Until the early 2000s, dominant interpretations of how multiparty presidentialism functions focused primarily on the legislative powers of the president. Agenda-setting authority, control over cabinet formation and political appointments, and dominance over the budgetary process were identified as the key mechanisms driving presidential success in legislative arena (AMORIM NETO, 2006; CHEIBUB, 2007; FIGUEIREDO and LIMONGI, 2001). High levels of control over the legislative process, combined with strong party discipline among pro-government parties, were seen as sufficient to ensure governability. The National Congress approved, 78% of bills initiated by the Executive, between 1989 and 1998 (FIGUEIREDO and LIMONGI, 2001, p. 102). This high approval rate left little doubt about the president’s ability to consistently advance their legislative agenda – a finding reaffirmed by more recent studies (PINHEIRO et al., 2024).

As the debate evolved, it branched out in several directions. Building on the initial framework, research developed primarily along two main lines. The first focused on coalition formation and coordination processes, as well as the composition of the president’s allied party forces. This line of inquiry examined factors such as the size of party delegations (INÁCIO, 2018); the ideological range within the coalition (ARAÚJO et al., 2018; GAYLORD and RENNÓ, 2015); the flexibility of the majorities formed (FIGUEIREDO, 2012; PRAÇA et al., 2012); and the negotiation structures shaped by the ‘formateur’ (AMORIM NETO, 2006; ANSOLABEHERE et al., 2005; WARWICK and DRUCKMAN, 2001). The second strand of the literature explored the command structures, within the presidency and executive cabinet. Including studies on ministerial appointments (AMORIM NETO and SAMUELS, 2010); the prominence of different ministries (ZUCCO JR. et al., 2019); budget management within executive agencies (BATISTA, 2017; FIGUEIREDO and LIMONGI, 2008); the appointment of commissioned officials (BATISTA and LOPEZ, 2021; PRAÇA et al., 2012); and the expansion and contraction of presidential offices (INÁCIO and LLANOS, 2015). Both lines of research have significantly advanced our understanding of presidential systems.

Building on that initial line of research, the focus shifted toward the internal dynamics of governing coalitions, highlighting the direct ties between the Executive and Legislative branches. In this context, the formation of coalitions and the relationships among their member parties came to be seen as essential for advancing the president’s agenda within a multiparty presidential system (e.g., FREITAS, 2016a, 2016b). Whereas earlier discussions emphasized conflict between the Executive and Legislature, attention had now turned to the mechanisms through which these two branches interact and cooperate.

In this context, when the literature states that the government forms a coalition to govern, it implies a power-sharing arrangement among the parties involved. The Executive’s dominance in lawmaking should not be attributed solely to the president’s so-called legislative powers (FREITAS, 2016a, p. 17). Rather, the Executive and Legislature operate through an integrated process. Competing interests adjust and manifest throughout the various stages of the legislative process: from the drafting of bills by Ministries – each headed by different parties – to committee review, where rapporteurs can significantly amend proposals, and finally to the plenary vote. This pathway offers multiple opportunities for parties to exert influence.

Incorporating the legislative process into the analysis shifts the focus from presidential success to the role of the coalition in shaping the Executive’s legislative output. The coalition not only participates in drafting legal proposals but also possesses both the motivation and the capacity to intervene. By holding agenda-setting power and the prerogative to make changes, coalition members gain a distinct advantage. Parties that support the president tend to modify more provisions, underscoring their greater influence over the final content of approved bills (FREITAS, 2016a; 2016b). However, this influence does not guarantee that the coalition has sufficient votes to pass the legislation it supports. Simply knowing whether the coalition holds a majority is not enough to conclude that enacted laws fully reflect the Executive’s policy preferences. For instance, bills proposed by the Executive are sometimes vetoed (FREITAS, 2016a, p. 104), clearly demonstrating that the legislative process itself plays a decisive role.

In similar vein, the size of each coalition can be assessed in relation to the voting thresholds required to approve legislation in the Chamber of Deputies. Drawing on data from NECI’s Legislative Database covering 1995 to 2016, Graph 01 shows the gap between the number of votes held by twenty presidential coalitions – measured at the beginning and end of their terms under ideal conditions – and the number of votes needed to pass legislation: an absolute majority (257 votes) and a qualified majority (308 votes). In other words, Graph 01 estimates the potential autonomy of presidential coalitions at their peak, assuming full party discipline and unanimous support from all coalition members in favor of the Executive’s agenda.

