Patronage and Presidential Coalition Formation

Abstract

Effective democratic governance rests on the executive’s ability to forge coalitions that can advance policy and sustain the government against challengers. Scholars have long focused on cabinet appointments to understand how executives build coalitions with their legislative allies. In many democracies, however, cabinet appointments at the ministerial level may only represent the tip of the iceberg. We show that administrative political appointees (APAs) beneath the ministerial level constitute one of the most important ways that cooperation between legislative and executive is forged. Leveraging a unique and comprehensive database of an average 2600 Brazilian APAs per year over two decades, we evaluate their effect on coalition unity in critical legislative votes. We demonstrate that these APAs, which we collectively term the “patronage coalition,” have a significant effect on legislative support and thus are a critical tool for presidents. Our results are particularly relevant to a new emphasis in the political science literature on the “toolbox” that presidents utilize to address the challenges of simultaneously maintaining legislative support while implementing policy. These results demonstrate that the patronage coalition is a fundamental tool that should be more widely integrated into models of legislative-executive bargaining.

Keywords

patronage coalitional presidentialism political appointments legislative coalition bureaucracy

Coalition discipline has far reaching consequences for government stability, policy outputs, and government quality. In their efforts to build coalitions, presidents and prime ministers face many conflicting objectives, such as the need to approve desired legislation, implement policies, and defend against rival coalitions.

Scholars of parliamentary democracies have found that coalition discipline in legislative voting depends to a large extent on “cabinet coalescence,” “the amount of deviation from proportionality between seats and votes that a given election produces” (Rose 1984). In presidential systems, similar findings hold: Amorim Neto’s (2002) influential work on Brazil, for example, showed that the fairness of the coalition agreement, as measured by cabinet distribution, helped explain legislative success.¹

While cabinet positions serve as public signals of the executive’s policy priorities and those of their top coalitional allies, presidents in coalitional presidential systems draw on a much broader range of tools to hold together their coalitions (Chaisty et al. 2018; Chaisty et al. 2014; Raile et al. 2011). Within the realm of appointed positions, ministers are only the tip of the iceberg. Appointments beneath the cabinet level can be extremely relevant to achieving a president’s goals. Patronage appointments serve both a control and a reward function (Panizza et al. 2018). Administrative political appointees (APAs) fill the layer of positions between the minister and civil servants, serving as policy gatekeepers, watchdogs, and influencers. APAs offer prestige and material rewards to parties or factions who align themselves with the president. Furthermore, APAs need not mirror the partisan identity of cabinet appointees, meaning that they provide controls and rewards distinct from, and often not aligned with, the partisan identity of the cabinet minister.

Scholars have speculated that APAs could have an impact on legislative support (Grindle 2012, 208; Hassan and Sheely 2016). However, there has been little systematic evaluation of their actual effect. We analyze the Brazilian case to explore the effect of patronage coalitions on legislative coalition unity in critical legislative votes. Theoretically, we call attention to the neglected role of the patronage coalition in the toolbox of coalitional presidentialism, demonstrating that patronage appointments serve as an important complement to other tools used to hold the coalition together. Empirically, we illustrate the importance of the patronage coalition as a mechanism binding fractious coalitions, and demonstrate how the partisan composition of the patronage coalition, relative hierarchy of appointments, relative ideology of appointees’ parties, and presidential characteristics jointly condition the effect of the patronage coalition on legislative support.

Although studying one country in isolation limits the extension of our arguments, the approach allows for more rigorous testing of the causal mechanisms linking indicators to each other in ways that could not be explored in cross-national research (Mershon 2001, 280). It also permits extremely data-intensive analysis: we leverage a comprehensive database of an average 2600 APAs per year in Brazil over the course of two decades, incorporating partisan affiliation data that is not available in many other countries.²

The paper provides empirical backing for the emerging consensus that the strategies of coalition formation effectively extend well beyond cabinet appointments alone, and that the patronage coalition is an essential component in presidential coalitional strategies. The first section describes the literature on coalition formation and legislative voting. The second theorizes how the patronage coalition may influence legislative voting. The third discusses the Brazilian case used to test the theory. The fourth presents our data, method, analyses, and findings, before concluding.

Coalition formation and legislative voting in presidential systems

Scholars of democracy have paid great attention to coalition-building, especially within parliamentary systems, where coalitions determine prime ministers’ political survival. Scholars of presidential systems increasingly also point to its coalitional logic. As one comprehensive study found, more than half of presidents today are “minority presidents” governing with multiparty coalitions (Chaisty et al., 2018, 2). Even in majority presidential systems, presidents often administer coalitions of factions within the dominant party. As Lewis (2008, 8) notes with reference to the United States, the “giving and withholding of jobs is used to maintain party discipline or as a bargaining chip to help presidents get their way in Congress...jobs represent a crucial political resource for presidents.”

What is the logic that guides the distribution of these jobs? Gamson (1961) hypothesized that participants in a coalition will demand “a share of the payoff which is proportional to the amount of resources they contribute to it.” This hypothesis has guided much of the subsequent work on coalition formation. The participant “contribution” is usually understood as legislative support, and the “payoff” as participation in the executive cabinet.

Research on participation in executive coalitions has found empirical grounding for Gamson’s model, suggesting that coalition discipline depends to a large extent on cabinet coalescence (Rose 1984, Morelli 1999; Warwick and Druckman 2001, 2006; Browne and Franklin 1973; Martin and Stevenson 2001). Prime ministers in parliamentary democracies can often count on strong, disciplined support for extended periods, and in many Westminster systems, failure to maintain party or coalitional unity on critical votes may cost the governing coalition its mandate, triggering new elections. The question of survival serves as a powerful disciplining mechanism.

Coalition formation in presidential systems may be more complex, given that fixed terms and separate origins mean that there are fewer immediate incentives for party discipline on the executive’s legislative priorities. Further, in multiparty presidential systems, fragmentation may weaken the president’s potential legislative support: the simple existence of multiple parties makes it less likely that the president’s party will hold a legislative majority (Amorim Neto 2002, Mainwaring 1993, Figueiredo et al. 2012). It is therefore critical to understand the tools which presidents use to build floor coalitions, “durable cross-party alliances that presidents form to support their agenda for meaningful periods of time” (Chaisty et al. 2018, 10).

The effects of legislative fragmentation may be partially overcome by coalition agreements that compensate parties for their legislative support by providing seats in the ministerial cabinet. One of the most influential studies of the relationship between cabinet formation and coalition discipline in presidential systems is Amorim Neto (2002), which finds that legislative support may be secured through cabinet appointments, and the more “coalescent” the cabinet, the more successful the presidential agenda.³ This finding is echoed by a number of scholars studying coalitional presidentialism in Latin America who have refined Amorim Neto’s findings about the importance of cabinet portfolio distribution (Alemán and Tsebelis 2011, Bertholini and Pereira 2017, Raile et al. 2011, Zucco 2009). Minority presidents have been found to use cabinet appointments to create a legislative shield against impeachment (Pérez-Liñán 2007, Cheibub et al. 2004). Cross-national research finds similar cabinet dynamics in other contexts, such as Ukraine (Chaisty and Chernykh 2017; Bucur 2020).

