle likewise has a positive impact, with the strongest statistical significance of any variable, strongly supporting H5. The president's popularity has a negative coefficient and is statistically significant, thus corroborating H6.

Among the institutional variables, MPVs (base category) have a probability of success that is substantially higher than PLs. PLNs, by virtue of how they are classified in the data, are measured in relation to the MPVs in Model 01, and appear with a negative coefficient, but are not statistically significance. Thus, the findings support H1.

Finally, looking at the type of matter, the highest probabilities of approval were for those of a budgetary nature. The exceptions were in Model 02, which has a positive coefficient and high statistical significance, and Model 03, which has a negative coefficient, but without significance. The remainder are positive with no statistical significance in relation to the base (administrative). Tributes and social policy are those with the lowest probability of success, having statistical significance, especially the latter. Political-institutional matters have a lower chance of approval, although this is not statistically significant, while economic matters, have even less chance of being approved, though also without statistical significance. Exceptions for the latter are in Model 02, in which it appears with positive coefficient and is statistically significant, and Model 04 with a positive coefficient, but without statistical significance. These findings support H2, and we can also point out that economic matters have a good chance of success despite not being exclusive to the executive.

However, an analysis based only on regression coefficients may be insufficient, since the results are in log of the odds ratios. One way to improve the analysis is by exponentiating the estimated coefficients to describe the results of the odds ratios. Once we have done this, we get the results in Table 04 below from the coefficients generated in Model 01 (complete):

Table 04
Odds ratios

Variable / Odds ratio		Confidence interval
		2.50%	97.50%
(Intercept)	0.289	0.101	0.831
FHC I	1.535	1.056	2.232
FHC II	1.557	1.036	2.34
Lula I	2.22	1.608	3.067
Lula II	3.69	2.394	5.689
PLN	0.731	0.467	1.146
Ordinary	0.111	0.086	0.146
Coalition	1.066	1.052	1.081
Popularity	0.962	0.951	0.974
Electoral Cycle	16.355	9.243	28.941
Economic	0.827	0.599	1.143
Tribute	0.272	0.082	0.905
Budgetary	1.134	0.687	1.875
Political-Institutional	0.599	0.211	1.704
Social	0.546	0.414	0.723

• Source: Elaborated by the author.

Above, we have the results of coefficient exponentiation at 95% confidence. The odds ratios are interpreted as the likelihood of a given event occurring divided by the chances of it not occurring. With this in mind, starting with the institutional variables, we see that the probability of success for budgetary matters are 0.73 times that of a MPV. That is to say, PLNs are less likely to be approved than MPVs, but still have a high chance. Meanwhile, the chances for ordinary bills are 0.11 times those of MPVs, a low probability of success. Therefore, the results are in agreement with the coefficients, with the chances of MPVs being higher than those of budgetary matters and especially ordinary bills. This result shows that, in fact, the concentration of agenda-setting power in the hands of the presidents increases their chances of success, in keeping with the different theories of legislative organization in the Brazilian Congress (AMES, 2003; FIGUEIREDO and LIMONGI, 1999) and findings on the Chilean case (ALEMÁN and NAVIA, 2009).

As for the types of material, we see that the economic ones are 0.82 times less likely to succeed than projects of an administrative nature. Therefore, despite their weaker probability compared to the administrative matters, they still have a good chance of success, supporting the arguments of Santos and Almeida (2011) and Amorim Neto and Santos (2003). Similarly, institutional-political matters are 0.60 less likely than administrative ones to be approved, which, as well as reinforcing the argument about the greater success of exclusive matters, indicates that the chances of approval are not as low as those of economic ones. Tributes and social policy have 0.27 and 0.54 less likelihood than administrative ones, respectively. Therefore, the chances of success are considerably lower, especially for social matters. As can be seen in Graph 03 below, the variance of tributes is greater, meaning greater chances of success compared to social matters. This corroborates the arguments of Amorim Neto and Santos (2003), since deputies tend to prioritize social matters signed by their peers. Finally, budgetary matters are 1.13 times more likely to succeed than administrative ones, thus having the highest odds, but without robust differences in relation to the intercept. Again, these observations support the arguments of Alemán and Navia (2009), Ames (2003) and Figueiredo and Limongi (1999).

