Abstract
Latin American legislatures have gone largely unstudied, with the functioning of the Argentine Chamber of Deputies prior to the 1980s being an entirely unexplored subject. This paper fills that gap by examining the organization of the Chamber, with particular focus on its standing committee system from 1946 to 2001. We assess the portability of two U.S.-based theoretical approaches to legislative organization by applying them to committee assignments. An original data set of Argentine deputies was constructed and a way of measuring political power in committees was devised for this study. Despite weak democratic governments, military interventions, and changes to the electoral system, we find that ruling parties have consistently influenced the committee system, shaping its structure and securing an over-proportion of their deputies in key committee positions. These results support the applicability of the U.S. originated Cartel Theory of legislative organization to understanding and studying legislatures outside that country.
Introduction
With the exception of the United States Congress, the activities of most legislatures go un-theorized, and in the case of Latin America, understudied. As scholars have slowly begun to fill that void, debate has emerged as to whether theories developed to study the appointment process in the stable U.S. House of Representatives can be applied to other countries, particularly countries where the legislatures have an intermittent committee system composed mostly of inexperienced members.
The Argentine Chamber of Deputies is a particularly interesting case study as it has been closed, due to military coups, five times since the 1930s. Its instability and mostly neophyte membership offer an opportunity to test the main U.S. originated theories of legislative organization in order to see whether they are applicable to legislatures outside of the United States. In the process, we can increase our understanding of how the Argentine Chamber of Deputies operates.
There has been only limited attention paid by the academic community to Argentine legislative politics, and what little study has been done focuses only on the period of legislative stability since the end of the last military government in 1983 (e.g., Jones 2002; Mustapic 2002; Saiegh 2004; Alemán 2006; Spiller and Tommasi 2007). Within this limited literature, few studies examined the internal organization of the Chamber; some of them applied the U.S.-based Cartel Theory to their analysis of party discipline using roll-call votes from the 1989–2003 period (Jones and Hwang 2003; 2005); and only one, to our knowledge, examined the committee system during the 1985–1997 period (Jones et al. 2002).
With the goal of making sense of the organization of the Chamber through time, this paper 1 explores the structure of its standing committee system for the period 1946–2001. Three aspects characterize our work. First, we assessed the relevance of the main claims of two central theoretical approaches used to analyse the organization of the U.S. House of Representatives (i.e. the committee-government and the party-centred models) to explain the appointment process in the Argentine lower chamber. Second, we constructed an original data set of Argentine deputies. This data set comprises, for each unit of observation (deputy/ legislative period), variables for party affiliation, region, number of terms, committee assignments, and positions in the Chamber, within caucuses, and in committees. Third, we created an original model to test the applicability of the two theories. Our methodological strategy considers the influence each deputy exerts within the committee system, accounting for the legislative importance of both committees and authority positions. This is a unique approach to the study of committees, and its strength above previous tests lies in the fact that it accommodates the political reality that committees and committee positions have different importance in terms of influence and status.
We thank the Social Sciences and Humanities Research Council of Canada for research support, André Blais, Bruce M. Hicks, and one of the anonymous reviewers for useful comments.
Our findings are consistent with the partisan view of legislative organization described by Cox and McCubbins (2007 [1993]) who brought forth the idea that political parties are invented to solve the collective dilemma that legislators face. In this sense, despite contextual and institutional differences between the U.S. House of Representatives and the Argentine lower chamber, we found that Argentine party leaders managed to control their committee system and secured a firm grip on key committee positions.
The paper is organized similar to the approach just outlined. Section 1 describes the institutional context of the Chamber and presents a qualitative analysis that conveys preliminary evidence of the influence of parties over the committee system. Before assessing our empirical evidence more systematically, section 2 introduces the theoretical ground that underlies the predictions. Section 3 outlines the data and explains the methodology. Our empirical findings are presented and discussed in section 4. A final section concludes.
Unlike studies on the U.S. House of Representatives (hereafter the U.S. House), it is necessary to examine the institutional characteristics of the understudied Argentine Chamber of Deputies (hereafter the Chamber) 2 before any testing of theories on organization can be undertaken. In this section we provide basic background information on contemporary Argentine politics. Then, we turn to a qualitative assessment of the evolution of the committee system and to a description of rules and practices on committee assignments, highlighting the main differences between the Chamber and the U.S. House.
Except otherwise indicated, all data employed in this article was obtained from official publications of the Argentine Chamber of Deputies: a) Rules of the Chamber, 1955 to 2000 editions (Reglamento de la Honorable Cámara de Diputados de la Nación); b) Composition of the Chamber and Committee Membership, 1964 to 2000 editions (Cámara de Diputados de la Nación – Composición y Comisiones), and c) Journals of Sessions (Journal), printed editions – 1946 to 1997– (Diario de Sesiones de la Honorable Cámara de Diputados de la Nación), and online debates – 1998 to 2001 –, online: <http://www.hcdn.gov.ar/dependencias/dtaquigrafos/frames.html> (Versiones Taquigráficas Online).
