Parliamentary rules,party norms,and legislative speech

Abstract

This article examines speech participation under different parliamentary rules: open forums dedicated to bill debates, and closed forums reserved for non-lawmaking speeches. It discusses how electoral incentives influence speechmaking by promoting divergent party norms within those forums. Our empirical analysis focuses on the Chilean Chamber of Deputies. The findings lend support to the view that, in forums dedicated to non-lawmaking speeches, participation is greater among more institutionally disadvantaged members (backbenchers, women, and members from more distant districts), while in those that are dedicated to lawmaking debates, participation is greater among more senior members and members of the opposition.

Keywords

Legislative speech electoral incentives legislative behavior party norms

Introduction

On August 4, 2009, Carolina Goic, a legislator from the Christian Democratic Party, stood up on the floor of the Chilean Chamber of Deputies and delivered a speech calling attention to the environmental damage caused by salmon farming and aquaculture in the Magallanes region, where her electoral district was located. She asked her fellow legislators to take action to protect the biodiversity of the region and enact appropriate regulations. She was immediately followed by another legislator, Jorge Sabag, who demanded that the Ministry of Public Works upgrade a rundown bridge in the town of San Carlos, within his district, so its inhabitants, partially isolated due to the construction of a new highway, could make their way into and out of the town. These speeches, as well as many others that are regularly delivered at the end of sessions, called attention to local problems and showed deputies standing up for their constituents. Of course, these are not the only types of speeches legislators deliver in the chamber; earlier in the day, deputies debated a major bill focused on cracking down on money laundering.

Bill debates and individual speeches are essential tools used by legislators to exercise their roles as representatives. Speechmaking, however, has not been scrutinized as often as other legislative activities, such as roll call votes, bill initiation, and cosponsoring, particularly in presidential countries. Examining legislative speech participation can illuminate not only unexplored patterns of representation but also more specific behaviors derived from electoral incentives and legislators’ demands for position-taking.

In this article, we examine congressional speeches delivered on the floor of the Chilean Chamber of Deputies. Our contribution to the literature on legislative speechmaking is two-fold. First, we shed light on legislative speeches during different stages of the chamber’s daily session, some that give leaders control over who may talk and others during which no formal screening is possible. Second, we examine differences between bill debates and non-lawmaking speeches, which are usually examined separately. We find that institutional rules and endogenously determined party norms result in substantial differences in speech participation. During party-controlled settings dedicated to non-lawmaking speeches, participation is greater among more institutionally disadvantaged members, while participation in lawmaking debates occurs more frequently among more senior legislators and members of the opposition.

The remainder of this article is organized into four parts and a brief conclusion. The next section discusses the differences between open and closed speech rules. Then we examine legislators’ electoral incentives and institutional rules and derive a series of expectations about legislative speeches in the Chilean Chamber of Deputies. This is followed by a description of the data we collected, the variables used in the empirical analysis, and the statistical models employed. We then outline the result of our analysis, followed by a brief conclusion.

Speech participation

Legislators use speeches to convey messages to constituents, interest groups, other politicians, and the general public. They are tools that representatives can employ to advocate for particular policies, advertise their positions, signal their efforts, and differentiate themselves from their opponents (Kellerman, 2012; Maltzman and Sigelman, 1996; Martin and Vanberg, 2008). Both individual legislators and party leaders value the opportunities offered by forums dedicated to debating bills and delivering speeches. The media often reports on the various oral statements made by legislators on the parliamentary floor, and debates are sometimes televised. However, participation in these forums is not equal. As is often noted, in legislatures, time is a scarce resource, which results in norms and regulations that constrain speaking time.