But to what extent is ministerial allocation the only, or even the most important, factor at work in holding together coalitions? Contrary to much of the literature on parliamentary systems, the presidential cabinet is often distant from what Gamson’s principle of proportionality would suggest. Especially in highly fragmented party systems, some parties in the president’s floor coalition may not even be represented in the cabinet. Chaisty et al. (2018,10) note, for example, that “the three Brazilian presidents between 1995 and 2015 sometimes enjoyed the support of 10–12 parties on the floor—that is, parties formally declaring their support for the executive and taking the government whip in Congress—while accommodating only 5–7 of these parties in the cabinet.” Some parties may not even wish to be in the cabinet: parties in Ecuador’s “ghost coalitions,” for example, avoided ministerial level cabinet positions while joining floor coalitions (Mejía Acosta 2006).

Compelling evidence suggests that while cabinet coalitions are important, presidents may look beyond cabinet appointments to cement fractious coalitions. In the wake of various high-profile corruption scandals, for example, scholars of Brazilian presidential coalition formation suggested that perhaps cabinet disproportionality had generated the necessity for corrupt side-payments to compensate losers (Pereira et al. 2011). Somewhat less nefariously, the “inefficient secret” is that “presidents may actually benefit from the fact that some parties do not in fact seek policy but rather narrow, particularistic benefits” (Chaisty et al. 2018, 17; Shugart and Carey 1992). Such benefits might include budgetary allocations to legislators’ home districts or legislation favoring campaign backers (Carazza 2018, Mancuso 2012, Samuels 2002, Mancuso and Speck 2015). Chaisty et al. (2018, 19–21) suggest that the toolbox presidents use to hold coalitions together includes agenda-setting powers (the president’s direct legislative powers that allow them to dominate the legislature), cabinet powers (membership in the presidential cabinet), partisan powers (powers that presidents gain from domination of their parties), budgetary authority (often translated into pork), and informal institutions (the exchange of favors). One tool not explicitly mentioned in this analysis of the presidential toolbox is the patronage coalition.

Theorizing patronage coalitions in dynamic coalition management

The “patronage coalition” refers to all APAs within government: the appointees within the executive bureaucracy who are not cabinet members but who are not hired via merit examination and are members of political parties.⁴ The number of potential APAs varies widely, from zero in Denmark, to hundreds in France, Britain and Germany, to thousands in the United States and Brazil (Lewis 2008; Peters and Pierre 2015). As these magnitudes suggest, even in countries with strong civil service protections, such as the United States and Brazil, the number of individuals serving at the president’s pleasure is significant.

There are many reasons that the ability to influence the selection of APAs is valuable to legislative parties. APAs play a fundamental role in the supply of coalitional goods such as policy and pork. APAs serve as gatekeepers for public spending, prioritizing and routing the benefits that make their way through the bureaucracy. In multiparty systems, furthermore, it is seldom the case that APAs merely extend the partisan or factional interests of the ministerial appointee. The strategy of mixing and matching partisans may provide a valuable perk by permitting members of the cabinet, patronage, and floor coalitions to monitor each other (Pereira et al. 2017). APAs serve as watchdogs upward, keeping tabs on the minister above them so as to ensure that coalition bargains are upheld (Lipsmeyer and Pierce 2011; Thies 2000; Verzichelli 2008). APAs serve as watchdogs downward, ensuring that the civil service is effectively delivering information, implementing policy, and delivering coalition goods (Lewis 2008, 66–68). APAs are also often important information brokers in the legislative process, providing information from within the bureaucracy to their party or coalition, which can be used to evaluate the costs, benefits and viability of distinct policy alternatives.

With these various functions of APAs in mind, our theory anticipates that the patronage coalition will help explain legislative success by providing incentives for cooperation between actors in the legislative and executive branches. Gamson (1961) summarized the overall logic well: “Coalitions are temporary, means oriented, alliances among individuals or groups which differ in goals. There is generally little value consensus...This makes the pursuit of power itself, i.e. control over future decisions, an ideal basis for coalition formation since it is an instrument for the achievement of widely ranging and even incompatible goals.”

How do presidential strategies for appointment of the patronage coalition affect floor coalition discipline? On the demand side, APAs may be sought both for policy influence and control over patronage. One Brazilian president noted in the runup to a key vote that vacillating legislators were about equally divided between those who wanted his attention, those who wanted pork, and those who wanted patronage (Salomon 1997). Parties are not homogenous, and appointments may help address the parochial concerns of cliques within parties. Cabinet-level appointments that satisfy senior party leaders may be insufficient to ensure the votes of more junior legislators, or “backbenchers,” whose discipline may be improved through APAs (Mershon 2001, 280). Once their preferred APAs are in government, legislators will have strong incentives to preserve them there by cooperating with the executive.

On the supply side, the president is concerned to control governmental decisions for which she will be held responsible by both voters and the floor coalition, and appointments may help to ensure better alignment between the members of the coalition represented in the executive and the floor coalition. Further, she faces a continual tradeoff between various potential payouts to her coalition: not all demands can be met with pork, budget amendments, policy concessions, or cabinet slots, and appointments provide a valuable alternative. An additional motivation is that cabinet appointments often reflect inter-party negotiations between senior coalitional allies at the beginning of a cabinet. As the relative alignment and strength of parties shifts, APAs may allow for subtle readjustments in coalition payoffs, without the need to reconstitute the entire coalitional bargain through a cabinet reshuffle.

If cabinet realignments are akin to a sailboat changing tack, patronage realignments may be more like tweaking the sail trim, essential to keeping the boat moving efficiently, but less apparent to the casual observer. Our argument is not exclusionary: in the same way that the cabinet does not explain the totality of executive-legislative bargaining, the patronage coalition is complementary to other variables, including cabinet appointments. Coalition management is a dynamic exercise that requires simultaneous efforts to fit together all the pieces of what one high-ranking official told us was a complex “jigsaw puzzle” (Anonymous 2013). This has important implications: the overall relationship between patronage appointments and legislative success is less a quid pro quo (support on particular votes, for example) and instead part of an extended multilevel negotiation about the direction of the presidential administration. The potential scale of APAs, however, suggests that their omission distorts estimates of coalition goods that coalition members receive. Failure to incorporate the patronage coalition could lead to overestimation of the error term, and as a consequence, to poor models of coalition coalescence. Further, there is no a priori theoretical reason to believe that the patronage coalition is less important than other hypothesized variables influencing legislative support. It is thus worth evaluating the patronage coalition, rather than continuing to rely solely on cabinet appointments as a proxy for complex and dynamic processes of coalitional bargaining.

The patronage coalition in Brazil

Brazil exemplifies many of the elements of the presidential toolbox described by Chaisty et al. (2018), but a quick inventory of the five tools they describe also suggests the importance of more systematically considering the patronage coalition’s effect on legislative support for the president’s agenda.