The timeframe covered by this paper includes periods of democratic and unelected de facto governments. The main political parties in the period were the middle-class Radical Civic Union (Unión Cívica Radical – UCR), and the urban working-class and petite bourgeoisie Peronist/ Justicialist Party (Partido Peronista/Justicialista – PJ).
As Table 1 shows, ten presidential elections were held from 1946 to 2001 – six Presidents belonged to the PJ, and four to the UCR – and only three elected Presidents completed their terms in office. For de facto governments, the armed forces held power for four intermittent periods that covered almost 19 years. According to the country's Constitution, Argentina is a federal republic consisting of 23 provinces and an autonomous Federal Capital. It has a presidential regime and a bicameral legislature. The 257 deputies in the Chamber are elected for four-year terms from 24 multimember districts. One-half of the Chamber is renewed every two years. Like U.S. representatives, Argentine deputies can be re-elected indefinitely.
Democratic Presidential Terms in Argentina, 1946–2001
Democratic Presidential Terms in Argentina, 1946–2001
Source: Own elaboration upon data from Potash 1959, Hodges 1988, and Jones 2002.
From 1946 to 2001, three electoral systems determined the allocation of seats in the Chamber. The first system (incomplete list) gave two-thirds of the seats to the party that got the most votes in a district, and the remaining third to the second largest party (1946–51/1958–62). 3 The second was the plurality system (Law 14.032), in effect from 1951 to 1955. Since 1963, seats are assigned via closed-list proportional representation (PR) system using the d'Hondt formula, with a legal threshold of 3 percent of the registered voters in each district.
This system is known as the “Sáenz Peña Law”, named after the President who introduced it in 1912 (President Roque Sáenz Peña).
The number of deputies changed through time as a result of constitutional reforms, census updates, the achievement of provincial status by national territories, and de facto decrees. The successive re-openings and midterm renewals of the Chamber led to 18 legislatures in the period 1946–2001 (Table 2). 4 These legislatures differ not only in their membership but also in the length of their terms. Constitutional reforms and military coups explain these differences.
Number of Deputies and Ruling Party Types in the Chamber, 1946–2001
Source: Own elaboration upon data from Potash 1959, Hodges 1988, and Jones 2002. For the codification of RP types see Appendix A.
We use the label “legislatures” to refer to each of these 18 compositions.
Different ruling parties (hereafter RPs) were in charge of the 18 legislatures. The RP refers to the party or coalition in control of the Chamber. We use this label because, in contrast with the U.S. House, which has always been controlled by a “majority party”, the Chamber has been ruled by eight single-party majorities, seven single-party pluralities, two majority coalitions, and one plurality coalition (Table 2). 5 Peronists as well as Radicals were four times the single-party majority in Chamber. Two Peronist majority coalitions controlled the Chamber (1946–1948; 1973–1976). The only plurality ruling coalition was the ALIANZA (1999–2001). Peronists and Radicals were the single-RP plurality four and three times respectively.
See our codification of ruling party types in Appendix A.
The Development of the Committee System
Reviewing the studies on committee size manipulation in the U.S. Congress, Eulau states that “[p]olitical units are more likely to increase than decrease in size, and efforts to reduce their size are invariably offset by tendencies toward growth, as if growth were a law of nature” (Eulau 1984: 592). If we were to find confirming evidence of such a law of committee growth, and of the influence of RPs on the structuring of a committee system, a good place to start would be the Chamber. The following data tells the story.
From 1946 to 2001 the number of standing committees increased from 19 to 45. This expansion was moderate from 1946 to 1976, when nine committees were created and two were dissolved. From 1983 to 2001, by contrast, the committee system increased by 19 units. The period 1990 to 1998 experienced the greatest expansion (14 committees). RPs controlled the creation of committees during the whole period (members of RPs authored all but one resolution for the creation of committees), chaired all new committees until 1976 and 68 percent of the committees created after 1983. From 1983 to 2001, committees were created by occasionally unclear summary proceedings, most of which occurred after the Chamber's biannual renewal (during the negotiation of committee assignments). The realpolitik reasons that justified their creation (e.g., the need to carry out political deals about the distribution of Chamber positions) were explicitly stated by RP leaders. 6
See the creation of standing committees in 1987 and 1998 (resolution 2353-D-87 and speech of Eduardo Camaño (PJ) in the session of 11 March 1998). Before 1983, the reasons that justified the creation of committees were the aim at dealing with social problems or at matching the ministries of the Executive Power.
In the Chamber, committee size is established in the Reglamento (the equivalent to the U.S. Rules of the House). From 1946 to 1973, committees were assigned a fixed number of deputies. Since 1983, however, minima and maxima have been established. While the number of deputies increased by 63 percent during these 55 years (from 158 to 257), the majority of committees have almost tripled their membership from a mean of nine members in 1946 to 26 in 2001. The committees whose membership has increased the most are Budget (41 members), Foreign Affairs (37) and Constitutional Affairs (33) – the most important committees of the Chamber. 7 This expansion at the top of the committee system contrasts with the U.S. House, where enlargement is not concentrated among the control committees (i.e. Rules, Ways and Means, Appropriations, and Budget). Even though RPs controlled the enlargement of committee size during the whole period, differences exist both before and after 1983. The increase in the number of deputies in 1955, 1963, and 1983 (Table 2) was followed by the enlargement of committee size. By contrast, the three consecutive enlargements of 1994, 1998, and 2000 did not follow any increase in the number of deputies. This suggests that RPs may have had strategic reasons to modify the committee structure. These issues are addressed in the following sections.