Graph 01 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

The results are clear: in most cases, the coalitions formed during the period did not command an absolute majority. Across the six presidential terms analyzed, it was only during Lula’s second term that the government consistently relied solely on coalition votes to pass its agenda. Presidents lacked sufficient votes to guarantee full legislative autonomy in more than half of the coalitions (12 out of 20). This shortfall affected both absolute and qualified majority thresholds in five other instances – FHC II (02), Lula I (01), Dilma II (3), Dilma II (04), and Dilma II (05). In the other seven cases, coalitions fell short only with respect to qualified majority thresholds – FHC (01), Lula I (03), Lula I (04), Dilma I (01), Dilma I (03), Dilma II (01) and Dilma II (02). More precisely, except for Lula’s second term, the data suggest that unrestricted governability through the coalition was only attainable at certain moments during presidential terms – FHC I (01), FHC I (02), Lula I (02), Lula I (05), and Dilma I (02) – and only under the assumption of full party discipline among all coalition members. Existing theories that explain why presidents form minority governments (FIGUEIREDO et al., 2012) do not fully account for this variation, as they rely on institutional factors, such as presidential veto power, which remained constant across all the cases analyzed. Figueiredo et al. (2012) suggest that party fragmentation negatively correlates with the likelihood of minority governments: the greater the number of parties, the stronger the incentive to form a majority coalition.

Given the coalitions’ intermittent autonomy throughout presidential terms – and the persistence of institutional factors used to explain the formation of minority governments – the literature on coalition presidentialism gains further nuance: how do temporary majorities translate into lasting governability? Two possibilities emerge. Either, presidents advance their agendas only when the timing is favorable. Or there exists a consistent and complementary flow of votes from parties outside the formal coalition that helps secure legislative backing for the government.

Despite broad theoretical consensus regarding the process of coalition formation, the literature still lacks clear frameworks for understanding the bargaining dynamics that unfold throughout the legislative process – both within the coalition and in interactions with parties outside it (FREITAS, 2016a, p. 97). For instance, it remains unclear what determines the ‘price’ of opposition party support or how the diversity of their demands influences the formation of legislative majorities (PERES and BEZERRA, 2020). Most existing studies have focused on the powers exercised by the president within the system of government and on the internal dynamics of how the coalition operates (AMORIM NETO, 2006; CHEIBUB, 2007; FIGUEIREDO and LIMONGI, 2001; SANTOS, 2003, 2002).

Despite significant theoretical advances over the past two decades, the literature on coalition presidentialism continues to struggle with accounting for the behavior of individual parties, both inside or outside the governing coalition. While the Executive’s central role in the legislative process is well established, the effects of multiparty competition are still largely neglected. Variations in party behavior, as well as the diversity and specificity of the government’s agenda, are rarely incorporated into the strategy of coalition agreements. Instead, these actions are typically guided almost exclusively by presidential preferences and institutional powers. Based on assumptions made by Executive actors, the areas of interest and the political costs of pursuing the government’s agenda are treated as constant, and party alliances are assumed to remain stable over time (AMORIM NETO, 2006). The preferences and demands of parties in negotiations with the Executive are generally peripheral in most theoretical models, except insofar as they relate to ministers’ capacity to draft legislative proposals submitted to Congress (e.g., ARAÚJO, 2017).

To account for party interests, Figueiredo and Limongi (2001) and Cheibub (2007) incorporated competitive interactions among party actors – particularly party leaders – into their analyses, examining how the legislative process and agenda are shaped through Executive–Legislative relations. These pioneering approaches laid the foundation for subsequent research that emphasized legislative procedures and the dynamics of party negotiations with the Executive, both in forming parliamentary majorities and in shaping distinct policy agendas (FREITAS, 2016a, 2016b; SANTOS, 2003, 2002).

Among the few studies that have moved in this direction, findings reveal that coalitions play a key role in modifying legal provisions throughout the legislative process. First, it is important to emphasize the legislature’s capacity to intervene and amend bills under consideration: during the FHC, Lula I and II administrations, the legislature modified 10,729 provisions, compared to 16,878 amended by the Executive (FREITAS, 2016b, p. 15). This evidence underscores not only the significance of the legislative process as a whole but also the legislature’s substantial influence over lawmaking. Furthermore, on average, 36% of all approved amendments in enacted bills originated within the legislature (FREITAS, 2016b, p. 21). In other words, the legislature actively modifies legislation, and these modifications become law.