Internal rules of procedure may give party leaders the right to control access to the parliamentary floor. This is the case in the German Bundestag, for example, in which party leaders exert considerable control, screening members and often denying speaking time to party rebels. In other contexts, internal rules of procedure create specific forums wherein party leaders lack the formal authority to block legislators from participating in debates. Proksch and Slapin (2012) explained these differences by theorizing that electoral incentives are paramount in determining the organization of legislative speech; decisions regarding who may speak reflect parties’ strategic calculations over the value of presenting a cohesive party message. In scenarios where electoral incentives emphasize the role of parties over individuals, parties strive to control legislative speech. But in settings where the individual reputations of candidates are particularly salient, parties impose fewer constraints on legislators’ speech participation (Proksch and Slapin, 2012).

In some cases, however, the organization of legislative speech does not align quite neatly with electoral incentives. Giannetti and Pedrazzani (2016), for example, noted that, in the Italian Chamber of Deputies, parliamentary rules establish both types of forums: restricted access to floor debate when addressing ordinary bills and unrestricted access when addressing law decrees. They argued that speech participation primarily depends on these formal rules. Their findings revealed no substantial differences between frontbenchers and backbenchers in the open forum. In the restricted forum, however, they found more frequent participation by government frontbenchers and less frequent participation by opposition rebels. In their view, these findings reflect the limits to leadership control in the absence of rules that restrict access to the plenary.

The combination of open and restrictive speech rules is also present in presidential assemblies, such as the US House of Representatives. There, bill deliberations may take place under restrictive or open debate rules, a decision that is usually at the discretion of agenda setters (e.g. Rules Committee). In practice, restrictive rules prevail. In addition, non-lawmaking speeches in the US House of Representatives, such as the so-called one-minute speeches, occur in an open forum that allows any legislator to participate.

Chile, which is the focus of our study, displays noteworthy differences from the previous examples. As in Italy and the US, the internal rules of the Chilean Chamber of Deputies allow for both restricted forums in which participation is controlled by parties and open forums that provide all legislators with the right to participate. However, conversely to the norm in these two chambers, the Chilean legislature establishes an open forum for bill debates and a restricted forum for non-lawmaking speeches. What should we expect in this case?

The implications of previous studies are at odds regarding this scenario. On one hand, Proksch and Slapin’s (2012) theory maintains that electoral incentives will prevail even when party leaders lack the formal authority to screen participants, while Giannetti and Pedrazzani’s (2016) perspective casts doubt on the ability of party leaders to influence speech participation in the presence of open rules. Moreover, the evidence from the US House of Representatives regarding speeches in non-lawmaking open forums is either inconsistent with the previously mentioned expectations―for example, party leaders appear to participate more often―or inconclusive―for example, regarding participation by new members and committee leaders (Maltzman and Sigelman, 1996; Rocca, 2007).

In the next section, we discuss legislators’ electoral incentives and the institutional structure regulating speechmaking in the Chilean Chamber of Deputies and derive a series of expectations about participation in both open and closed speech forums.

Congressional speechmaking in Chile

Chile is a presidential country with a congress composed of career-oriented legislators. Reelection rates are among the highest in Latin America; close to two-thirds of members return to their seats after each election. Comparative studies typically rank Chile’s Congress as the most institutionalized in Latin America, possessing high technical expertise, a professional membership, and a robust committee system (Palanza et al., 2016). Chilean parties are among the most ideologically consistent in the region (Rosas, 2005) and exhibit high levels of unity in roll call votes (Toro Maureira, 2007).

During the period analyzed in this article (2006–2018), the electoral rule in place was open-list proportional representation in two-member districts. Voters selected one candidate from a two-candidate list. Votes for candidates in the same list were pooled, and seats were allocated to lists according to the vote total in the district. If a list won only one of the two seats, it would go to the candidate receiving the most votes. Typically, the two largest coalitions would both win one seat. A consequence of this dynamic was to exacerbate intra-list competition, which usually occurred between candidates from different parties of the same coalition.