See our stratification of committees based on their size and the number of authority positions in Appendix A.
Also in contrast with the U.S. House, there is neither a formal stratification of committees in the Chamber (i.e. there are no “exclusive”, “semiexclusive”, or “nonexclusive” committees, and no sub-committees), nor are there any restrictions regarding multiple assignments. In fact, multiple assignments are the norm in Argentina. Our case goes well beyond the observations of Gawthrop (1966), who saw a gradual but steady increase in the number of double committee assignments in the U.S. House (1947–1966). Indeed, while the average number of assignments by deputy is 2.6 over the whole period in the Chamber (1946–2001), assignments per deputy increased from 1.05 in 1948–1952 to 4.71 in 1999–2001.
Besides multiple assignments and the increase in the number and size of committees, the 1946–2001 period is also marked by increases in the number of committee leadership positions. Whereas in 1946 committees had one chair and one secretary, in 2001 they had one chair, two vice-chairs, and three secretaries (important committees had four secretaries). The proliferation of these positions began in 1963 with the creation of a vice-chair (VC1) for all committees, and had a peak in 1987, when a second vice-chair (VC2) and a second secretary for all committees were approved. RPs stated the need to improve the direction of committees as the reason for creating these positions. This reason is curious, particularly if we consider that the Reglamento establishes no functions for vice-chairs and secretaries and that, in practice, they have no decision-making power. Which parties benefited from the creation of these positions? RPs secured over-representation in the VC2 (more than 70 percent), leaving most of the VC1 to the caucus of the second party – in number of deputies – (1987–2001), and received the lion's share of the secretaries (1963–2001).
According to the Reglamento (rule 29), every two years during the preparatory sessions, deputies decide whether the allocation of committee seats will be made “by the Chamber,” or by “its President” (i.e. the President of the Chamber and not the country, a position equivalent to that of the Speaker of the U.S. House). 8 Committee assignments will be made “to the fullest extent feasible” in proportion to the seats held by parties in the Chamber (rule 105). Once committee assignments are made, each committee is to elect its authorities with a plurality of votes (rule 106).
The main activities of the preparatory sessions are the Chamber's partial renewal and the election of its authorities.
It should be noted that, despite several reforms to the Reglamento, those three rules were never modified. In addition, those are the only rules that guide the assignment process. In contrast with the U.S. House, in the Chamber there are no initial freshman committee assignment requests or non-freshman transfer requests. There is neither a committee on committees devoted to preparing the assignment slate for each party, nor caucus or floor approval of the committee nominations.
The 18 preparatory sessions and the debates about committee assignments that took place between 1946 and 2001 show three features of this process: a) Presidents (of the Chamber) are always entitled by the Chamber to allocate committee seats (rule 29); b) once the Chamber approves this delegation, the process is controlled by its President; c) this delegation implies consultation with party leaders (we call features b and c the “non-codified phase of the process”).
Presidents of the Chamber are the most influential actors of the institution. They exercise legislative, administrative, and agenda-setting functions that are uncommon for the presiding officer of a lower chamber in the 21st Century. Except for the right of removal, which formally resides in the Chamber (rule 107), as we will see with respect to appointments, the Chamber functions like the ‘pre-revolt’ U.S. House. 9
We refer to the period preceding the revolt against Speaker Joseph G. Cannon (R-Ohio, 1903–1911), when speakers were entitled to make the committee assignments, remove committee personnel, and considered seniority as only one of several criteria (Cox and McCubbins 2007 [1993]; Groseclose and King 2001).
Regarding the non-codified phase of the process, Jones et al. note that the President of the Chamber
decides, in consultation with [party] leaders, which committee leadership positions [chair, vice-chair and secretary] correspond to which parties. Once this allocation is decided, each party's leadership distributes its committee assignments (Jones et al. 2002: 660).
An analysis of the Chamber's debates allows us to make five observations with respect to this phase:
Consultation is discretionary: it depends on who sits in the President's chair. In 1952, a UCR deputy addressed the President (PJ) and claimed that “it is understood that this authorization [given by the Chamber to the President to assign committee positions] implies a consultation with party leaders.” The President answered that this consultation was “a simple parliamentary practice […], if it were an imposed condition, it would limit the authorization the Chamber gives to the Presidency.” 10
Our translation from Journal of Sessions 05-05-52, p. 57.
There is no agreement on the subject matter of consultation. Deputies sometimes expect consultations to fix or change committee ratios, or to name the specific deputies proposed in each committee. 11 In 1987, the President denied having the responsibility to allocate committee leadership positions, explaining that this was not an aspect of the consultations. 12
See Journals 11.05.60, pp. 45/138 and 15.06.61, p. 961.