Agenda-setting powers are quite strong for Brazilian presidents (Figueiredo and Limongi 1999, 2002) but vary little across time and thus seem unlikely to explain medium and short-term changes in legislative support. Cabinet powers are significant, but given that Brazil reached 35 parties during this period, cabinet appointments could not satisfy all coalition members’ demands. Partisan powers within the president’s party are significant, but since the late 1980s, the president’s party has never held a majority in the lower house. Coalition party leaders are given significant power by congressional rules and can whip party members to ensure party discipline (Figueiredo and Limongi 1999, 2002; Limongi and Cortez 2010). This, though, begs the question of what drives party leaders toward the executive in the first place. Budgetary authority is significant, but the strapped fiscal conditions of much of the past generation have limited – without eliminating – the budget as a coalition-building tool. The executive branch actively controls the size and timing of budget outlays (Figueiredo and Limongi 2002) and until 2019 had the power to withhold outlays at will: Congress could only authorize spending but not mandate it. Budgetary control meant that pursuit of pork through individual budget amendments was far less significant than might have been imagined (Figueiredo and Limongi 2002; Vasselai and Mignozotti 2014). With regard to informal tools for the exchange of favors, evidence from congressional votes suggests that executive inducements are at least as relevant as ideology in determining roll call behavior (Zucco and Lauderdale 2011, 365–7), although these inducements have not been clearly delineated in the literature, and the disbursement and oversight of such transfers is not transparent.

Even in the aggregate, then, the analysis of these five tools leaves open questions about the determinants of coalition discipline in Brazil. From a purely analytical perspective, none of these tools effectively incorporates the patronage coalition or evaluates its relative effect. Chaisty et al. (2018) do not mention APAs in discussing cabinet powers, except under the broader category of “the spoils of executive office” and “patronage” (121, 138). Although Zucco and Lauderdale (2011) discuss “cabinets,” “portfolios,” and “second and third-level posts” as inducements, their paper was not designed to evaluate the independent effect of these appointments nor to separately evaluate the effect of cabinet appointments from APAs. In their discussion of the informal exchange of favors, Chaisty et al. (2018, 190) mention appointments to “influential positions outside of the legislature and the cabinet,” but they appear to be thinking less of APA appointments within the executive bureaucracy and instead of appointments to “parastatals and boards of business.” Further, these appointments are analyzed in a bundle together with other favors, such as access to electoral resources, favorable treatment for allies’ businesses, and even illicit exchanges. The significant scale of APAs in the Brazilian state, the lack of detailed attention to their effects on executive-legislative relations, and evidence that they may follow a distinct logic suggest that the role of the patronage coalition should be evaluated as a separate, sixth tool in the presidential toolbox.

Qualitative evidence demonstrates that the patronage coalition has played a role in holding together the coalition in governments of all stripes. During the Cardoso administration (1995–2002), news reports showed intense bargaining over the patronage coalition as pension, civil service, and tax reforms were being considered (Toledo 1996; Salomon 1997). After Congress approved a constitutional amendment on reelection, President Cardoso concluded that representatives had heard the “hoarse voice of the streets.” A congressman quipped, “in addition to [that] hoarse voice...they hear the murmur of appointments” (Gaspari 1997).

Similar patronage bargaining took place under President Luiz Inácio Lula da Silva (2003–2010). During a critical pension reform, one party leader complained that his rank-and-file members were being led by the “scent of appointments” (Costa 2004) and another complained that appointments were being handed out to people that belonged to the “yes, sir” coalition (Maia 2004). When the leadership of the Chamber of Deputies threatened to fall into the wrong hands in September 2005, Lula rejiggered the cabinet, reassigning three cabinet seats away from two core coalition parties (PT and PCdoB), and rewarding the PSB, PP, and PMDB with one seat each. The shift within the patronage coalition was even more significant, with the share of high-level appointees allocated to those three parties and two allies increasing by 25% even as the PT and PCdoB lost appointees in the 3 months immediately before and after the vote. The government’s choice of leader was duly elected.

The logic of the patronage coalition and the manner by which it enhances party responsiveness was well exemplified during Rousseff’s first term. Investigative journalists obtained a spreadsheet of more than 300 federal jobs, which listed nominees and the “godfathers” who nominated them. Asked to explain, the President’s staff responded, “It is natural that the allied base, a foundation of support for the government, should legitimately demand the division of leadership slots in the federal sphere” (Patury and Loyola 2011).

One final descriptive example comes from the PP, whose leaders were deeply implicated in the Carwash corruption investigations that emerged in March 2014. Yet the party was a key coalition ally, holding 8% of lower house seats. As Rousseff’s coalition became more unstable in the midst of a combination of massive public protests, economic deterioration, and scandal (Hunter and Power 2019), and as members of key parties such as the MDB and PDT threatened to leave the coalition, she faced a difficult choice. It would be unpopular to the broader public and likely inflammatory to other coalition allies if she appointed another PP cabinet minister in the midst of the scandal. Yet she needed the party onside. After allowing PP’s proportion of the patronage coalition to slowly decline throughout her first term, the party’s share rose steadily throughout her second term relative to its share of the cabinet (Figure 1). As shown later, this contributed to robust PP legislative support.

Most countries place formal limits on the use of APAs. In Brazil, the 1988 Constitution establishes merit-based hiring as the rule for all civil servants. The primary positions open for non-merit hiring are so-called High-Level Direction and Advisory (Direção e Assessoramento Superior, DAS) slots. The DAS positions are organized in a six-level hierarchy from junior (DAS-1) to senior (DAS-6).

The bars on the left of Figure 2 show the distribution of appointees at every level of the hierarchy. The numerical columns show the average annual number of possible appointees at each level, the average number of slots at that level for each minister, and the relative weight of each slot, relative to a minister, over the period.

Even given hiring constraints, there is still ample room for political appointments. Admittedly, most DAS appointees are not partisans: roughly three quarters of the DAS appointees are typically civil servants, a proportion determined by a combination of custom and legislation (Bersch et al. 2017b; Lopez and Silva 2019; Taylor 2020, 298). But the President directly selects around 1200 of the DAS appointees, and her ministers select several thousand more, often with presidential guidance. At the more senior DAS 4–6 level, all appointments must be approved by the presidential chief of staff, ensuring tight political oversight, whether these APAs are registered party members or not. Indeed, one of the reasons the Brazilian case is valuable is precisely because the party membership of appointees can be identified through electoral court records.⁵

It is not feasible to reconstruct the manner of delegation between the president and cabinet minister in each APA appointment, but across the two decades of this study, one important pattern holds: seldom does the partisan identity of APAs align fully with the party of the minister. Within each ministry, APAs are appointed with multiple party affiliations. Bersch et al. (2017a) note that in the government agency with the highest level of partisan dominance, for example, the largest party held only 35% of the DAS slots in 2011, and even in that case, other parties were represented among the DAS appointees.⁶ As potential partners negotiate their legislative support, their understanding of the fairness of the agreement is likely to be guided by the allocation and seniority of posts they receive relative to other parties.

Method, data, analysis

Does the patronage coalition matter for legislative discipline? Does it matter as much as the cabinet coalition? We address these questions in four logically sequential analytical stages.

First, we replicate one of the most influential studies of how cabinet coalitions influence legislative discipline, by Amorim Neto (2002).⁷ We use his model of the cabinet coalition as a baseline to evaluate the relative effect of the patronage coalition on legislative discipline. Having successfully replicated his analysis for 1985–1999, we extend his same model to 1999–2016, a period encompassing six presidential administrations, four presidents, and 28 presidential coalitions.⁸ The results are in keeping with those for the 1985–1999 period, providing additional evidence for Amorim Neto’s conclusion that “cabinets with a higher coalescence rate maximize the coalitional discipline of governing parties, and coalition discipline decreases later in the presidential term” (2002, 69).

The second stage of research introduces a few friendly amendments to Amorim Neto’s model. We adopt improved measures of proportionality and ideology not available to him two decades ago. We further test the effect of presidential approval ratings and presidential dummy variables that account for the differing strategies employed and contingencies confronted by distinct executives, finding that some presidents are significantly more capable coalition managers.