After returning to democracy in 1990, Chile had two legislative coalitions that captured most of the congressional seats. The center-left coalition, called the Concertación until 2013, comprised the Socialist Party (PS), the Party for Democracy (PPD), the Christian Democrats (DC), and the smaller Radical Social-Democratic Party (PRSD). In 2013, its name was changed to New Majority, and the Communist Party (PCCh) was incorporated into its ranks. The center-right coalition, primarily comprising the Independent Democratic Union (UDI) and the National Renewal Party (RN), has had a variety of names, but is typically referred to as Alianza.

As elections approached, coalitions not only had to choose a presidential candidate but also craft congressional lists of just two candidates per district, knowing that only one would likely be elected. This task was particularly complicated for the Concertación coalition, which was composed of multiple parties. While there are variations across parties in the extent of local input allowed, the candidate selection process was relatively centralized at the leadership level (Siavelis, 2002).¹

Several scholars have stressed that Chile’s two-member electoral system fostered personal vote-seeking incentives (Angell, 2003; Morgenstern et al., 2014; Navia, 2008). The centralized process of candidate selection allowed party leaders to moderate intra-coalition struggle to some degree, but the need to cultivate a personal vote remained vital. During electoral campaigns, candidates would go to great lengths to highlight their own attributes. Following their detailed analyses of 24 congressional electoral campaigns, Díaz Rioseco et al. (2006: 169) concluded that the most effective way for challengers to beat an incumbent was ‘to spend more than, and aggressively confront, their own coalition’s incumbent candidate.’ While party identification helped mobilized base voters, successful candidates had to establish a personal following (Angell, 2003). Public opinion polls have revealed that only a small share of the Chilean electorate characterizes their vote as partisan, rather than being based on candidates’ attributes (Luna, 2008).

In short, Chile’s electoral rules, like those of the UK and the US, create incentives to seek a personal vote. In this context, Proksch and Slapin (2015: 80) expected the institutional structure regulating speech participation to reflect the basic incentive structure derived from electoral rules: a setting that ensures that all legislators can take the floor to deliver a speech without party approval. The rules regulating lawmaking debates in Chile follow such an open forum, guaranteeing all legislators the right to speak.

However, the setting in which constituency-focused speeches are more likely to be delivered, one in which the linkages with district issues are patently visible, is a closed forum wherein parties formally control speech participation. Thus, the formal organization of debate participation in Chile’s Chamber of Deputies does not necessarily coincide with electoral incentives. How do parties manage this closed forum dedicated to non-lawmaking speeches? One could interpret the insights derived from Giannetti and Pedrazzani’s (2016) analysis of the Italian case as suggesting that leaders would take full advantage of their formal control to privilege certain kinds of legislators (e.g. frontbenchers). But unlike Italy, Chile’s closed forum is dedicated to non-lawmaking speeches, so leaders’ calculations likely differ.

Even if parties are privileged in terms of the formal allocation of speaking time, when the incentives to cultivate a personal vote are strong, so too are internal pressures to develop norms that are electorally valuable to the party. We believe this is the case in Chile.² Preliminary and anecdotal evidence is consistent with the view that, during the stage of the session dedicated to non-lawmaking speeches, party leaders do not restrict access to the floor, nor do they control what deputies say (Alemán et al., 2017; Dockendorff, 2018; Gamboa and Toro, 2018).

The Hour of Incidents (Hora de Incidentes) is the forum dedicated to short non-lawmaking speeches.³ It provides a crucial arena for legislators to stand up for their districts’ concerns and strengthen their connections to their constituents (Gamboa and Toro, 2018). Speeches delivered at that time address a variety of issues and may include information requests to the executive branch (oficios). Time is allotted to party groups based on their share of seats, and then parties distribute slots to their respective members.

The Hour of Incidents provides legislators with a valuable opportunity to showcase their district-oriented activities via speeches that address regional or local matters (Alemán et al., 2017; Gamboa and Toro, 2018).⁴ It is not uncommon for legislators to request that transcripts of their speeches be sent to relevant actors in their constituency, such as interest groups or social organizations. National and local media frequently report on the content of these speeches. The TV channel of the chamber, which is carried by cable TV providers, regularly broadcasts the Hour of Incidents. Many deputies post the speeches they deliver in this forum on their Facebook pages, websites, and Twitter feeds. Alemán et al. (2017) showed that more frequent participation in the Hour of Incidents is associated with a greater chance of being reelected.