In 1987 a problem arose regarding the number of authorities the PJ would get in committees. The PJ caucus leader José L. Manzano asked for a special session to make the appointments. President Juan C. Pugliese (UCR) explained that the Presidency had nothing to do with the appointment of committee authorities, that this was an issue each committee should arrange. Manzano said that “the PJ could not delegate committee appointments.” Federico T. Storani (UCR) motioned to respect the traditional procedure. Manzano tried to respond but he was denied the right to speak and the President closed the debate (Journal 27.11.87, pp. 4354/56).
Consultation is limited to a handful of party leaders. Although Presidents are expected to consult all party leaders, practice shows that most are completely unaware of the negotiations. Only decisive allies may be invited to have a say in assignments, especially when a plurality rules the Chamber.
Negotiations about committee leadership positions are entangled with negotiations regarding the vice-presidencies of the Chamber. Usually, leaders from minor parties denounce and criticize these agreements during the preparatory sessions. 13
See Journals 27.11.87, pp. 4345/54 and 29.11.89, pp. 5648/56, and official records of debates, online: <http://www.hcdn.gov.ar/dependencias/dtaquigrafos/frames.html> 03.12.97 and 01.12.99.
Negotiations sometimes take more time than expected. Although committee members must be appointed before the opening of ordinary sessions, it is not unusual to observe delays of two or three months.
Thus, a rule enforcing proportional representation in committees is complemented with a discretionary and non-uniform practice that gives RPs ample power to decide on assignments. Evidence of the relationship between rules and practice in the period 1946–2001 can be seen in a comparison of Tables 3 and 4. Table 3 presents RPs’ aggregate shares of committee positions as compared to their shares of seats in the Chamber. Table 4 shows the proportion of chairs in the most important committees coming from RPs. At first, the shares seem rather equal; in fact, the difference is statistically significant only for the last two legislatures (Table 3). Yet, Table 4 shows that RPs control pivotal committee seats: more than 94 percent of the chairs of the most important committees were controlled by RPs. In short, RPs rewarded their members with high-profile positions, even though the aggregate shares of assignments remained proportional to party representation.
Share of Seats vs. Share of Assignments of the Ruling Parties, 1946–2001
Note: The shares are those of the RP in the Chamber. Significance testing: Two-sample proportion tests (one-tailed);
= p < 0.05
Source: See footnote 2 of this article.
Distribution of Chairmanships in Some Important Committees, 1946-2001
Note: The chair of the Committee on Education was vacant during one legislature.
Source: See footnote 2 of this article.
Evidence regarding chairs is relevant in the Chamber for two reasons. First, in contrast to the U.S. House, where the majority always gets all chairs, opposition parties in the Chamber have held chairs since 1963. Second, the rule in force to elect chairs (rule 106) states that a plurality of votes, as opposed to a majority, is needed in the committee. Hence, we must consider that non-RP pluralities may manage to elect chairs in some committees.
In sum, this section describes how RPs controlled the development of the committee system and how they came to be over-represented in certain positions, despite the presence of a proportionality rule. In addition, it shows that Presidents have the last word on the non-codified phase of the appointment process, thus the President's final decision – adopted after discretionary consultations – does not need (and never received) floor approval. In other words, the election of committee members is controlled by RPs. The next sections are devoted to making sense of the decisions adopted in this particular process.
Can we find patterns in the appointment process notwithstanding the mix of rules and practices, and the Chamber's fluctuating institutional environment? We focus on two lines of investigation. The first is that senior deputies play a leading role in structuring an otherwise disorganized committee system. The second is that RPs control the process, securing key positions in coveted committees for their members in order to foster party discipline. Consequently, we are interested in exploring two main approaches used to analyse the organization of the U.S. House to determine whether one or either can be used to explain the organization of the Chamber. These two approaches are the committee-government and party-centred models. The former describes the U.S. House as a professionalized institution characterized by decentralized power with weak parties and leadership structures, an autonomous and strong committee system, powerful chairs, and strict adherence to seniority norms (see Nelson 1968; Huitt and Peabody 1969; Fenno 1973; Fiorina 1974; Mayhew 1974; Shepsle and Weingast 1987, 1995; and Weingast and Marshall 1988). In contrast, the latter portrays the U.S. House as an organization composed of subservient committees and a majority party as the locus of decision-making. The claim of party-centred analyses is that parties are crucial to explaining the organization of the U.S. House (see Cooper and Brady 1981; Rohde 1991; Cox and McCubbins 1994, 1997, 2002, 2005, 2007 [1993]; Aldrich 1995; Sinclair 1995; Aldrich and Rohde 2001; and Finocchiaro and Rohde 2008).
A well-known party-centred theory is the Cartel, or Party Government, Theory introduced by Cox and McCubbins in Legislative Leviathan (2007 [1993]). This work, a direct challenge to the committee-government models, presented a view of parties as solutions to collective dilemmas that rational, re-election-oriented legislators face. If they care only for re-election, the resulting lack of cohesion would adversely affect their collective fate. But legislators, the authors argue, also desire majority status for their party and internal advancement in the hierarchy of the legislature. Therefore, parties’ primary methods of solving collective dilemmas are the creation of a central authority (e.g. the Speaker) that possesses and distributes selective incentives to discourage non-cooperative behaviour, and the structuring of the committee system. It should be noted that the control of the committee system together with the scheduling power are the key structural advantages of majority parties. As the authors put it,
congressional parties are a species of legislative cartels [that] usurp the rule-making power of the House in order to endow their members with differential power (e.g., the power of committee chairs), and to facilitate and stabilize legislative trades that benefit their members (Cox and McCubbins 2007 [1993]: 257).