The third stage uses our enhanced model to evaluate the effect of the patronage coalition. We discover that the patronage coalition has a larger effect on legislative discipline than the cabinet coalition.

The fourth stage analyzes how payoffs that individual parties receive via ministerial and APA appointments influence legislative discipline. Analyzing three prominent coalition parties in the Workers’ Party administrations that governed from 2003 to 2016, we demonstrate that membership in the patronage coalition has consistently larger effects than the cabinet coalition. Throughout, we provide brief descriptions of the variables used; further details on calculation and sources can be found in the online methodological appendix.⁹ All models are ordinary least squares.

Replicating and extending Amorim Neto’s model to 2018

Space does not permit us to reproduce here the careful description of the hypotheses in Amorim Neto (2002). To summarize, he tests three core hypotheses about the impact of cabinet coalescence on legislative voting: 1) the more closely the cabinet coalition mirrors the floor coalition, the more likely parties are to support the executive in legislative votes; 2) because the president’s term is fixed, as her term elapses, the value of holding cabinet posts declines, and so does legislative support; and 3) the greater ideological diversity in the cabinet coalition, the lower coalition discipline. From these hypotheses, Amorim Neto derives the following regression equation:

CD = $β_{0}$ + $β_{1}$ CABINET + $β_{2}$ TIMEREMAIN + $β_{3}$ IDRANGE + $ε$

The dependent variable, coalition discipline (CD), is measured as the percent of the total legislative membership of the parties represented in the cabinet voting in accordance with the orientation of the government leader on each roll call vote. Roll call votes are a good site to measure the effect of coalitional agreements and observe party conflict and coalition discipline, especially when the preferences of the president are clear, because they permit us to observe the extent to which legislators in parties represented in the cabinet take unified action. Not all roll calls are equally important, and roll calls with near consensus will be a less relevant test of coalition unity than those in which the majority and minority are more evenly matched. For this reason, Amorim Neto restricts the analysis to roll calls that meet a certain threshold of attendance and closeness, jointly referred to as “critical-ness” (Amorim Neto 2002, Carey 2000). We follow his convention throughout, defining critical votes as roll call votes in which at least 60% of legislators vote and at least 20% of legislators dissent from the majority. The cutoff value for this formula is 0.24 (see Amorim Neto 2002, 64).¹⁰

With regard to the independent variables, CABINET is the measure of cabinet coalescence, calculated as:

C o a l e s c e n c e = 1 - \frac{1}{2} \sum_{i = 1}^{n} (| S_{i} - M_{i} |)

where S_i represents the number of seats a party holds in the lower house, M_i represents the number of cabinet ministries, and i indexes party. TIMEREMAIN is measured as the difference in days between the day that a roll call vote is taken and the constitutional end of a given administration.¹¹ To calculate IDRANGE, Amorim Neto allocated each party to five standard positions along the left-right dimension and stipulated a constant distance between each, which he acknowledged was probably “too crude a measure” (2002, 69).

Table 1 includes four models. Model 1.1 is Amorim Neto’s original “best model.” Model 1.2 is our replication. Having fully replicated the results, we then apply his model to new data for the period 1999–2018 (Model 1.3), encompassing a large number of roll call votes: 1478 roll calls meet the test of critical-ness, by contrast to 324 in Amorim Neto’s original sample. Model 1.4 reports the results for the combined period 1989–2018.

Table 1.

Replication and Extension of Amorim Full Model. Dependent Variable: Coalition Discipline of Cabinet Parties.

Independent variables	1.1Amorim Neto model1989-1998	1.2Authors’Replication1989-1998	1.3Extension to later years1999–2018	1.4Amorim Neto model applied to full period1989-2018
Cabinet coalescence	0.874***	0.874***	0.107*	0.274***
Time remaining in term	0.00007***	0.00007***	0.00006***	0.00004***
Ideological range	0.012	0.012	−0.12∗∗∗	-0.05***
Constant	0.107	0.107	0.723***	0.512***
Adjusted R-square	0.151	0.151	0.055	0.028
N	324	324	1478	1802

Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01.

Amending Amorim Neto’s model

The second stage takes advantage of advances in the literature and data availability to amend Amorim Neto’s model. Four major changes are shown in Table 2. The first is to employ a cosine measure of cabinet proportionality (Almeida 2018; Koppel and Diskin 2009), that better meets some of the desirable properties of disproportionality measures (see online appendix). It is calculated as:

P r o p o r t i o n a l i t y = \frac{\sum_{i = 1}^{n} (S_{i} {*M}_{i})}{\sqrt{\sum S_{i}^{2} * \sum M_{i}^{2}}}

where i indexes party, S_i represents the number of seats a party holds in the lower house, and M_i represents the number of cabinet ministries it holds.

Table 2.

Amendment of Amorim Model. Dependent Variable: Discipline of Cabinet Parties (2.1, 2.1b) and Coalition Parties (2.2, 2.3, 2.2b, 2.3b).

	Models using ideological range			Models using ideological heterogeneity
Independent variables	2.1Amended AN model 1989–1998	2.2Extended amended model (EAM)1999–2018	2.3Extended amended model (EAM)1989–2018	2.1bAmended AN model 1989–1998	2.2bExtended amended model (EAM)1999–2018	2.3bExtended amended model (EAM)1989–2018
Cabinet coalescence	1.06***			1.076***
Cabinet proportionality		0.142**	0.125***		0.111*	0.101***
Time remaining in term	0.0001**	0.0001***	0.00004***	0.0001**	0.0001***	0.00005***
Ideological range	-0.092**	−0.095***	−0.094***
Ideological heterogeneity				.-0.397	−0.501***	−0.415***
Constant	0.108	0.560***	0.598***	0.064	0.547***	0.571***
Adjusted R-square	0.165	0.053	0.038	0.153	0.056	0.032
N	324	1478	1802	324	1478	1802

Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01.

Second, because some parties may be members of the coalition but not members of the cabinet, we adopt a measure of publicly-declared coalition membership (Almeida 2018).¹² Third, we employ a more precise measure of ideological coherence. Rather than allocating parties to five rigid categories, the dataset adopts Power and Zucco’s surveys of sitting legislators and their subsequent relative scoring of those positions (Power and Zucco 2009; Zucco 2009; Power and Zucco 2014; data from Almeida 2018, Appendix E). Models 2.1–2.3 use an ideological range measure akin to that used by Amorim Neto, derived from this newer data. Fourth, the ideological measures developed by Power and Zucco also allow us to derive a measure of coalition ideological heterogeneity, weighted by party size. Models 2.1b–2.3b uses this heterogeneity measure to test whether greater heterogeneity lowers discipline.

Extending the analysis to an additional two decades of data produces findings in line with Amorim Neto’s original results (Table 2). Beginning with this stage of research, however, we set aside the years prior to 1999 because there is no data available (nor, we believe, collected by relevant government agencies) for the period covered by Amorim Neto on the central variable that interests us in this paper: the patronage coalition. The remainder of this paper focuses on the years 1999–2018.