A closer look reveals that 82% of legislative amendments originate from coalition parties (FREITAS, 2016a, p. 95). This leads to two important observations: first, parties outside the coalition can also influence the process and introduce changes during the legislative procedure; second, different parties may behave differently throughout the legislative process.

These reflections on incorporating the legislative process open up avenues to a more nuanced understanding of party behavior in the legislature. If parties outside the coalition also have – even if limited – the capacity to amend bills, it becomes crucial to examine how they act when proposals reach the floor. Moreover, a deeper understanding of party bargaining dynamics is necessary, including the individual demands and strategies of each party. How do parties decide whether to join the coalition, and how do they vote on specific proposals? The impact of party diversity on agenda success within coalition presidentialism remains largely unexplored. In practice, the literature often assumes the feasibility of coalition presidentialism without factoring in party characteristics – whether parties are many or few, new or established, ideological or non-ideological – and presumes that the president is capable of forming coalitions and managing parliamentary majorities to pass their agenda.

In these analyses, these actors are often cast as secondary players: while they are part of the broader governance framework in multiparty presidential systems, parties and their leadership are largely seen as supportive of the Executive. With few exceptions, parties – regardless of their ideological leanings or size – are treated as organizations capable of self-organizing to maintain the legislative discipline observed in their behavior. From this perspective, party strategies are assumed to align around shared interests, and the cohesive behavior of party caucuses is viewed as merely a byproduct of the concentration of legislative power in the hands of party leaders. However, the lack of strong organizational structures and the pervasive personalism within parties have made it difficult for party leaders to exert firm internal control over their legislators’ behavior (AMES, 2001; FREIDENBERG and LEVITSKY, 2006; SAMUELS and ZUCCO JR., 2014). While studies such as Figueiredo and Limongi (2001), Cheibub (2007), and others acknowledge the role of party leaders in managing their parties through procedural authority in the legislature, they tend to largely overlook the influence of parties outside the coalition in shaping legislative majorities. As shown in Graph 01, the Executive’s formal alliances often fall short of securing the necessary votes in the Chamber of Deputies. While the Executive’s high legislative success rate is undeniable, in nearly every case, governing coalitions lack full self-sufficiency.

It is worth emphasizing that this article does not reject the established theories regarding the concentration of legislative powers and the president’s central role in managing coalitions (AMORIM NETO, 2006; CHEIBUB, 2007; FIGUEIREDO and LIMONGI, 2001). On the contrary, these are foundational premises and serve as the starting point for our analysis. Our goal is to incorporate the missing link in this debate: the impact of the legislative behavior of parties outside the coalition on the performance of the Executive’s agenda – thus offering a refinement of the existing model. In short, the core research question guiding this study is: how do parties, both within and outside the coalition, influence governability under Brazilian presidentialism?

At this stage, a significant empirical gap remains to be addressed. There is still a gap in relation to the adequate tools to fully capture the distinct legislative behavior of individual parties within the lawmaking process. It is essential to examine their actions in parliament on a case-by-case basis. To that end, we have developed a tool designed to measure the influence of both coalition and non-coalition parties on the advancement of presidential agendas: the Legislative Performance Index (LPI).

A broad array of theoretical frameworks and analytical perspectives exists on what enables the formation of an effective and functional government in Brazil (AMORIM NETO, 2006; CHEIBUB, 2007; FIGUEIREDO and LIMONGI, 2001; SANTOS, 2003, 2002). However, the types of empirical evidence typically used to support these arguments tend to fall into three main categories (AMES, 2001, AMORIM NETO, 2006; FIGUEIREDO, 2012; FIGUEIREDO and LIMONGI, 2001; INÁCIO, 2006; LAAKSO and TAAGEPERA, 1979; MAINWARING, 1993; PINHEIRO et al., 2024). The first includes metrics related to party system fragmentation, such as the effective number of parties, fractionalization, electoral volatility, and the territorial distribution of votes. The second focuses on the composition of presidential cabinets, using indicators such as coalition rates and the partisan makeup of the cabinet. The third encompasses measures of executive legislative success, such as party discipline, cohesion, legislative success rates, and dominance rates. While these indicators are valuable for analyzing the positions of political actors, coalition formation, and presidential powers, they do not directly capture the formation of parliamentary majorities – that is, the actual vote counts that determine legislative outcomes.