Regarding the committee assignment process in the U.S. House, the authors test this view of parties, providing statistical evidence about the limits of the main claims of the committee-government models (e.g., the “pure self-selection model” and the seniority norm), and analysing the tools with which majority parties influence the process (e.g., the creation and destruction of committees, regulation of their size, and control over appointments).
Considering the role played by RPs in the development of the Chamber's committee system and the substantial power its Presidents exercise over appointments (see Section I), we expect that the party-centred approach, and the Cartel Theory in particular, would have the strongest explanatory validity with respect to the committee assignment process. This expectation is also supported by evidence from recent research that a) identify the high level of party discipline and the over-representation of RPs in committees as traits of cartel behaviour (Jones and Hwang 2003; 2005); b) explain appointments and the increase in the number of committees as clientelistic rewards for service made by the executive power to secure party discipline (Mustapic 2002); and c) highlight the lack of committee autonomy and the influence of party leaders in the Chamber (Alemán 2006). 14
Jones et al. (2002) have demonstrated the inapplicability of the distributional and informational theories to the Chamber (1983–1997 period), and Jones and Hwang (2005) have shown the general applicability of the Cartel Theory to the Chamber by examining deputy behaviour through roll-call vote analysis (1989–2003 period).
Following the strategy used by Cox and McCubbins to analyse the appointment process, we first consider two expectations coming from the committee-government models that suppose a limit to the influence of party leaders: the seniority and apprenticeship norms. The former refers to committee-specific seniority and implies that the committee member of the majority (in our case ruling) party with the longest continuous service on the committee becomes chair (Cox and McCubbins 2007 [1993]: 44). The latter expectation suggests that committee appointments “operate under an ‘apprenticeship’ norm that guarantees mediocre assignments to incoming members” (Cox and McCubbins 2007 [1993]: 39). As explained below, in addition to committee-specific seniority, we consider the impact of seniority (i.e. total service in the Chamber) more generally. Then, to address assignments as an instrument of partisan control, we first test whether RPs managed to secure an over-proportional representation in important committees. Next, we consider the extent to which RPs monopolized chairmanships.
Finally, we consider the “margin hypothesis,” a prediction of Young and Heitshusen (2003), which is derived from the party-centred approach. The authors argue that RPs’ incentives to influence committee composition vary by margins of victory. This line of investigation is particularly interesting in our case because the introduction of PR in Argentina (1963) led to different margins of victory (see Table 2). Adapted to the Chamber, this expectation can be stated as follows: since pluralities and regular majorities face “more difficulties in overcoming the opposition delegation's obstructive tendencies,” they will exercise (in proportion) greater control of the standing committees’ composition than qualified or exceptional majorities do (Young and Heitshusen 2003: 663). 15
The original phrasing is as follows: “Aside from salience to the majority party, a final incentive involves margin. A majority party with a small margin faces more difficulties in overcoming the minority party's obstructive tendencies [Dion 1997] and thus has greater incentives to influence committee composition in ways favourable to the majority party [McCarty, Poole, and Rosenthal 2001]” (Young and Heitshusen 2003: 663).
Before turning to the empirical analysis, it is necessary to discuss both the data and methodology, as both were developed specifically for this research. We test the predictions using a new data set on Argentine deputies covering the 1946–2001 period. It comprises previously unreleased information on deputies (party affiliation; seniority; committee assignments; and Chamber, caucus, and committee positions) gathered during fieldwork in the Chamber (see footnote 1).
Our objective is to test the influence of RPs on the standing committee system, while accounting for the importance of each committee position. As noted, simply comparing shares of committee assignments with shares of seats, which is the usual approach of researchers, leads to misleading evidence. While proportional representation is the rule when aggregating all assignments, RPs enjoy overwhelming influence in appointing chairs and in securing appointments to the most prized committees for their deputies. As a result, considering all committee seats as equal in importance would miss the actual influence that RPs might exert over the committee system. This is why we developed a weighted measure of assignments, where the weights account for the importance of the position.
To construct this index of weighted assignments, we first rank committees according to their importance. Appendix A gives the detailed methodology used to create this measure of committee importance, as well as the resulting ranking. Second, we also create an ordinal ranking of the positions held within a committee. The four positions (regular member, secretary, vice-chair, and chair) are quantified by coding from one to four (respectively). Finally, the weighted sum of assignments for each deputy is obtained as follows:
where Pi is the power held in committees by the ith deputy, and where j stands for the jth assignment of that deputy (this summation therefore accounts for the fact that a single deputy may have multiple assignments, up to J). For instance, if deputy 1 is assigned chair of the Budget Committee, this assignment adds 22.7 points to her total score (four points for the chair position, times 5.67 points for being in the Budget Committee, as taken from Table A1 in Appendix A).
We test the predictions using the following set of explanatory variables:
Ruling Party. The ruling status of the party to which a deputy belongs (coded one for deputies from RPs, and zero for members of the opposition).