As expected, the magnitude of some coefficients differs given the use of the improved measures of cabinet coalescence and ideological range. Most important, the coefficient for cabinet proportionality is only one-sixth what it was for cabinet coalescence. This difference may be explained by the fact that the period covered by Amorim Neto encompassed two presidents with atypically low levels of coalescence (Collor and Franco, with coalescence rates from 22 to 59), as well as another with high levels (Cardoso, with rates from 50 to 72). Under these conditions, the coefficient may capture a great deal of noise from the different strategies adopted by each of these presidents.

For this reason, Table 3 also evaluates characteristics of individual presidents. The shaded columns (2.2 and 2.2b) simply duplicate the corresponding models in Table 2. Model 2.4 builds on the extended amended model by adding measures of net presidential popularity, per Altman (2000). Model 2.5 adds presidential dummies. This significantly boosts model fit, with coefficients in the expected direction. Given significant collinearity between ideological range/heterogeneity, net presidential popularity, and the presidential dummies, Model 2.6 excludes the first two variables.¹³ Model 2.6 shows that presidents Cardoso and Lula had about 15–19% more coalition unity than Temer, while Rousseff was about 5% higher. This comports with the historical record of significant turmoil under Temer and during the latter years of the Rousseff administration and also with scholarly findings that Rousseff relied on standard law-making procedures less than her predecessors (Amorim Neto 2019).

Table 3.

Amendment of Amorim Model, Including Ideological Range, Ideological Heterogeneity, and Popularity Dependent Variable: Discipline of Coalition Parties.

	Models with ideological range			Models with ideological heterogeneity			Model with presidential dummies
Independent variables	2.2Extended amended model (EAM)1999–2018	2.4EAM with popularity1999-2018	2.5EAM with popularity and presidential dummy1999-2018	2.2bExtended amended model (EAM)1999–2018	2.4bEAM with popularity1999-2018	2.5bEAM with popularity and presidential dummy1999-2018	2.6EAM with presidential dummy1999-2018
Cabinet proportionality	0.142*	0.012	0.187**	0.111*	0.018	0.157**	0.224*
Time remaining in term	0.0001***	0.0001***	0.0001***	0.0001***	0.0001***	0.0001***	0.0001***
Ideological range	−0.095***	−0.167***	0.045
Ideological heterogeneity				−0.501***	−0.642***	−0.093
Net presidential popularity		0.001***	0.0005***		0.001***	0.0004***
Cardoso			0.200***			0.191***	0.190***
Lula			0.103**			0.136*	0.153***
Rousseff			0.024			0.060	0.052**
Constant	0.560***	0.692***	0.234	0.547***	0.606***	0.302***	0.248***
Adjusted R-squared	0.053	0.101	0.149	0.054	0.080	0.149	0.145
Number of observations	1478	1478	1478	1478	1478	1478	1478

Note: ∗p < 0.1; ∗∗p < 0.05; ∗∗∗p < 0.01. Reference category for presidential dummy variables is Temer.

In all but one case (2.4/2.4b), these simple models confirm Amorim Neto’s general argument that as cabinet coalescence (or proportionality) increases, the likelihood of government unity on roll call votes increases and that as a presidential term elapses, coalition discipline frays.

Incorporating the patronage coalition

The third stage of analysis builds off Model 2.6 to evaluate the effect of the proportionality of the patronage coalition on coalition unity. The first column of Table 4 (Model 2.6) is duplicated from Table 3 to facilitate comparison. The findings of the next two columns (Models 3.1 and 3.2) demonstrate that the patronage coalition is more influential on coalition discipline than the cabinet coalition. This comports with our theory of the disciplining effect of participation in the patronage coalition, given the various incentives APAs provide to legislators when considering the president’s legislative initiatives.

Table 4.

Determinants of Coalition Discipline (1999–2018), with Patronage Coalition. Dependent Variable: Discipline of Coalition Parties.

Independent variables	Model 2.6 EAM with presidential dummy1999–2018	Model 3.1 EAM with all partisan DAS1999–2018	Model 3.2 EAM with senior partisan DAS61999–2018
Proportionality cabinet	0.224*	0.233***	0.165***
Proportionality DAS		0.360***
Proportionality DAS6			0.537***
Time remaining in term	0.0001***	0.0001***	0.0001***
Cardoso	0.190***	0.181***	0.124***
Lula	0.153***	0.178***	0.176***
Rousseff	0.052**	0.080***	0.068***
Constant	0.248***	−0.090	−0.130
Adjusted R-squared	0.145	0.149	0.152
Number of observations	1478	1478	1478

Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01. Reference category for presidential dummy variables is Temer.

As Figure 2 showed, there is a clear hierarchy of appointments within the federal civil service. This figure also showed that the 200 high-level DAS6 appointees are significant in the bureaucracy, sitting atop an average of 19,960 appointees.¹⁴ Previous studies have shown that more senior appointees are more likely to be party members, with the proportions ranging from around 12% among junior DAS-1 appointees to closer to 30% among senior DAS-6 appointees (Lopez and Silva 2019). Together, the rewards of participation in the patronage coalition and the hierarchy of appointments inspire two hypotheses. First, the more coalescent the patronage coalition, the more unified the behavior of the legislators belonging to the patronage coalition parties on roll call votes. Second, hierarchically higher APAs will have a greater effect on party discipline on roll call votes.

These hypotheses distinguish the “patronage coalition” of APAs from Amorim Neto’s “cabinet coalition” of ministers. We test both hypotheses using the variables PROPORTIONALITY_DAS, which covers all DAS appointees, weighted by the proportion of DAS to minister, and PROPORTIONALITY_DAS6, which includes only high-level DAS-6 appointees (data and weights updated each December). Table 4 reports the results.¹⁵

Model 3.1 shows that proportionality of DAS appointments has an effect 1.5 times as large as cabinet proportionality on coalition discipline. We are able to compare coefficient sizes directly because the underlying variable scale is the same (0–1). However, an F-test does not show a statistically significant difference between the coefficients for DAS proportionality and cabinet proportionality. Model 3.2 provides more robust findings, evaluating the effect of the proportionality of the highest level APAs (PROPORTIONALITY_DAS6). These positions wield the greatest budgetary and policy influence, so are presumably worth the most to politicians seeking access to executive patronage. Model 3.2 demonstrates that in the aggregate, DAS6 appointments are more than three times as important as ministerial allocations to coalition discipline. An F-test confirms that the difference between the coefficients is statistically significant. While cabinet proportionality certainly matters, the proportionality of the patronage coalition, especially in its upper echelons, provides an even more substantial gain.

These findings reveal the importance of incorporating the patronage coalition into models of executive-legislative relations in coalitional systems. To underscore the magnitude of the findings, consider the empirical reality in Brazil during this period. Model 3.2 shows a coefficient of 0.537 for the proportionality of DAS6 appointments. Moving from the minimum to the maximum observed values of DAS6 proportionality during this period would imply a gain in coalition discipline of 13%, all else being equal. If the coalition size were 350 deputies, this would be the equivalent of bringing 45 new legislators onside.

Evaluating party-level payoffs

The final stage of empirical analysis moves to the party level. Throughout, we are guided by Amorim Neto’s hypothesis “that individual parties will cooperate with the president if they see positive benefits for doing so” (2002, 70). He further suggests that different parties may have different objectives for participating in the governing coalition, which may mean that their perceived payoff varies in terms of cabinet seats, calculations about the next election, and ideology. This stage of analysis therefore tests how individual parties’ support is affected by these distinct payoff structures.