With this in mind, and aiming to contribute to the analytical tools developed in the literature on governability in Brazilian politics, we propose a new index focused on the legislative performance of presidential agendas during roll-call votes. The objective is to pinpoint the sources of votes that enable Executive-backed proposals to meet the various voting thresholds in the Chamber of Deputies (simple, absolute, and qualified majorities). The Legislative Performance Index (LPI) is based on three variables: party size, party discipline, and the required voting threshold. It is calculated in four steps.

Step 01 involves calculating the legislative performance of party p in each roll-call vote ‘I’ related to the presidential agenda:

For each party ‘p’ in a given roll-call vote ‘i’, the number of its members in Congress is multiplied by the level of party discipline observed. This yields the exact number of votes the party contributed in support of the Executive’s agenda. The proportion of support each party ‘p’ provided in that vote is obtained by dividing that number by the vote threshold required for ‘i’ – which varies depending on the type of legislation. For example, in the vote on PEC 637/99, held on March 18, 1999, 92 of the 99 PSDB members of Congress (93%) and 46 of the 59 PT members of Congress (78%) followed the government leadership’s recommendation to approve the bill. Thus, in this qualified threshold vote, the PSDB’s legislative performance accounted for 30% of the 308 votes needed, while the PT’s accounted for 15%.

This step makes it possible to assess each party’s relevance in supporting the Executive, as the data immediately show the relative contribution of each party in a given roll-call vote.

With this information, it is possible to aggregate the results to evaluate the collective support of coalition parties for the Executive’s agenda. In this study, Step 02 involves summing the legislative performance scores of all coalition parties in the votes analyzed, yielding the coalition’s overall legislative performance in vote ‘i’. This results in the Coalition’s Legislative Performance Index:

Step 03 applies the same process to parties outside the coalition, calculating the Legislative Performance of the Opposition (LPO) in vote ‘i’, which yields the following index:

Finally, Step 04 summarizes the LPI for the presidential term by summing the Legislative Performance Indices of the Coalition (LPIC) and the opposition (LPIO) for each vote ‘i’:

The indicator ranges from 0 to 02 for roll-call votes requiring a simple or absolute threshold (such as PL, PLC, and MPV — respectively, Bills of Law, Complementary Bills of Law, and Provisional Presidential Decrees), and from 0 to 1.7¹ for roll-call votes requiring a qualified threshold (such as Constitutional Amendments, or ‘PECs’ in the Brazilian Portuguese acronym). To simplify interpretation, the number of votes required for approval is normalized to 01. Thus, in both cases, an LPI value of 01 or higher indicates that the government’s legislative performance in a given vote surpassed the threshold needed to pass the Executive’s agenda. Unlike other measures of governability, this index captures the day-to-day dynamics of the Chamber of Deputies’ plenary activity, factoring in parliamentarians’ attendance, the different thresholds required for various types of votes, and fluctuations in party discipline during each roll-call vote. Importantly, unlike the criteria used by Figueiredo and Limongi (2001), we do not consider obstruction – whether by party leaders or individual parliamentarians – as an indicator of disciplined behavior, since deputies engaged in obstruction are formally excluded from the threshold count in roll-call voting sessions².

In addition to accounting for the daily workings of the Brazilian legislature, aggregating this index across all roll-call votes during a presidential term allows for a precise measurement of each party’s contribution to the President’s legislative performance – that is, the kinds of majorities that can be formed to ensure governability. The indicator reveals the relative weight of each group’s votes in the passage of individual bills. In doing so, the index effectively captures the legislative dynamics of both the coalition and the opposition in action.

To deepen the theoretical discussion surrounding governability data, it is useful to incorporate a typology of the majorities formed under coalitional presidentialism into the index scale, drawing on key concepts from the literature. This typology is presented in Figure 01.

Figure 01 Source: Elaborated by the authors.

As shown in Figure 01, four classic theories of coalitions are associated with four corresponding ranges of the Legislative Performance Index: minimally winning coalitions (GAMSON, 1961), majority governments (AMORIM NETO, 2006), minority governments (AMORIM NETO, 2006), and decision-making paralysis (SANTOS, 1986). While each of these four concepts drawn from the literature carries a degree of complexity and analytical depth that extends beyond the scope of this study, we offer a brief exploration of the analytical contributions they provide when integrated into the Legislative Performance Index.

The theory of minimally winning coalitions, a classic concept from the parliamentary literature (GAMSON, 1961), is used here to describe parliamentary majorities that operate within narrow legislative performance margins: coalitions that secure between 100% and 110% of the votes required to pass the Executive’s agenda. As Gamson (1961) argues, this model of governability aims to optimize the cost-benefit balance in the bargaining process between the Executive and the Legislature by conceding as little as possible in patronage practices while still ensuring effective majorities. This approach allows for the best of both worlds: broad control over the cabinet and a successful approval of the legislative agenda.