Graph 03
Reasons for the chances of legislative success of the executive in Brazil: 1995-2014

On the contextual factors, first we can see that each seat added to the governing coalition increases the chances of presidential legislative success by 6.6%. This result shows how majority coalitions do in fact enhance the chances of presidential success, which shows the strategy of presidents to seek support from ideologically distant parties is rational (CHEIBUB et al., 2004). For example, a coalition holding about 300 seats in the Chamber (about 60%) constitutes very favorable circumstances for the approval of presidential bills, since each of these seats represents a 6.6% chance of approval.

Popularity, by contrast, is 0.96 less likely to succeed, which, despite what the literature claims, fits with what we observe in the data on the Brazilian case and with what Alemán and Navia (2009) found for the Chilean case. Therefore, we can interpret this as meaning that in fact the popularity of the president does not weigh heavily on the success of Brazilian presidents in obtaining political support for their bills.

The electoral cycle, as observed in the analysis of coefficients, is the factor that weighs most greatly on the legislative success of Brazilian presidents. The chances of success increase by 35.5%, the further away the elections, as argued by Alemán and Navia (2009) and the wider literature.

What about the personal variable? We see that Cardoso has 1.56 higher probability compared to Rousseff in his first term and 1.54 in his second. We can thus note noted that overall Cardoso has about 1.55 greater chance of being able to approve a bill than Rousseff. Therefore, although Cardoso's success ratio is higher than Rousseff's, the difference is not great. Lula has considerably greater chances, with 2.22 greater probability than Rouseff in his first term and 3.69 in his second. This difference between the presidents was an expected result, since their capacity for persuasion certainly varies. However, there is an interesting aspect to the result. Cardoso is considered a skilled president in the literature of the 1990s (LOUREIRO and ABRUCIO, 1999), however, the results show that his chances of success are not much higher than those of Rousseff, a president considered by the press to be averse to dialogue. According to the empirical data, Cardoso also achieved a higher success rate than Lula, although according to the results it is the latter who is the most skillful. Thus, institutional variables seem to mitigate the influence of skill, and they are in fact the primary factors affecting the ability of presidents to garner political support, while others better explain variations in the rate, such as coalition size and the electoral cycle. Therefore, the predictability of political support that institutions bring to creating conditions for collective action, as argued by Figueiredo and Limongi (1999), seems to offer a more effective explanation than the argument about the difficulty of presidents to gain support in the legislature, as made by Ames (2003), and of the necessity of virtù proposed by Abrucio and Loureiro (1999).

In summary, through the statistical tests carried out using logit technique and odds ratios, we see that the exclusive initiatives – both provisional measures and budgetary and administrative matters – have high chances of approval, supporting the argument in the literature regarding the incentives for deputies to support the executive and the use of exclusive prerogatives as a means of solving collective action issues (AMES, 2003; FIGUEIREDO and LIMONGI, 1999). These results are in line with what Alemán and Navia (2009) find for the Chilean case, which strengthens the argument that institutionally strong presidents can take their agendas forward. The fact that economic matters are practically tied with administrative ones, offer empirical support to the argument about the executive's informational advantage (SANTOS and ALMEIDA, 2011). Furthermore, social matters show the lowest probability (AMORIM NETO and SANTOS, 2003), while other non-exclusive types, besides economic matters, especially the political-institutional, also have reasonable chances of success.