Amorim Neto’s view has been influential over the past two decades, appearing in party-oriented perspectives that focus on why parties join the governing coalition (Amorim Neto 2019, Freitas 2016, Inácio 2013). But as the previous analysis showed, these spoils of executive power are not solely cabinet positions: APAs play a role. For that reason, we revise Amorim Neto’s regression of the independent variables relating to a party’s support to the president in roll call votes to incorporate the patronage coalition:

SUPPORT_i = $β_{0}$ + $β_{5}$ CABPAYOFF + $β_{6}$ PATPAYOFF + $β_{7}$ TIMEREMAIN + $β_{8}$ DIST + $ε$

We analyze the three most important parties in the governing coalition of the Workers’ Party-led governments between 2003 and 2015: PMDB, PT, and PP. SUPPORT is the percent of the delegation voting in accordance with the government leader. CABPAYOFF is the cabinet payoff ratio for each of the three parties, calculated as the proportion of ministerial seats to legislative seats. PATPAYOFF is the payoff ratio for participating in the administration coalition, calculated as the proportion of DAS appointments to legislative seats.

The literature has found ambiguous results for cabinet payoff on individual party support. In his analysis of 1989–1998, Amorim Neto (2002, 73–74) finds strong evidence of such an effect. By contrast, in analyzing the 1989–2016 period, Almeida (2018, 188) reports no systematic effect. Our findings for the 2003–14 period demonstrate the same ambiguity with regard to the cabinet payoff: a model run with cabinet payoff but without patronage payoff (not reported here) leads to a mix of signs across the cabinet payoff coefficient and only limited statistical significance. Table 5 reports the results when cabinet payoff is incorporated concurrently with patronage payoff, showing that the cabinet payoff coefficient varies in signs and is only significant in one case.

Table 5.

Support by Individual Parties, 2003–2014. DV: Support for President in Roll Call Votes.

	Models using ideological range			Models using ideological heterogeneity
Independent variables	PMDB	PP	PT	PMDB	PP	PT
Cabinet coalition payoff	−0.0001	0.068	−0.122^***	0.015	0.057	-0.140^***
Patronage coalition payoff	0.309^**	0.555^***	0.071^**	0.263^**	0.538^***	0.059^*
Time remaining in term	0.0001^***	0.0001^***	0.0001^***	0.0001^***	0.0001^***	0.0001^***
Ideological range	−0.389^*	−0.146	−0.176
Ideological heterogeneity				-0.993^**	-0.568	-0.448
Constant	1.018^***	0.662^**	1.108^***	0.669^***	0.579^***	0.995^***
Adj. R-squared	0.058	0.053	0.043	0.056	0.050	0.038
N	830	830	857	830	830	857

Note: ^***p < 0.01; ^**p < 0.05; ^*p < 0.1.

Why might this be the case? Cabinet payoff may be simply too crude a measure of the benefits of participating in government: there are not enough cabinet slots to go around and not all parties may wish to hold ministries. Party leaders are the most direct likely beneficiaries of cabinet seats, and the desires of backbenchers are only indirectly and partially filled by meeting party leaders’ needs. The effect of cabinet payoffs may be limited to only a segment of the party, especially in large or internally factionalized parties. Additionally, because elevation to the cabinet greatly enhances the electoral opportunities for politicians, it simultaneously includes and excludes ambitious politicians from the same party, meaning that its positive influence on one faction may be offset by its negative effects on others.¹⁶

These inconclusive results for cabinet payoff make it even more notable that the coefficient for patronage payoff is consistently significant and substantive.¹⁷ To illustrate the magnitude of the patronage payoff coefficients, consider the case of the PMDB. Transitioning from minimum to maximum patronage payoff would imply, other things equal, an increase of legislative support of 11.6% by the PMDB, or from 60.09% to 67.03% of the party’s legislative membership. The PMDB held a minimum of 71 seats in the chamber over this period, meaning that the increased support would be the equivalent of eight more legislators voting with the government. A similar calculation for the PT would mean an additional six legislators (7.9% of their minimum delegation) voting with the government, while for the PP the figure is eight more legislators (20.9% of their minimum delegation).

Conclusion

This paper has demonstrated that the distribution of administrative political appointments has significant effect on the coalescence of coalition allies. This is not a trivial issue: presidents who lack a unified floor coalition are unlikely to get very far on their legislative goals, and as Amorim Neto (2019) points out, may even be unable to finish their terms. Our findings reinforce previous results about the political uses of the distribution of cabinet ministries in presidential regimes (e.g. Amorim Neto 2002, Altman 2000, Martinez-Gallardo 2014). But they also provide robust support for a growing literature on the broader toolbox of minority presidents in coalitional presidentialism, showing that management of the patronage coalition is an important and hitherto underappreciated tool. Our results show that patronage is a tool that is even more effective than the cabinet coalition in determining legislative support.

A decade ago, Lewis (2011, 60) noted that “although there is substantial evidence that presidents can use appointees to accomplish policy goals, very little research examines whether presidents accomplish other goals through personnel...[D]o...members of Congress whose appointment requests were honored by the president support the president more frequently?” This paper’s answer is strongly affirmative. Actively managing the patronage coalition is an essential tool for cementing the legislative coalition that contributes to the president’s success and survival.

The findings in this paper will shift our understanding of coalitional presidentialism away from the rarefied cabinet boardroom down into the wheelhouse of government bureaucracy. Future research might examine a number of related issues. It would be useful to find ways to further distinguish between the uses of appointments for policy, pork, and patronage. Scholars might consider how appointments are layered within and across agencies so that presidents ensure a modicum of policy coherence. Our understanding of how the partisan distribution of appointments is structured in ways that simultaneously reward and control coalition allies is still embryonic. Given space constraints, we have elided a few variables that would be worth incorporating in future analyses, including changes in the rules governing budget amendments and executive decrees, which may alter parties’ strategic incentives (Rennó 2010). Measurement issues have led us to omit APAs serving in state-owned banks and state-owned enterprises outside the executive bureaucracy (Chaisty et al. 2018). Scholars have evaluated the salience of particular ministerial portfolios (Batista 2018, Mauerberg and Pereira 2020, Zucco, Batista, and Power 2019), and similar evaluation of the relative salience of subministerial agencies headed by APAs would likely yield valuable insights. Finally, it would be useful to normatively evaluate the costs of these coalitional strategies. While this paper has demonstrated the benefits of the patronage coalition in terms of legislative coalition-building, it is still an open question whether the costs—whether from coordination problems, the expense of patronage, or possible inefficiencies—are outweighed by the potential governance gains from legislative cohesion.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Katherine Bersch

Notes

References

Alemán

Eduardo

Tsebelis

George

. 2011. “Political Parties and Government Coalitions in the Americas.” Journal of Politics in Latin America 3, 1: 3–28.

Almeida

Acir dos Santos

. Governo presidencial condicionado: delegação e participação legislativa na Câmara dos Deputados. Ph.D. Thesis, Universidade do Estado do Rio de Janeiro, 2018.

Altman

David.

2000. “The Politics of Coalition Formation and Survival in Multiparty Presidential Democracies.” Party Politics 6, 3: 259–283. https://doi.org/10.1177%2F1354068800006003001

Amorim Neto

Octavio

. 2002. “Presidential Cabinets, Electoral Cycles, and Coalition Discipline in Brazil.” In Legislative Politics in Latin America, edited by Scott

Morgenstern

Benito

Nacif

, eds, 48‐78. Cambridge: Cambridge University Press.