Lawmaking debates occur mostly during the Order of the Day. Rules give every deputy the right to participate in this forum. Each member is allowed to speak twice during the general discussion of a bill (up to 10 and 5 minutes respectively) and twice per article during the particular (i.e. detailed) discussion of a bill (each time up to 5 minutes).⁵

Bill debates, in comparison to non-lawmaking speeches, offer more limited opportunities for position taking on issues important to constituents. There are several reasons for this. Constitutional rules limit the ability of legislators to offer tax and spend amendments, thereby blocking opportunities to use the floor to push for pork-barrel projects targeted to specific districts (Baldez and Carey, 1999). Moreover, legislators are prohibited from addressing non-germane issues during bill debates. This rule, enforced by the president of the chamber and in place for over a century, means that legislators may only speak on whatever topic the chamber is discussing at that moment and cannot raise other issues. Furthermore, agenda setters,⁶ who control the types of bills that come to the floor, typically avoid bringing up controversial issues that would divide their coalition (Alemán, 2006). This limits opportunities for individual position-taking that could harm the party label, antagonize its leadership, and complicate coalition coordination during bill debates.

In short, institutional differences between forums dedicated to bill debates and those dedicated to non-lawmaking speeches, we believe, should result in significant differences in participation. We expect members of congress who are comparatively disadvantaged to be more likely than others to speak in non-lawmaking forums.⁷ Although the rules governing Chile’s Chamber of Deputies give parties control over such forums, internal norms of open participation should be conducive to party leaders playing a non-discriminatory coordinating role, accommodating members’ needs for constituency-oriented position-taking. Party leaders should facilitate and even encourage this behavior because it is electorally beneficial to the party (Gamboa and Toro, 2018). Voter-focused non-lawmaking speeches give parties an opportunity to enhance the reputation of their members while minimizing the risk of open challenges to their core policy positions. In other words, it is an optimal strategy for leaders concerned about the party’s electoral fortune. As a result, during this forum, individuals seeking to strengthen their reputations with voters should have an opportunity to pursue that specific goal.

We expect four types of legislators to participate more often in the deliverance of non-lawmaking speeches: backbenchers, junior members, women, and those coming from districts farther away from the seat of government. Rank and file legislators typically encounter greater obstacles to communicating their positions than do chamber authorities and party leaders, who enjoy several procedural privileges (Rocca, 2007). They are also less likely to be influential in the formation of policy and the process of amending government legislation. Furthermore, on average, backbenchers are not as likely to get media coverage on policymaking issues of national importance. Consequently, they are particularly motivated to take advantage of forums dedicated to non-lawmaking speeches, which offer an abundance of opportunities for position-taking. Party leaders, who may worry that rank and file members may deviate from the party position in bill debates, are much less concerned about this potential behavior during non-lawmaking speeches, which do not touch on the core policy positions of the party. Instead, they are likely to find the exposure of disadvantaged members particularly beneficial to the electoral fortunes of the party.

This difference in speech participation between rank-and-file and frontbenchers in non-lawmaking forums is less likely to be manifested in small parties. This is not only the result of a ‘small n’ generating wider confidence intervals. When parties are small, the share of members occupying a party leadership position is greater; their influence is comparatively limited, and, as a result, the expected differences between frontbenchers and the rank-and-file are likely to be attenuated.

Junior members are also likely to find non-lawmaking speeches particularly effective when attempting to enhance their reputations with district constituents (Dockendorff, 2018). In comparison to more senior members, they tend to have smaller networks, play a less prominent role in committees, and are not as well-known among voters (Rocca, 2007). Non-lawmaking speeches provide junior legislators an especially apt setting for constituency-oriented position-taking on the topic of their choosing.