The issue of the electoral cycle and the size of the coalition are also verified in the Brazilian case, thus corroborating these arguments (BOND et al., 2003; CHEIBUB et al., 2004; ALEMÁN and NAVIA, 2009). Therefore, it can be assumed that the further away the elections and the larger the coalition, the greater the incentives Brazilian presidents have for presenting bills to Congress, because chances of success are higher. As far as coalitions are concerned, ideological heterogeneity does not prevent most executive projects from being approved. In fact they guarantee a margin of political support for presidents in the legislature, a phenomenon that is not an anomaly of Brazilian presidentialism.

Counter to the argument in the literature regarding popularity (BOND et al., 2003; CALVO, 2007, CANES-WRONE and MARCHI, 2002), but in line with the findings of Alemán and Navia (2009), the statistical tests did not find evidence that this variable positively impacts on legislative success. We can thus assume that this factor does not influence presidents' ability to achieve political articulation when pursuing their projects.

Finally, despite the differences among presidents with regard to skill, and consequently their probability of success, institutional and contextual variables seem to have a more significant impact. These findings challenge the argument in the literature that Brazilian presidents would have a high bargaining cost for attaining support (ABRUCIO and LOUREIRO, 1999; AMES, 2003; STEPAN, 2000). The fact that Cardoso has about 55% greater chance of success than Rousseff indicates that there is no such a robust difference between them. On this factor Lula stands out, and yet when we look at the empirical data we see that Cardoso was able to approve more bills than he was. Thus, despite differences in the way each president negotiated with his or her support base, institutions have a key role in ensuring support for the executive, meaning that less-skilled presidents like Rousseff can also approve most of their bills. Therefore, the results support the argument that institutions encourage predictable political support (FIGUEIREDO and LIMONGI, 1999) rather than those about the need to form ad hoc coalitions and virtù (ABRUCIO and LOUREIRO, 1999; AMES, 2003).

Discussion

This article analyzes a set of factors that influence the legislative success of presidents in Brazil, focusing on the Cardoso, Lula and first Rousseff governments. Evidence was found that exclusive prerogatives lying with the president are important to legislative outcomes, in light of the remarkable success of the presidents in approving provisional measures compared to ordinary laws, and administrative and especially budgetary matters in comparison to other types. These results are in line with what Alemán and Navia (2009) found in the Chilean case and reinforces the argument that institutionally strong presidents can carry out their projects through the strong agenda-setting powers they possess. Alongside this, the way presidents manage to garner political support has been tested from two perspectives: one focusing on the ability of presidents and another on the role of institutions. It has been found, in the statistical results as well as through observation of empirical data, that institutions tend to outweigh the importance skill, although, as expected, there are differences in the persuasiveness of the three presidents. Thus, in addition to the exclusive prerogatives, the endogenous rules of the Legislative, as pointed to by Figueiredo and Limongi (1999), are fundamental for the predictability and guarantee of political support for the presidents.

The empirical data also show that the larger the governing coalition is, that is, the less the president depends on the opposition's votes and the distribution of preferences in Congress, the greater the chances of success. Therefore, the presidents, by increasing their base of support, act rationally, because they gain a margin of support that makes them less dependent on the votes of the opposition. On the other hand, there is no evidence that the president's popularity positively influences his or her success. The strategies and decisions taken by the political actors involved in the legislative process do not seem to take into account the popularity of presidents over the electorate. An example of this is Lula in 2010, an electoral year, who having reached an approval rating of 83% still had a low success rate in passing legislation.

There are strong indications that the electoral cycle greatly influences legislative success. Parliamentarians tend to support the government more strongly at the beginning of their terms, with their support tending to decrease according to the advance of the electoral calendar.

Finally, I believe that this study points to new directions for future research. A study that focuses the substance of the law would greatly enrich this deabte (BARRETT and ESHBAUGH-SOHA, 2007). An analysis of the influence of the urgency requirement in the executive and legislative on the likelihood of success, when the president and congress make use of it, would also be enriching.