Amorim Neto

Octavio

. 2019. “Cabinets and Coalitional Presidentialism.” In Routledge Handbook of Brazilian Politics, edited by Barry

Ames

, ed, 293‐312. New York: Routledge.

Anonymous . 2013. High Level Official in President's Casa Civil Ministry. Author interview, July 30.

Batista

Mariana

. 2018. “Who Gets What and How Does It Matter? Importance-Weighted Portfolio Allocation and Coalition Support in Brazil.” Journal of Politics in Latin America 10, 3: 99‐134. https://doi.org/10.1177%2F1866802X1801000304

Bersch

Katherine

Praça

Sérgio

Taylor

Matthew M

. 2017a. “State Capacity, Bureaucratic Politicization, and Corruption in the Brazilian State.” Governance 30, 1: 105‐124.

Bersch

Katherine

Praça

Sérgio

Taylor

Matthew M

. 2017b. “State Capacity and Bureaucratic Autonomy within National States: Mapping the Archipelago of Excellence in Brazil.” In State Building in the Developing World, edited by Miguel

Angel Centeno

Atul

Kohli

Deborah

Yashar

, eds, Vol. 157-83. Cambridge: Cambridge University Press.

10.

Bertholini

Frederico

Pereira

Carlos

. 2017. “Pagando o preço de governar: Custos de gerência da coalizão no presidencialismo brasileiro.” Revista de Administração Pública 51, 4Jul-Aug: 528‐550. DOI: 10.1590/0034-7612154969

11.

Browne Eric

Mark Franklin

1973. “Aspects of Coalition Payoffs in European Parliamentary Democracies.” American Political Science Review 67: 453‐469. DOI: 10.2307/1958776

12.

Bucur

Cristina

. 2020. “Presidents and Cabinet Payoffs in Coalition Governments.” Political Studies Review 18, 1: 30‐52. https://doi.org/10.1177%2F1478929919856875

13.

Carazza

Bruno

. 2018. Dinheiro, eleições e poder: As engrenagens do sistema político brasileiro. SP: Companhia das Letras.

14.

Carey

John M.

Gina

Yannitell Reinhardt

2004. “State-Level Institutional Effects on Legislative Coalition Unity in Brazil.” Legislative Studies Quarterly 29, 4: 23‐47. DOI: 10.1590/S0011-52582003000400005

15.

Carey

John.

2000. Party and Coalition Unity in Legislative Voting. St. Louis: Department of Political Science, Washington University. Unpublished Working Paper 376.

16.

Chaisty

Paul

Chernykh

Svitlana

. 2017. “How Do Minority Presidents Manage Multiparty Coalitions? Identifying and Analyzing the Payoffs to Coalition Parties in Presidential Systems.” Political Research Quarterly 70, 4December: 762‐777.

17.

Chaisty

Paul

Cheeseman

Nic

Timothy Power

. Coalitional Presidentialism in Comparative Perspective: Minority Presidents in Multiparty Systems. Oxford: Oxford University Press, 2018.

18.

Chaisty

Paul

Cheeseman

Nic

Power

Timothy

. 2014. “Rethinking the 'Presidentialism Debate': Conceptualizing Coalitional Politics in Cross-Regional Perspective.” Democratization 21, 1: 72‐94. DOI: 10.1080/13510347.2012.710604

19.

Cheibub José

Antonio

Przeworski Adam M

Saiegh Sebastian

. 2004. “Government Coalitions and Legislative Success under Presidentialism and Parliamentarism.” British Journal of Political Science 34, 4October: 565‐587.

20.

Costa

Raymundo

. 2004. Racha no PMDB dificulta apoio total. Folha de S. Paulo. January 6.

21.

Figueiredo

Angelina Cheibub

Limongi

Fernando

. 2000. “Presidential Power, Legislative Organization, and Party Behavior in Brazil.” Comparative Politics 32, 2: 151‐170.

22.

Figueiredo

Angelina Cheibub

Canello

Júlio

Vieira

Marcelo

. 2012. “Governos minoritários no presidencialismo latino-americano: determinantes institucionais e políticos.” Dados 55, 4: 839‐875. DOI: 10.1590/S0011-52582012000400001

23.

Figueiredo

Argelina Cheibub

Limongi

Fernando

. 1999. Executivo e legislativo na nova ordem constitucional. Rio de Janeiro: Editora FGV.

24.

Figueiredo

Argelina Cheibub

Limongi

Fernando

. 2002. “Incentivos eleitorais, partidos e política orçamentária.” Dados 45, 2: 303‐344. DOI: 10.1590/S0011-52582002000200005

25.

Freitas

Andréa

. 2016. O presidencialismo da coalizão. Rio de Janeiro: Konrad Adenauer Stiftung.

26.

Gamson

William

. 1961. “A Theory of Coalition Formation.” American Sociological Review 26, 3June: 373‐382.

27.

Gaspari

Elio

. 1997. Cinco vinhetas do balcão de FFHH. Folha de S. Paulo. January 29.

28.

Grindle

Merilee S.

2012. Jobs for the Boys: Patronage and the State in Comparative Perspective. Cambridge, MA: Harvard University Press.

29.

Hassan

Mai

Sheely

Ryan

. 2016. “Executive-Legislative Relations, Party Defections, and Lower Level Administrative Unit Proliferation: Evidence from Kenya.” Comparative Political Studies 50, 12: 1595‐1631.

30.

Hunter Wendy J

Power Timothy

2019. “Bolsonaro and Brazil’s Illiberal Backlash.” Journal of Democracy 30, 1: 68‐82.

31.

Inácio

Magna

. 2013. “Escogiendo ministros y formando políticos: los partidos en gabinetes multipartidaristas.” América Latina Hoy 64: 41‐66. DOI: 10.14201/alh.10231

32.

Koppel

Moshe

Diskin

Abraham

. 2009. “Measuring Disproportionality, Volatility, and Malapportionment: Axiomatization and Solutions.” Social Choice and Welfare 33, 2: 281‐286. DOI: 10.1007/s00355-008-0357-1

33.

Lewis

David E.

2008. The Politics of Presidential Appointments: Political Control and Bureaucratic Performance. Princeton: Princeton University Press.

34.

Lewis

David E.

2011. “Presidential Appointments and Personnel.” Annual Review of Political Science 14, 1June: 47‐66. DOI: 10.1146/annurev-polisci-042009-121225

35.

Limongi

Fernando

Cortez

Rafael

. 2010. “As eleições de 2010 e o quadro partidário.” Novos Estudos CEBRAP 88: 21‐37. DOI: 10.1590/S0101-33002010000300002

36.

Lipsmeyer

Christine S.

Pierce.

Heather Nicole

2011. “The eyes that bind: Junior ministers as oversight mechanisms in coalition governments.” The Journal of Politics 73, 4: 1152‐1164. DOI: 10.1017/S0022381611000879

37.

Lopez

Felix

Silva

Thiago

. 2019. “Party affiliation and appointment for positions in the Brazilian Federal Bureaucracy (1999-2018).” Revista de Administração Pública 53, 4: 711‐731.

38.

Maia

Caio

. 2004. Só participa do governo quem tem fidelidade "canina", diz deputado petista. Folha de S. Paulo. February 5.

39.