Majority and minority governments refer to the presidential strategy framework developed by Amorim Neto (2006). According to the author, both forms of governance depend on the President’s policy priorities and the institutional tools available to secure legislative majorities in Congress. When advancing a legislative agenda is not a primary concern and the President’s unilateral powers are deemed sufficient, coalition-building often results in a minority government. Within the framework of the Legislative Performance Index proposed here, this form of governability corresponds to a performance range between 31% and 99% of the votes required to approve a roll-call vote. Conversely, when legislative productivity is a central priority and patronage practices guide presidential strategy, coalition-building tends to aim at forming supermajorities. Translated into the scale of the Legislative Performance Index, this corresponds to scenarios of majority-based governability, where the President commands more than 110% of the votes needed to pass the Executive’s agenda.

The final kind of governability captured by the index corresponds to government failure, or decision-making paralysis (SANTOS, 1986). This scenario arises when a significant number of roll-call votes during a presidential term receive only minimal legislative support – defined in the Legislative Performance Index as 30% or less of the votes required to pass the Executive’s agenda. Adapted from the classic work ‘Sixty-Four: Anatomy of the Crisis’ (SANTOS, 1986), this concept describes a context in which the government loses its ability to form parliamentary majorities, leading to dysfunctional coalitions incapable of ensuring even minimal levels of governability.

The governability descriptions integrated into the index are designed to classify patterns of roll-call votes across an entire presidential term, rather than to characterize isolated events. Occasional successes or temporary setbacks in coalition-building are to be expected and, therefore, do not determine the overall level of governability for the presidential term as a whole.

To evaluate the weight of coalitions in advancing presidential agendas in Brazil, we compared legislative performance across six presidential terms from 1995 to 2016. This period covers a broad range of two-term presidencies and encompasses a diverse sample of twenty coalitions, that were formed and reshaped in comparable institutional and political contexts (FIGUEIREDO, 2012; INÁCIO, 2006). The dataset includes 2,030 roll-call votes drawn from the NECI Legislative Database: 353 with a qualified threshold, 368 with an absolute threshold, and 1,309 with a simple threshold. Table 01 presents the distribution of votes by presidential term.

Table 01

Qualified threshold votes
FHC I	FHC II	Lula I	Lula II	Dilma I	Dilma II	Total
100	77	47	56	50	23	353
Absolute threshold votes
FHC I	FHC II	Lula I	Lula II	Dilma I	Dilma II	Total
79	112	48	42	49	38	368
Simple threshold votes
FHC I	FHC II	Lula I	Lula II	Dilma I	Dilma II	Total
129	93	256	385	192	254	1309

Source: Elaborated by the authors based on Roll-call Votes – NECI Legislative Database (2021).

To avoid overestimating governability in the Chamber of Deputies, we selected roll-call votes based on two criteria: active involvement of the government leader and a minimum level of competitiveness among parliamentarians – defined as at least 10% dissenting votes. As Figueiredo and Limongi (2001) emphasize, these criteria help prevent mistaking unanimous congressional votes for genuine government strength, an important safeguard against bias when comparing different administrations.

Graphs 03, 04, and 05 display the Legislative Performance Index (LPI) for the six presidential terms from 1995 to 2016, covering votes with simple, absolute, and qualified thresholds, respectively. Consistent with existing literature, the LPI data clearly show that governability is the norm: from FHC I through Dilma II, 92% of roll-call votes aligned with presidential agendas, with 84% of those classified as majority governments (achieving more than 110% of the votes required to meet the threshold).

Graph 03 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

Graph 04 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

Graph 05 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

Despite the persistent pattern of governability, the mean difference tests³ reinforce the indications from the previous graphs, confirming significant variation in legislative performance across the six presidential terms and among the different types of roll-call votes. For instance, when examining votes with a qualified threshold, contrasting trends emerge between the two-term presidencies of FHC and Lula. While the PSDB president’s average LPI increased from 1.09 in his first term to 1.20 in his second, the PT president experienced a decline, with an LPI falling from 1.21 in his first term to 1.11 in his second.