Bearing in mind that coalition size matters, a study that also considers how the ideological heterogeneity of government coalitions affects legislative success could also make an interesting contribution. Lula and Rousseff had more ideologically heterogeneous coalitions than Cardoso, and this fact may have been fundamental in allowing greater space for bills originating from congress. Almeida (2017) points out a reversal in this regard beginning in 2005 and this factor may be important in explaining why Lula and Rousseff's legislative success rates, although generally high, were lower than Cardoso.

Finally, understanding the factors that influence the approval of the legislative measures of the Executive allow us a deeper understanding of the functioning of presidentialism. In this paper, I have sought to contribute to this line of study by testing a series of hypotheses proposed in the literature on the subject, drawing especially on the work of Alemán and Navia (2009) on the Chilean case, a country which like Brazil has institutionally strong presidents, and using data on the Brazilian case.

References

• ABRUCIO, Fernando Luiz and LOUREIRO, Maria Rita (1999), Política e burocracia no presidencialismo brasileiro: o papel do Ministério da Fazenda no primeiro governo Fernando Henrique Cardoso. Revista Brasileira de Ciências Sociais Vol. 14, Nº 41, pp. 69-89.
• ALEMÁN, Eduardo and CALVO, Ernesto (2008), Analyzing legislative success in Latin America: the case of democratic Argentina. In: New voices in the study of democracy in Latin America Edited by O'DONNELL, Guillermo; TULCHIN, Joseph S., and VARAS, Augusto. Washington: Woodrow Wilson International Center for Scholars. pp. 07-38.
• ALEMÁN, Eduardo and NAVIA, Patricio (2009), Institutions and the legislative success of 'strong' presidents: an analysis of government bills in Chile. The Journal of Legislative Studies Vol. 15, Nº 04, pp 401-419.
• ALMEIDA, Acir (2017), Delegação, representação e atividade legislativa: explicando o novo protagonismo do Congresso. Paper presented in the 41º Encontro Anual da ANPOCS. Caxambu.
• AMES, Barry (2003), Os entraves da democracia no Brasil Rio de Janeiro: Editora FGV. 409 pp..
• AMORIM NETO, Octavio and SANTOS, Fabiano (2003), O Segredo Ineficiente Revisto: O que Propõem e o que Aprovam os Deputados Brasileiros, Dados Vol. 46, Nº 04, pp. 661-698.
• BARRETT, Andrew W. and ESHBAUGH-SOHA, Matthew (2007), Presidential success on the substance of legislation. Political Research Quartely Vol. 60, Nº 01, pp. 100-112.
• BOND, Jon R. and FLEISHER, Richard (1990), The president in the legislative arena Chicago: University of Chicago Press. 267 pp..
• BOND, Jon R.; FLEISHER, Richard, and WOOD, Dan B. (2003), The Marginal and Time-Varying Effect of Public Approval on Presidential Success in Congress. Journal of Politics Vol. 65, Nº 01, pp. 92-110.
• CALVO, Ernesto (2007), The responsive legislature: public opinion and law making in a highly disciplined legislature. British Journal of Political Science. Vol. 37, Nº 02, pp 263–280.
• CANES-WRONE, Brandice and MARCHI, Scott de (2002), Presidential approval and legislative success. The Journal of Politics Vol. 64, Nº 02, pp. 491-509.