Mainwaring

Scott

. 1993. “Brazilian Party Underdevelopment in Comparative Perspective.” Political Science Quarterly 107, 4Winter: 677‐708.

40.

Mancuso

Wagner Pralon

Speck

Bruno Wilhelm

. 2015. “Financiamento empresarial na eleição para Deputado Federal (2002-2010): determinantes e consequências.” Teoria & Sociedade 23, 2Jul-Dec: 1‐23.

41.

Mancuso

Wagner Pralon

. 2012. “Empresas e financiamento de campanhas eleitorais de candidatos a deputado federal pelo estado de São Paulo nas eleiçoes de 2002 e 2006.” In Desenvolvimento e crise na América Latina: estado, empresas e sociedade, edited by Dalla Costa

Minella

A.C.

Gros

D.B.

Mick

Etcheverry

M.S.

Iglecias

, eds Curitiba: CRV.

42.

Martin Lanny

Stevenson Randolph

2001. “Government Formation in Parliamentary Democracies.” American Journal of Political Science 45, 1January: 33‐50.

43.

Martinez-Gallardo

Cecilia

. 2014. “Designing Cabinets: Presidential Politics and Ministerial Instability.” Journal of Politics in Latin America 6, 2: 3‐38.

44.

Mauerberg

Arnaldo

Jr. Pereira.

Carlos

2020. “How Valuable is a Presidential Cabinet? Measuring Ministries’ Political in Brazil Attractiveness.” Latin American Politics and Society 62, 1February: 25‐45. DOI: 10.1017/lap.2019.44

45.

Mejía Acosta

Andrés

. 2006. “Crafting Legislative Ghost Coalitions in Ecuador: Informal Institutions and Economic Reform in an Unlikely Case.” In Informal Institutions and Democracy: Lessons from Latin America, edited by Gretchen

Helmke

Steven

Levitsky

, eds, 69‐86. Baltimore, MD: Johns Hopkins University Press.

46.

Mershon

Carol

. 2001. “Contending Models of Portfolio Allocation and Office Payoffs to Party Factions: Italy, 1963-79.” American Journal of Political Science 45, 2April: 277‐293.

47.

Morelli

Massimo

. 1999. “Demand competition and policy compromise in legislative bargaining.” American Political Science Review 93, 4December: 809‐820.

48.

Panizza

Francisco

Peters Guy

Ramos Larraburu Conrado

2018. “Roles, Trust and Skills: A Typology of Patronage Appointments.” Public Administration 97, 1: 147‐161. DOI: 10.1111/padm.12560

49.

Patury

Felipe

Loyola

Leandro

. 2011. A lista secreta de Dilma. Época.

50.

Pereira

Carlos

Batista

Mariana

Praça

Sergio

Lopez

Felix

. 2017. “Watchdogs in our Midst: How Presidents Monitor Coalitions in Brazil's Multiparty Presidential Regime.” Latin American Politics and Society 59, 3: 27‐47.

51.

Pereira Carlos J

Power Timothy

Raile Eric

2011. “Presidentialism, Coalitions, and Accountability.” In Corruption and Democracy in Brazil: The Struggle for Accountability, edited by Power Timothy J. and Taylor Matthew M., eds, 31‐55. Notre Dame: University of Notre Dame Press.

52.

Pérez-Liñan

Aníbal

. 2007. Presidential Impeachment and the New Political Instability in Latin America. Cambridge: Cambridge University Press.

53.

Peters

B. Guy

Pierre

Jon

. 2015. Politicization of the Civil Service in Comparative Perspective: the Quest for Control. New York: Routledge.

54.

Power

Timothy J.

Zucco

Cesar

. 2009. “Estimating Ideology of Brazilian Legislative Parties, 1990-2005: A Research Communication.” Latin American Research Review 44, 1: 218‐246.

55.

Power

Timothy J.

Zucco

Cesar

. 2014. Brazilian legislative surveys (1990-2013).” Database on Harvard Dataverse. DOI: 10.7910/DVN/T2PW7F. accessed July 31, 2019.

56.

Raile Eric

Pereira Carlos J

Power Timothy

2011. “The Presidential Toolbox: Building Legislative Support in a Multiparty Presidential Regime.” Political Research Quarterly 64, 2: 323‐344. DOI: 10.1177/1065912909355711

57.

Rennó

Lúcio

. 2010. “Executive-Legislative Relations in Brazil: Is 2009 the First Year of the Rest of Our Lives?” Revista de Ciencia Política 30, 2: 213‐230.

58.

Rose

Richard.

1984. “Electoral Systems: A Question of Degree or of Principle?” In Choosing an Electoral System: Issues and Alternatives, edited by Arend

Lijphart

Bernard

Grofman

, eds New York: Praeger.

59.

Salomon

Marta

. 1997. FHC detona caça ao voto de 100 deputados. Folha de S. Paulo.

60.

Samuels

David J.

2002. “Pork Barreling is not Credit Claiming or Advertising: Campaign Finance and the Sources of the Personal Vote in Brazil.” Journal of Politics 64, 3August: 845‐863. https://www.jstor.org/stable/1520116

61.

Shugart

Matthew Soberg

John M

Carey

. Presidents and Assemblies: Constitutional Design and Electoral Dynamics. New York: Cambridge University Press, 1992.

62.

Taylor

Matthew M.

2020. Decadent Developmentalism: The Political Economy of Democratic Brazil. Cambridge: Cambridge University Press.

63.

Thies

Michael

. 2000. “Keeping Tabs on One’s Partners: The Logic of Delegation in Coalition Formation.” American Journal of Political Science 45, 3: 580‐598.

64.

Toledo

José Roberto de

. 1996. PFL pede a presidente fim de ameaças a deputados. Folha de S. Paulo.

65.

Vasselai

Fabrício

Mignozzetti

Umberto

. 2014. “O efeito das emendas ao orçamento no comportamento parlamentar e a dimensão temporal: velhas teses, novos testes.” Dados 57, 3Jul-Sep: 817‐853.

66.

Verzichelli

Luca.

2008. “Portfolio Allocation.” In Cabinets and Coalition Bargaining: The Democratic Life Cycle in Western Europe, edited by Stom

Kaare

Müller

Wolfgang C.

Bergman

Torbjorn

, eds, 237‐265. New York: Oxford University Press.

67.

Warwick Paul

Druckman James

2001. “Portfolio Salience and the Proportionality of Payoffs in Coalition Governments.” British Journal of Political Science 31, 4: 627‐649. http://www.jstor.org/stable/3593295

68.

Warwick Paul

Druckman James

2006. “The Portfolio Allocation Paradox: An Investigation into the Nature of a Very Strong but Puzzling Relationship.” European Journal of Political Research 45, 4: 635‐665.

69.

Zucco

Cesar

. 2009. “Ideology or What? Legislative Behavior in Multiparty Presidential Settings.” The Journal of Politics 71, 3July: 1076‐1092. DOI: 10.1017/S0022381609090896

70.

Zucco Cesar Batista Mariana J

Power Timothy

2019. “Measuring Portfolio Salience Using the Bradley-Terry Model: An Illustration with Data from Brazil.” Research & Politics 6, 1: 1‐8.

71.

Zucco

Cesar

Benjamin

Lauderdale

. Distinguishing Between Influences on Brazilian Legislative Behavior. Legislative Studies Quarterly 2011; 36(3 August): 363–396, doi:10.1111/j.1939-9162.2011.00019.x.