Freshman and Seniority. Either a freshman dummy that equals one if the deputy is appointed to committees for the first time (to test for an apprenticeship norm), or the count of four-year terms of the deputy (both measures were not included at the same time in the empirical models due to collinearity).
Multiple Assignments. The total number of assignments of a deputy.
Chamber Position. A control variable that measures whether the deputy holds an authority position in the Chamber (i.e. President or a vice-President of the Chamber).
Caucus Position. A control variable that measures whether the deputy holds an authority position in her caucus.
We also include legislature dummies as regressors. Those are required to correct for the temporal increase in the number of committees, so that the data from several legislatures can be pooled together. Notice that the data is not in panel format, but are rather pooled cross-sectional: since re-election is atypical in the Chamber, few deputies appear more than once in the sample.
Table 5 shows the descriptive statistics. It will be noted that our main dependent variable, the weighted sum of assignments, varies from 0 to about 40. In other words, the most disadvantaged deputies had no assignment at all in committees, whereas the most powerful managed to score 40 by accumulating important committee assignments. The average deputy has a score of 11 points. 79 percent of deputies in the sample are in their first four-year term in the Chamber. The maximum length of service for a deputy has been five terms.
Descriptive Statistics
Note: Committee-Specific Seniority is relevant only when taking assignments as the unit of analysis.
Source: See footnote 2 of this article.
Those data support the work of Jones et al. (2002), which characterized Argentine deputies as “amateur legislators”. Note that, since half of the Chamber is renewed every two years, the Freshman variable identifies the newly elected deputies more precisely (i.e. those deputies who are appointed to committees for the first time). Freshmen are a subset of the 79 percent of deputies in their first mandate.
Our main empirical tests consist of estimating OLS regressions, building upon the following baseline model:
where Pi is the committee power held by deputy i, and where […] refers to the set of above-mentioned control variables. If RPs are seizing an over-proportional share of meaningful positions in committees, then the estimate of β1 should turn out to be positive and statistically significant. Indeed, the null hypothesis is that each deputy captures a share of committee power that is proportional to the share of seats garnered by her group (either the RP, or the opposition), so that rejecting the null hypothesis would indicate that RPs are successful in securing an effective over-proportional representation.
As a final step, we replace the seniority-in-Chamber measures by committee-specific seniority, for a more robust test of the seniority norms. Instead of considering legislators as units of analysis, we focus on assignments, so that committee-specific seniority can be matched with the corresponding assignments. We then use a binary dependent variable that equals one for chair assignments, and zero otherwise. This alternative model is estimated with maximum likelihood.
In the following lines, we first report on our test of whether RPs manage to control the composition of the committee system, after accounting for the importance of each assignment. This test is implemented at the deputy level, which allows controlling for other deputy-specific characteristics, such as seniority. In the second subsection we examine the margin hypothesis. We look at detailed results depending on the majority or minority status of RPs. In the third subsection we turn our attention to chairmanships. We test whether chair assignments are determined by party affiliation, controlling this time for committee-specific experience (i.e. the seniority ranking of deputies within a specific committee).
Determinants of Committee Assignments at the Deputy-Level
Table 6 presents the results of two models predicting the weighted sum of assignments in committees at the deputy level. The first model includes Freshman as an explanatory variable, whereas the second includes Seniority, measured by the count of four-year terms. As explained in the previous section, we include legislature dummies to control for the temporal increase in the total number of positions available (although the corresponding estimates are not reported in the table).
Power Allocation in Argentine Committees, 1946–2001
Power Allocation in Argentine Committees, 1946–2001
Notes: OLS regression with the weighted sum of assignments of a deputy as the dependent variable. Legislature dummies are included in each regression.
= p < 0.001
= p < 0.01
* =p < 0.05.
Source: See footnote 2 of this article.
First, we consider whether deputies from RPs secure an over-representation in important committee positions. The positive and significant estimate for the RP variable indicates that, all else being equal, members of RPs do enjoy ascendancy in committees. This evidence supports the idea that RPs tend to have the upper hand in the strategic composition of committees, assigning super-proportions of their own deputies to important positions. If the deputy belongs to the RP, she may expect to have about two more points than the average opposition deputy (Model 2, Table 6). Those two points are the equivalent of a regular member position in an average committee. Put another way, a typical RP member has 120 percent of the influence of an average opposition deputy in committees.
Second, we compare the impact of party dominance against that of apprenticeship and seniority norms. The Freshman variable's estimate is negatively and significantly related to the committee power held by deputies (Model 1, Table 6). In other words, freshmen are typically assigned to less relevant positions than seniors. To assess the relative importance of RPs compared with the apprenticeship norm, consider Table 7, which contains the scores predicted by our first model, for all combinations of the RP and Freshman variables (all other variables held at their means). As can be seen, a senior deputy from the opposition is expected to have no more influence in committees than a freshman from the RP (contrast the predicted scores of 10.7 against 11, respectively). In other words, although seniority can help to push a deputy up the ladder, it does not trump the benefit of belonging to the RP. As for our second model (Model 2, Table 6), we find a similar pattern for seniority in Chamber. Deputies are more likely to get high-profile positions in committees when they have more experience in the Chamber. Yet, the impact of an additional term is significantly lower than the impact of belonging to the RP (as confirmed by a Wald test: F = 26.99; p = 0.00). Most importantly, recall that very few deputies in Argentina are seniors. Indeed, the mode of the count of legislative terms over the whole time period is one, and only 5 percent of the deputies between 1946 and 2001 were in office for more than two terms. Hence, even though seniors benefited from their status, they represent a minority in the Chamber.