In other words, both administrations followed opposite trends in variation between their terms, while maintaining average legislative performance levels consistent with the governability described in the literature. From FHC’s first to second term, coalition performance improved, shifting from minimally winning coalitions to a majority government. Conversely, Lula’s coalition showed a declining trend, moving from a majority government in his first term to minimally winning coalitions in his second, especially in votes requiring higher thresholds.

Regarding majorities formed in simple threshold votes, all six administrations recorded average LPI scores consistent with majority governments. The lowest averages were observed in FHC I (1.38) and Dilma II (1.39), while the highest value appeared in Lula I (1.64) and Lula II (1.67). The strongest overall legislative performance occurred in votes requiring an absolute threshold, with an average LPI of 1.65 across all terms. Dilma I registered the lowest average in this category (1.49), and Lula I the highest (1.86).

But to what extent can presidential legislative performance be attributed to the coalitions formed? To address this question, the LPI was recalculated for each term, considering only the votes of coalition members (LPIC). As illustrated in Graphs 06, 07, and 08, the levels of governability achieved solely by the president’s allies differ noticeably from the overall legislative performance across the terms, while coalition parties made up the majority of votes.

Graph 06 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

Graph 07 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

Graph 08 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

When considering only the votes cast by coalition members, the government would have succeeded in just 66% of roll-call votes. In other words, comparing legislative performance between presidential terms and coalitions, as shown in Table 02, reveals a 26% drop in the number of votes the governments would have won — out of 2,030 votes, the success rate fell from 1,864 (92%) to 1,348 (66%). The gap is even more pronounced in votes requiring a qualified threshold, where presidential terms had a 93% success rate (330 out of 353), while coalitions secured only 16% (56 out of 353). For votes with the less stringent simple majority threshold, coalition forces performed better – winning 75% (982 out of 1,309) – but still fell significantly short of the 91% success rate under full terms (1,186 out of 1,309).

Table 02

Table 2 – Roll-call Votes by Type of Governability, 1995–2016
		Qualified Threshold					Absolute Threshold					Simple Threshold
		Term		Coalition		% Coalition - %Presidency	Presidency		Coalition		% Coalition - %Presidency	Presidency		Coalition		% Coalition - %Presidency
		n	%	n	%		n	%	n	%		n	%	n	%
FHC 1	Decision-Making Paralysis	1	1%	1	1%	0%	0	0%	1	1%	1%	0	0%	0	0%	0%
FHC 1	Minority Government	14	14%	55	55%	41%	2	3%	5	6%	4%	8	6%	19	15%	9%
FHC 1	Minimally Winning Coalition	38	38%	30	30%	-8%	2	3%	10	13%	10%	3	2%	21	16%	14%
FHC 1	Majority Government	47	47%	14	14%	-33%	75	95%	63	80%	-15%	118	91%	89	69%	-22%
FHC 2	Decision-Making Paralysis	0	0%	0	0%	0%	2	2%	2	2%	0%	0	0%	1	1%	1%
FHC 2	Minority Government	1	1%	66	86%	84%	0	0%	10	9%	9%	3	3%	18	19%	16%
FHC 2	Minimally Winning Coalition	17	22%	11	14%	-8%	5	4%	2	2%	-3%	6	6%	2	2%	-4%
FHC 2	Majority Government	59	77%	0	0%	-77%	105	94%	98	88%	-6%	84	90%	72	77%	-13%
Lula 1	Decision-Making Paralysis	0	0%	0	0%	0%	0	0%	0	0%	0%	6	2%	7	3%	0%
Lula 1	Minority Government	2	4%	46	98%	94%	0	0%	9	19%	19%	6	2%	43	17%	14%
Lula 1	Minimally Winning Coalition	9	19%	1	2%	-17%	0	0%	10	21%	21%	11	4%	37	14%	10%
Lula 1	Majority Government	36	77%	0	0%	-77%	48	100%	29	60%	-40%	233	91%	169	66%	-25%
Lula 2	Decision-Making Paralysis	1	2%	1	2%	0%	0	0%	0	0%	0%	1	0%	4	1%	1%
Lula 2	Minority Government	3	5%	55	98%	93%	0	0%	0	0%	0%	23	6%	36	9%	3%
Lula 2	Minimally Winning Coalition	23	41%	0	0%	-41%	0	0%	0	0%	0%	5	1%	9	2%	1%
Lula 2	Majority Government	29	52%	0	0%	-52%	42	100%	42	100%	0%	356	92%	336	87%	-5%
Dilma 1	Decision-Making Paralysis	0	0%	0	0%	0%	2	4%	2	4%	0%	7	4%	6	3%	0%
Dilma 1	Minority Government	1	2%	50	100%	98%	7	14%	13	27%	12%	24	13%	52	27%	15%
Dilma 1	Minimally Winning Coalition	6	12%	0	0%	-12%	1	2%	12	24%	22%	7	4%	19	10%	6%
Dilma 1	Majority Government	43	86%	0	0%	-86%	39	80%	22	45%	-35%	154	80%	113	59%	-21%
Dilma 2	Decision-Making Paralysis	0	0%	0	0%	0%	0	0%	1	3%	3%	0	0%	1	0%	0%
Dilma 2	Minority Government	0	0%	23	100%	100%	7	18%	15	39%	21%	45	18%	138	54%	37%
Dilma 2	Minimally Winning Coalition	4	17%	0	0%	-17%	2	5%	15	39%	34%	25	10%	49	19%	9%
Dilma 2	Majority Government	19	83%	0	0%	-83%	29	76%	7	18%	-58%	184	72%	66	26%	-46%