• CHEIBUB, José Antonio; PRZEWORSKI, Adam, and SAIEGH, Sebastián M. (2004), Government coalitions and legislative success under presidentialism and parliamentarism. British Journal of Political Science Vol. 34, pp. 565-587.
• COHEN, Jeffrey E.; BOND, Jon R., and FLEISHER, Richard (2013) Placing presidential- congressional relations in context: a comparison of Barack Obama and his predecessors. Polity Vol. 45, Nº 01, pp. 105-126.
• COX, Gary W. and MCCUBBINS, Mathew D. (1993), Legislative Leviathan: Party Government in the House. California: University of California Press. 342 pp..
• DAVIDSON, Roger H. (1984), The presidency and the congress. In: The presidency and the political system Edited by NELSON, Michael. Washington: CQ Press. pp. 363-391.
• EDWARDS III, George C.; BARRETT, Andrew, and PEAKE, Jeffrey (1997), The legislative impact of divided government. American Journal of Political Science Vol. 41, Nº 02, pp. 545-563.
• FIGUEIREDO, Argelina and LIMONGI, Fernando (2005), Processo Orçamentário e Comportamento Legislativo: Emendas Individuais, Apoio ao Executivo e Programas de Governo. Dados Vol. 48, Nº 04, pp. 737-776.
• FIGUEIREDO, Argelina and LIMONGI, Fernando (1999), Executivo e legislativo na nova ordem constitucional Rio de Janeiro: Editora FGV. 231 pp..
• KREHBIEL, Keith (1992), Information and legislative organization University of Michigan Press. 326 pp..
• LIMONGI, Fernando (2006), A democracia no Brasil: presidencialismo, coalizão partidária e processo decisório. Novos Estudos - CEBRAP Nº 76, pp. 17-41.
• LOUREIRO, Maria Rita and ABRUCIO, Fernando (1999), Política e Burocracia no Presidencialismo Brasileiro: o papel do Ministério da Fazenda no primeiro governo Fernando Henrique Cardoso. Revista Brasileira de Ciências Sociais. Vol. 14, Nº 41, pp. 69-89.
• MAINWARING, Scott (1990), Presidentialism, multiparty systems and democracy: the difficult equation. Working paper Nº 144. Notre Dame: Kellogg Institute.
• MAYHEW, David R. (1987), The electoral connection and the Congress. In: MCCUBBINS, Matthew D.; SULLIVAN, Terry (Ed.). Congress: structure and policy. Cambridge University Press. 576 pp..
• NEUSTADT, Richard (1960), Presidential power: the politics of leadership. Nova York: The New American Library. 286 pp..
• PEREIRA, Carlos and MUELLER, Bernardo (2002), Comportamento estratégico em presidencialismo de coalizão: as relações entre executivo e legislativo na elaboração do orçamento brasileiro. Dados Vol. 45, Nº 02, pp. 265-301.
• PESSANHA, Charles (1997), Relações entre os poderes executivo e legislativo no Brasil: 1946-1994. Doctoral thesis Faculdade de Filosofia, Letras e Ciências Humanas. São Paulo: USP.
• RAILE, Eric D.; PEREIRA, Carlos and POWER, Timothy J. (2011), The executive toolbox: building legislative support in multiparty presidential regime. Political Research Quarterly Vol. 64, Nº 02, pp. 323-334.
• SANTOS, Fabiano and ALMEIDA, Acir (2011), Fundamentos informacionais do presidencialismo de coalizão Curitiba: Editora Appris. 216 pp..
• STEPAN, Alfred (2000), Brazil's decentralized federalism: bringing government closer to the citizens? Daedalus Vol. 129, Nº 02, pp. 145-169.