We expect legislators elected from districts more distant from the seat of government to also find non-lawmaking speeches particularly useful to overcome some of the challenges they face. Citizens from more remote regions are more likely to rely on the state for their health care and children’s education (UNDP, 2018). Yet, they usually confront greater challenges in accessing the government bureaucracy, and most interest groups lobbying members of congress are based in the metropolitan region.⁸ Matters of importance to their districts are not often highlighted by the national media and are less likely to be discussed during bill debates. These districts tend to be more rural and more sparsely populated. More than 60% of high-income households are located in Santiago’s metropolitan regions, with the percentage of high-income households decreasing as one moves away from the center of the country (UNDP, 2018).

Legislative studies focused on other countries have revealed that representatives from more distant districts are more likely to engage in constituent-focused behavior than those from the center (André and Depauw, 2014; Brouard et al., 2013; Heitshusen et al., 2005; Martin, 2011). Furthermore, Alemán et al. (2018) found that, in Chile, deputies from more remote regions are more likely to present written parliamentary questions related to their district to the executive branch than deputies from regions closer to the metropolitan area.

Women have also encountered an adverse context in Chile’s Congress. During the period examined in this study, only about 15% of legislators were women. As in many other countries, Chilean women have also been underrepresented in leadership positions within the chamber, and appear to have faced greater challenges accessing campaign funds. This unfavorable context has led some scholars to expect women to deliver fewer speeches than their male colleagues (Bäck et al., 2014). However, in the case of non-lawmaking speeches, our expectation differs. Along the lines of Pearson and Dancey’s (2011) argument about the US House of Representatives, we expect the adverse context previously described to motivate female legislators to be more active, visible, and incentivized to prove themselves to their constituents.

These propositions can be summarized into four testable hypotheses. We expect:

H1: Rank-and-file legislators to deliver more non-lawmaking speeches than frontbenchers.

H2: Junior members to deliver more non-lawmaking speeches than more senior members.

H3: Legislators from more remote districts to deliver more non-lawmaking speeches than legislators from districts closer to the seat of government.

H4: Female legislators to deliver more non-lawmaking speeches than male legislators.

During bill debates, however, we expect a different dynamic. While the rules of Chile’s Chamber of Deputies give all members the right to participate in bill debates, participation is unequal. As in most other legislative bodies, the policy positions of parties are typically decided before the actual debate takes place. While parties benefit from presenting a cohesive stance that enhances the value of the party label, incentives to cultivate a personal vote may threaten a unified party message. Institutional constraints, however, can ameliorate this potential collective action problem (Proksch and Slapin, 2015: 25). We have noted the implications of germaneness rules in bill debates, rules that preclude legislative particularism, and agenda control. Given these restrictions, legislators should find lawmaking debates less suitable for district-oriented position-taking behavior than the non-lawmaking forums previously discussed.

Lawmaking debates are more likely to address national policy and sectoral issues than local matters. Parties benefit from coordinating their public stances on these matters and developing norms that privilege expertise in lawmaking debates (Alemán et al., 2017; Proksch and Slapin, 2012). Even if, formally, leaders are unable to prevent those members who wish to take the floor from participating in the debate, they have other tools (e.g. nominations and influence over bills) to discourage challenges. Further, marginalized members have fewer incentives to defy party norms regarding participation in this forum, given the limited opportunities offered for district-oriented position-taking.

More specifically, we expect more senior members and those who have held committee chairs to speak more often than others during lawmaking debates. In addition, we expect that, holding all else equal, being in the opposition should increase a legislator’s likelihood to participate in bill debates.