Predicted Power in Committees, by Party Affiliation and Freshmanship
Note: Predicted values are based on the estimates of Model 1, Table 6. All other explanatory variables are held at their means.
Source: See footnote 2 of this article.
At this point, it would be tempting to conclude that the hypothesis of RP dominance is supported by the data. However, observing the whole period (1946–2001) may disguise important differences across time. For example, under PR (that is, from 1963 on), RPs tended to have slighter margins of victory. Cartel-like behaviour may differ depending on margins of victory.
The margin hypothesis (Young and Heitshusen 2003) states that smaller shares of victories lead majorities (or in our case RPs) to secure even higher super-proportions in committees as compared to larger majorities (RPs). Small shares of victories became frequent in Argentina after the introduction of PR in 1963, yielding suitable data to test this hypothesis. Using dummies for five main RP types plus the opposition, we can test whether smaller RPs (pluralities) exhibit different behaviour than larger RPs (majorities).
In the first two columns of Table 8, we compare the coefficient estimates for five RP types against the reference of all opposition deputies. At first glance, there seems to be a pattern consistent with the margin hypothesis. The advantage of deputies from smaller RPs (regular and exceptional pluralities) is slightly higher than that of deputies from majorities, although the relationship is not clearly linear (the estimate for qualified majorities happens to be larger than for other majority types).
Power Allocation in Argentine Committees, by Margin of Victory, 1946–2001
Power Allocation in Argentine Committees, by Margin of Victory, 1946–2001
Note: OLS regression with the weighted sum of assignments of a deputy as the dependent variable. All control variables from the first model are included in this regression, although only the output for ruling party types is reported.
= p < 0.001;
= p < 0.01;
* = p <0.05.
Source: See footnote 2 of this article.
However, a stricter test would consist of assessing whether coefficient estimates for RP variables significantly differ from each other. We first performed ten tests, one for each possible pairing of RP type estimates, to confirm whether those coefficients are equal. In all cases, the results indicate that we cannot reject the null hypothesis of equal coefficients (results not shown). To illustrate this, we ran the regression again, the results of which appear in the second column of Table 8, this time using qualified majorities as the reference category. It can be seen that other RP types are not significantly different from qualified majorities. Therefore, the margin hypothesis is not supported by the data: RPs tend to exert an equally over-proportional control on important committee assignments, despite variations in the electoral system and in the magnitude of their electoral victories.
Overall, electoral reforms, particularly the introduction of the PR system in 1963, did not affect committee assignment patterns. Certainly, PR has spawned different types of RPs compared to majority systems (see Table 2), but they exhibit similar super-proportions in important committee positions over the whole time period.
While these initial results offer empirical evidence in favour of the cartel hypothesis, until now we have ignored the impact of committee-specific seniority. Since prior experience within committees can hardly be indexed for a single deputy (i.e. committee-specific experience relates to advancement within that specific committee, but is not relevant for assignments to other committees), we shift the focus to assignments as units of analysis. We consider the pool of assignments that took place during the time period. Then we are able to match each assignment to a committee with the deputy's previous experience in that specific committee. We consider a dichotomous dependent variable, coded one for chair assignments, and zero for all other assignments. Committee-specific experience and the party status of the deputy can then be considered as explanatory factors of chairmanship.
In addition to the inclusion of a measure of committee-specific seniority, we modify our previous set of explanatory variables as follows. First, we introduce an interaction term between party status (the RP variable) and committee-specific seniority. This interaction serves the useful purpose of examining whether committee-specific seniority is a selection criterion used within the RP, or if it overrides RP's ascendency. Second, the Chamber Position variable leads to perfect failure predictions: no deputy holding an authority position in the Chamber ever became chair of a committee. Therefore, those 69 cases were dropped. Finally, we rescaled all explanatory variables between zero and one, so that coefficient estimates in our latent equations may be compared in size even if we use logit regressions.
Table 9 reports the results. The second model (reported on the right) includes the interaction term between RP and committee-specific seniority, whereas the first model (on the left) does not. When focusing on the model without an interaction term, committee-specific seniority appears to play a huge role in explaining the likelihood of being assigned chair, if we consider that the size of the estimate for this variable is much larger than the size of the estimate for RP (again, because all explanatory variables have the same scale, estimates can be compared in size). The portrait becomes clearer when considering the interaction term in model two. The constitutive term for the committee-specific seniority variable shows the impact of committee experience for deputies of opposition; for RP members, both the constitutive term and the interaction term must be added. Consequently, committee experience does matter in Argentina, but mostly within the RP.