• 1
  Projetos de Lei do Congresso Nacional (Bills of the National Congress, PLN) are bills that involve budget directives.
• 2
  All data were obtained through the online portal of the Núcleo de Estudos Comparados e Internacional (Group for Comparative and International Studies, NECI) at the University of São Paulo. The basis for the site is the Legislative Database of the Centro Brasileiro de Análise e Planejamento (Brazilian Center for Analysis and Planning, CEBRAP).
• 3
  Source: CEBRAP Legislative Database.
• 4
  The US is considered to be a case of weak presidents because the latter do not possess substantive prerogatives capable of manipulating the calendar for presenting bills or exclusive matters.
• 5
  Theories of legislative organization help to understand how collective action between parliamentarians and presidents is structured. They are divided into 3 approaches: distributivist (MAYHEW, 1987), informational (KREHBIEL, 1992) and partisan (COX and McCUBBINS, 1993). In Brazil, despite differences in ways that political support is exchanged, in general, the different visions emphasize the significant weight held by the executive in attracting the support of parliamentarians (AMES, 2003; FIGUEIREDO and LIMONGI, 1999; RAILE et al., 2011; SANTOS and ALMEIDA, 2011).
• 6
  After the electoral reform of 2015, Chile adopted a proportional electoral system using open lists.
• 7
  Throughout the article, I examine this topic in greater depth.
• 8
  In the results presented here, it is noted that even with this amendment the probability of a provisional measure being approved remains high.
• 9
  The validity period for an MPV is sixty days, which can be extended once for a further sixty days. If it is not approved within 45 days of its publication, the MPV blocks the entry of new votes into the house in which it is located (congress or senate) until it has been voted on. In such a case, the congress can only vote on certain types of proposal in an extraordinary session. For more details about MPVs. See: http://www2.camara.leg.br/comunicacao/assessoria-de-imprensa/medida-provisoria, accessed on October 9, 2017.
• 10
  'Skill', or the ability to garner political support, can be understood as an individual characteristic possessed by presidents. Neustadt (1960) argued that in the United States, in order to gain support presidents must be highly persuasive. Therefore, this argument is based on the need for negotiation between the president and members of congress on a case-by-case basis, so the cost for support is high and there is no guarantee that it will in fact be sustained.
• 11
  Except for 'skill'.
• 12
  The idea of 'office-seeking' lies in the assumption that deputies seek to hold political office as a way of gaining power to carry forward their political agendas.
• 13
  Referring to the ability of leaders to discipline their teams.
• 14
  Regarding the budget issue, Pereira and Mueller (2002) argue that there is a conflict between the executive and legislative and that the distribution of individual amendments to the budget solves this question. For Figueiredo and Limongi (2005), there are no conflicting interests. These authors argue that amendments made by deputies complement the law issued by the executive. This is able to channel the demands of deputies and accommodate them within the programs they define as priorities. The executive's control of this process is such that not even opposition deputies can propose amendments that represent an alternative agenda to that proposed by the government.
• 15
  Legislative success is measured as the rate of approved bills relative to unapproved bills.
• 16
  Recalling that from the distributivist point of view deputies behave reactively, approving executive projects by diluting their preferences through pork barrel deals (AMES, 2003). On the other hand, from the party perspective, the executive's projects are negotiated ex-ante via deliberation in congress between the party leaders and the presidents, the former disciplining their party benches to vote according to the agreed position. Therefore, there is no distance between executive and legislative preferences in the final result of the bill (FIGUEIREDO and LIMONGI, 1999). In spite of the differences between these perspectives, in both, the executive is a strong political actor and is able, with greater or lesser difficulty, to pass most of its bills.
• 17
  Amorim Neto and Santos (2003) argue that approved projects that are of a social nature tend to be more linked to deputies than to the executive. This is due to the executive's tendency to focus on broader and more aggregated issues indirectly affecting the lives of citizens. The deputies act to fill this gap and by mostly proposing and approving bills that affect the daily life of the average citizen.
• 18
  On the issue of minority governments, some studies such as Cheibub et al. (2004) point out that such governments do not necessarily pose a risk to democracy and demonstrate that at certain moments presidents may form minority coalitions without democracy being put in risk.
• 19
  The criterion for a coalition to be considered 'official' is that positions are occupied within the executive.
• 20
  All data on coalitions are shown in Table 04.
• 21
  Source: Datafolha.
• 22
  Among the budgetary matters, 1,693 are PLNs and 205 are distributed between MPVs and PLs. When presidents present budgetary matters in the form of MPVs and PLs, they are in order to add credits to the existing budget law. For the purposes of analysis I considered all as 'Orc', because this matter lies exclusively with the president and the theme is the same.

Publication Dates

• Publication in this collection
  13 May 2019
• Date of issue
  2019

History

• Received
  16 Mar 2018
• Accepted
  19 Jan 2019