In Chile’s presidential system, members of the opposition face substantial hurdles when attempting to influence policymaking. The executive originates most major bills, has exclusive rights to initiate amendments on various topics, and maintains a firm grip on budgetary policy (Baldez and Carey, 1999). Lawmaking debates give members of the opposition a visible platform to convey their policy stances to constituents (Martin and Vanberg, 2008). They can criticize government bills and publicize their alternative proposals, even if, in practice, they have a slim chance of being enacted. In contrast, governing parties tend to place a premium on quick passage (Pedrazzani, 2017). Thus, we have three additional hypotheses to be tested in our empirical analysis. We expect:

H5: Legislators with higher tenure to give more speeches than more junior members during forums dedicated to lawmaking debates.

H6: Committee chairs to give more speeches than others during forums dedicated to lawmaking debates.

H7: Legislators in the opposition to give more speeches than legislators in the government coalition during forums dedicated to lawmaking debates.

To summarize, our expectations about speech participation vary according to the stage of the chamber’s daily session. In forums where parties control participation but are dedicated to non-lawmaking speeches, we expect relatively disadvantaged members to participate more often and leaders to not stand in the way of their doing so. In forums dedicated to lawmaking debates (where each legislator has the formal right to speak), we expect party norms to favor more senior members and committee chairs. In these instances, we also expect greater participation from members of the opposition. The next section discusses the data we collected and the models employed to test these propositions.

Data, variables, and model

To analyze speech participation, we collected data from three legislative periods covering 12 years beginning in 2006. During this period, there were 44,239 individual speeches. The information made available by the chamber excludes short procedural interventions by chamber authorities (e.g. to assign speaking time, petition for a roll-call, or answer a point of order) and legislators (e.g. questions or petitions to the chamber’s president), and identifies the name of the speaker, and the date and part of the session where the speech was delivered.⁹ To analyze the data, we have aggregated speeches by deputy for each period and stage of the session.

We grouped speeches in two categories. The first refers to individual participation in lawmaking debates and includes speeches delivered during the Order of the Day, the main arena for bill debates, and the Simple Dispatch stage, dedicated to the quick passage of non-controversial bills. It also includes speeches delivered during Special Sessions held to continue a bill debate started in an ordinary session.

The second category includes speeches delivered during the Hour of Incidents, the Agreements & Resolutions stage, and other Special Sessions. Time in the Hour of Incidents and Special Sessions¹⁰ is distributed to parties according to their seat share. Special Sessions are usually called as a result of an agreement or resolution issued by the chamber to address a matter considered of great importance.¹¹ Agreements are directed at the executive and tend to relate to governmental acts, whereas resolutions focus on other salient national or international issues. During the Agreements & Resolutions stage, time is allotted to a maximum of four deputies, all from different parties: two from parties supporting the agreement or resolution and two from those that oppose the agreement or resolution.

The dependent variable in all our models is a count reflecting the number of speeches given by each legislator. During the time covered in our analysis, Chile’s Chamber of Deputies was composed of 120 members. We run separate negative binomial regressions for the two categories of speeches,¹² which allows us to investigate differences in participation within each of the two categories of speeches. In addition to running each model for the full sample, we run separate estimations for the last period of the legislators’ term, which starts 10 months before the election, and prior years. Bäck et al. (2019), in their examination of six European parliaments, found that speeches were less uniformly distributed among members of parliament when elections were approaching. By splitting our data, we seek to examine whether the findings from the full sample change at the end of the term, as Bäck et al. (2019) suggested. Lastly, we investigate whether the effect of a particular variable varies across similar models with a different dependent variable (i.e. lawmaking versus non-lawmaking speeches).

The models have six key independent variables. Authority indicates whether the deputy was a member of the chamber’s directorate (mesa) or the leader of a party bloc (jefe de bloque). Committee chair indicates whether the deputy held that position at some point during the legislative period.¹³ Seniority captures a legislator’s tenure in office. Female indicates whether the deputy is a woman. Government refers to deputies who belong to the government coalition (both major coalitions alternated in the presidency during the time analyzed in this article). Distance refers to the log kilometers between the deputy’s district and the official seat of government, Chile’s Presidential Palace (Palacio de la Moneda), in the city of Santiago.