Determinants of Chairmanship in Argentina (Logit Regression)
Determinants of Chairmanship in Argentina (Logit Regression)
Note: Binary logit regression with committee chairmanship as the dependent variable. Legislature dummies are included in each regression, but the output is not shown. Also, the first legislature (1946–48) had to be dropped since we have no data on previous experience for this legislature.
= p<0.001;
** = p<0.01;
= p<0.05.
Source: See footnote 2 of this article.
To illustrate our last contention, Figure 1 plots out-of-sample predicted probabilities of being assigned chair against committee-specific seniority, for opposition deputies (left panel) and for deputies of the RP (right panel). Those probabilities are computed using the estimates of model two in Table 9. As can be seen, increasing committee experience when a deputy belongs to the opposition has a trivial impact on the probability of becoming chair: the predicted probability varies from 0.02 to 0.13, for a change from the minimum level of experience to the maximum level in the sample. In contrast, the predicted probability of being assigned chair increases from 0.04 to 0.93 when the deputy belongs to the RP, for a similar change in the level of committee experience. Thus, having committee-specific experience is not a sufficient condition for a deputy aspiring to become chair; the ruling status of her party definitely matters most if this aspiration is to be fulfilled.
Predicted Probability of Chairmanship, by Committee-Specific Seniority and Party Status
This paper examined the previously unexplored subject of committee appointments in the institutionally unstable Argentine Chamber of Deputies from 1946 to 2001. In the context of Argentine politics, the institutional characteristics of the Chamber and the development of its committee system were reviewed. An original data set was assembled and a new methodology was developed to examine this legislature's organization and to test the applicability of theories created to study the U.S. Congress. Evidence emerged that RPs determined the key processes of the Argentine Chamber of Deputies’ committee system, across time, and that committee assignments were not independent from party influence.
The qualitative exploration showed, in particular, that the proportionality rule in force to allocate committee seats was circumvented through strategic interpretation of the rules by the President of the Chamber. The over-representation of RPs in key committee positions confirms the power of Presidents over appointments. Our empirical evidence shows that party dominance in committees transcends RP types, and different margins of victory. Irrespective of variations, when accounting for the importance of each assignment, RPs were able to secure over-representation in the committee system during the entire period. Although we found evidence of a seniority norm in the appointment process, such a norm remains trivial in the face of the simple fact that very few deputies in Argentina are senior. Seniority does not override the importance of being a member of a RP in order to obtain strategic committee positions. Committee-specific experience is, at most, a criterion used within RPs to select the chairs from among their members.
Our findings suggest that, despite differences between the U.S. House and the Argentine Chamber, political parties face similar challenges and strategies, seizing key structural powers in order to overcome collective action problems. In addition to the benefits for policy-making, the control of the committee system is a powerful tool to ensure party cohesion as RP leaders are then able to maintain the support of their deputies through the incentive of prestigious committee assignments. This strategy may even be more important in Argentina, where RPs had to deal with an ever freshman Chamber, shaken several times by institutional breakdowns.
The empirical evidence presented in this paper confirms the portability of theories that were developed for the study of the U.S. House. The demonstration that specific institutional paths and regime histories are not obstacles to the application of the Cartel Theory contributes to the development of a general theory of legislative organization. Finally, the methodology and approach developed for this paper offer a new way for analyzing committee systems, one that includes measures for the importance of committee assignments, something lacking from previous institutional studies.
Footnotes
Appendix A
The ranking of committees is based on three indicators. Variations in those three variables stem directly from the reforms to the Reglamento approved by deputies. Therefore, each of the three measures should reflect the importance of committees as perceived by the deputies themselves. First, the timing of the increase in size of a committee captures the demand for being member of that committee. Although all committees tend to grow in size through time, we could easily distinguish between six different patterns related to the timing of such increases. Committees that first increased in size in the time-period (holding constant proportional increases in the size of all committees over time) received the highest score (six) on a [1,6] discrete scale. Committees having never benefited from a relative increase in their size received the lowest score (one).
Second, the timing of the increase in the count of Vice-chairs and Secretaries within a committee is computed in a similar way. Again, we were able to distinguish six different patterns (since those changes occurred after the same reforms to the Reglamento). Although correlated with the first indicator, there is some variation captured by this second measure: very important committees (such as Budget) benefited from earlier increases in the number of high-rank positions. Again, this second indicator varies from one to six, with six indicating the highest score. Third, the longevity of a committee is also considered as a measure of its importance. The rationale is simple: Committees which have persisted over time are more important than newly created committees or than committees that have been dismantled. For this indicator, we simply measured the count of legislatures for each committee. This indicator has been rescaled from one to six, in order to match the two others. The importance of a committee is the average score on those three indicators, and is therefore scaled from one to six.
The construction of our index relies on an interval scale measure of committee importance. We have tested a number of alternative functional forms, and the empirical results presented in this paper are not substantially affected. Specifically, we considered binary measures in which the most important committees (those receiving a score above 4 in Table A1, or alternatively those receiving a score above 3.33) are coded one, while other committees are coded zero. Replicating the results with those alternative measures, we found that all our key estimates have consistent signs, and remain statistically significant. Therefore, we restrict our attention to the results based on the measure of committee importance described above.