Source: Elaborated by the authors, based on NECI Legislative Database (2021).

When breaking down the legislative performance of coalitions compared to that of full presidential terms, the limitations of presidential alliances become more complex and varied, as Table 02 suggests. The most effective coalitions were those during Lula’s second term and FHC’s second term in votes requiring simple and absolute thresholds, as they exhibited the smallest performance gaps relative to their terms. Conversely, the coalitions under Dilma II and Lula I were more situational and contributed the least to their administrations’ overall legislative success in this context. Overall, the persistent gaps between the performance of the presidencies and that of the coalitions, shown in Table 02, underscore how fragile governability in Brazil’s presidential system would be if coalitions were the Executive’s sole source of legislative support.

In short, the discrepancies between the legislative performance of presidencies and their formal coalitions are not adequately explained by existing literature. No theory satisfactorily reconciles the coalitions’ voting shortfalls with the presidents’ high legislative success rates. If, as demonstrated, formal coalitions alone cannot account for the governability observed during these presidencies, then the alternative explanation – that legislative victories result merely from strategically timed votes within those coalitions – also fails to hold up, especially given the consistently insufficient majorities in the most demanding votes (see Graph 08). It leads us to explore a second alternative: the existence of informal coalitions – groups of parties outside the formal alliance that consistently contribute votes in favor of the presidential agenda.

In general, the literature on coalitional presidentialism assumes that parties outside the governing coalition will not support the government. Whether due to ideological differences or the absence of patronage-based arrangements with the Executive, these parties are expected to resist the government’s legislative agenda (AMORIM NETO, 2006; CHEIBUB, 2007; FIGUEIREDO and LIMONGI, 2001). Theories of minority governments (FIGUEIREDO et al., 2012; STRØM, 1990) attribute the decision to form a minority government to cross-national institutional variations. While these explanations are compelling in comparative analyses, they are less applicable when studying a single case, where institutional conditions tend to remain relatively stable – at least within the timeframe of a single presidential term.

This reasoning about the misalignment between ruling and opposition parties closely mirrors assumptions in the literature on coalition-building and parliamentary behavior in Brazil’s Chamber of Deputies (AMES, 2001; FIGUEIREDO, 2012; INÁCIO, 2006; MAINWARING, 1993). Parties outside the coalition are generally expected to support the government only occasionally. Either under specific circumstances or when their policy preferences align relatively closely with the Executive’s. However, the actual voting patterns reveal a more complex reality: across the six presidential terms analyzed, there was a fairly consistent flow of votes from opposition parties to the government. Some opposition parties systematically voted in line with government leadership recommendations in Congress. As shown in Graph 12, during simple threshold votes, the FHCI and FHCII administrations received, on average, 15% and 23% of the votes needed to pass their agendas from outside their formal coalitions, while Lula I and Dilma II received nearly half – 41% and 44%, respectively.

Graph 12 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

Graph 11 shows that solely the votes from parties outside the coalition provided 30% or more of the votes needed for the government to pass qualified threshold votes. Except in the cases of Lula II and FHC I. Occasionally, and in extreme cases (as seen in Graphs 09 and 10), during the FHC II, Lula I, Lula II, and Dilma II administrations, the government could have reached simple or absolute thresholds relying solely on opposition votes. However, these instances are exceptions rather than the rule. As Graphs 12 and 13 indicate, coalitions make up the core of roll-call votes throughout these terms, while parties outside the coalition help secure comfortable majorities.

Graph 11 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

Graph 09 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

Graph 10 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

Graph 13 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

For votes requiring simple and absolute thresholds, the coalitions during the FHC I, FHC II, Lula I, and Lula II administrations operated as majority governments, reinforced by additional vote flows that significantly boosted their legislative performance – especially in Lula’s case. In contrast, during Dilma’s terms, opposition votes played a crucial role. In her first term, the President’s coalition operated within the margins of minimally winning coalitions (1.10 and 0.99 LPIC), only reaching the status of majority government with opposition support (1.48 and 1.57 LPIC). In her second term, opposition votes were decisive in transforming a minority coalition performance (0.96 and 0.96 LPIC) into that of a majority government (1.40 and 1.49 LPIC).

The opposition consistently played a crucial role as the swing vote ensuring governability in votes requiring a qualified threshold. As shown in Graph 14, from 1995 to 2016, no coalition alone averaged the minimum threshold required for governability – that is, the level of a minimally winning coalition (LPIC > 1). The FHC I administration came closest but still depended on additional support from the opposition (0.14 LPIO) to advance its presidential agenda. In the cases of FHC II, Lula I, Lula II, Dilma I, and Dilma II, it was the backing from outside the coalition that ensured comfortable winning margins in qualified threshold votes – effectively elevating these governments from minority to majority status, on average. Even if such episodes were sporadic, coalitions capable of reaching the 3/5 vote threshold in the Chamber of Deputies have become increasingly rare over time, as illustrated in Graph 08.

Graph 14 Source: Elaborated by the authors, based on NECI Legislative Database (2021).

As counterintuitive as it may seem given the current state of research on coalitional presidentialism, the role of the opposition as a driver of governability has long been recognized in the literature. In his classic analysis of the factors contributing to decision-making paralysis, Santos (1986) compared roll-call voting patterns during the administrations of Juscelino Kubitschek and João Goulart. He identified the withdrawal of the main opposition party (UDN) from the informal coalition (which had been a strong supporter of the PSD’s reformist agendas)⁴ as a major destabilizing factor during that period. Considering the crucial role of these parties’ votes in securing the comfortable majorities needed to pass contemporary presidential agendas, and following Santos’s reasoning (1986), Table 03 lists the parties belonging to informal coalitions, defined here as those contributing, on average, 10% or more of the votes required to meet threshold levels. The findings reveal some surprising and unexpected dynamics in the patterns of informal support.

Table 03

Presidency	Qualified threshold votes	Absolute threshold votes	Simple threshold votes
FHC I	PPR	PT/ PR	PPR
FHC II	PT	PT/ PTB	PP/ PTB
Lula I	PMDB/ PFL/ PSDB/ PPB	PMDB/ PFL/ PP/ PSDB	PP/ PSD
Lula II	PSDB/ DEM	PSDB/ DEM	---
Dilma I	PSD/ PSDB	PSDB/ PSD/ DEM	PSD
Dilma II	PSDB/ PSB	PSDB/ PP/ PSB	PSDB

Source: Elaborated by the authors.

Despite their longstanding electoral rivalry in presidential races, the data reveal an informal complementarity between the PT and PSDB coalitions: when one party holds the presidency, the other tends to provide legislative support. Another surprising finding is that the PFL/DEM (a traditional right-wing party) informally backed Lula’s administrations during times of greatest need – specifically in votes requiring the most demanding thresholds (qualified and absolute). Lula’s second term also stands out for the fragmented support from parties outside the coalition in simple threshold votes, with no single party clearly acting as an informal ally (none reached an LPO of 0.1). This scattering of outside-party votes likely reflects the strength of Lula’s formal coalition, which reduced the need for informal backing.

A noteworthy case deserving further investigation is the phased integration of the PMDB and PSD into governing coalitions. Initially, both parties acted as informal supporters – PMDB during Lula’s first term and PSD during Dilma’s first. However, following the presidents’ re-elections, these parties formally joined the governing coalitions in Lula’s second term and Dilma’s second term, respectively. These cases are particularly valuable as controlled scenarios for examining executive-legislative negotiations involving parties transitioning from informal to formal coalition members across consecutive presidential terms.

In sum, voting patterns in Brazil’s coalitional presidentialism suggest that groups outside the formal governing coalitions play a decisive role in ensuring governability. These actors are not merely dispersed throughout the Chamber of Deputies; rather, certain parties consistently support presidential agendas and can be considered informal members of the governing coalitions.