In addition to time dummies for each period examined, we include the following control variables that have appeared in previous works. Party size measures the total number of seats for the deputy’s party. Because differences between backbenchers and frontbenchers are not likely to manifest in very small parties, we interact the authorities variable with this one measuring the size of parties expecting more substantial effects as party size grows. Margin is the difference between the deputy’s share of the list vote and the share of the list vote received by his or her list partner. It is a measure of electoral vulnerability. Retire tells us that the deputy did not run for legislative office at the end of his or her term. Retiring members should be less interested in position-taking. We also control for coalition effects. Alianza refers to deputies belonging to the center-right coalition, and Other Party is indicative of independent deputies or those not aligned with any of the main coalitions. The baseline used for comparing coalition effects is an omitted variable indicating membership in the center-left coalition (Concertación later renamed Nueva Mayoría). A table of summary statistics appears in the appendix in the supplementary material online. The next section discusses our findings.

Results

The results for non-lawmaking speeches appear in Table 1, while those for lawmaking speeches appear in Table 2 (coefficients appear in bold, and standard errors are shown below). The first two models are for the full sample; the third excludes data from the last period of the four-year term; and the last column focuses only on the last period.

Table 1.

Count models of the number of non-lawmaking speeches.

	Full sample		First sessions	Last session
	Model 1	Model 2	Model 3	Model 4
Authorities	–0.154*	0.208	0.165	0.454*
Authorities	0.092	0.195	0.195	0.250
Authorities × party size		–0.018**	–0.016*	–0.031***
Authorities × party size		0.009	0.009	0.011
Party size	0.011*	0.016**	0.016**	0.018**
Party size	0.006	0.006	0.006	0.008
Committee chair	0.008	–0.008	0.008	–0.090
Committee chair	0.090	0.090	0.090	0.116
Female	0.307***	0.279**	0.249**	0.396***
Female	0.116	0.115	0.115	0.144
Seniority	–0.026	–0.020	–0.027	0.011
Seniority	0.035	0.035	0.035	0.044
Distance	0.115***	0.109***	0.110***	0.104***
Distance	0.021	0.021	0.021	0.027
Government	–0.232**	–0.245**	–0.235**	–0.291**
Government	0.097	0.097	0.097	0.124
Alianza	–0.659***	–0.684***	–0.667***	–0.747***
Alianza	0.125	0.124	0.124	0.160
Other party	0.247	0.227	0.238	0.142
Other party	0.252	0.250	0.250	0.306
Margin	0.063	0.023	0.012	0.022
Margin	0.289	0.287	0.289	0.353
Retire	–0.353***	–0.354***	–0.279**	–0.747***
Retire	0.120	0.120	0.120	0.163
Period 2010–2014	–0.151	–0.165*	–0.141	–0.282**
Period 2010–2014	0.095	0.095	0.095	0.120
Period 2014–2018	–0.062	–0.066	–0.076	–0.014
Period 2014–2018	0.096	0.095	0.096	0.119
Constant	3.340***	3.278***	3.067***	1.638***
Constant	0.191	0.192	0.193	0.241
Observations	360	360	360	360
Log likelihood	−1610.8	−1608.6	−1537.3	−1046.1

p < .1, **p < .05, ***p < .01.

Table 2.

Count models of the number of lawmaking speeches.

	Full sample		First years	Last year
	Model 5	Model 6	Model 7	Model 8
Authorities	−0.045	0.117	0.109	0.136
Authorities	0.080	0.180	0.180	0.213
Authorities × party size		−0.008	−0.008	−0.010
Authorities × party size		0.008	0.008	0.010
Party size	0.008	0.010*	0.008	0.016**
Party size	0.005	0.005	0.006	0.007
Committee chair	0.127	0.124	0.102	0.198**
Committee chair	0.081	0.081	0.082	0.096
Female	−0.085	−0.095	−0.139	0.066
Female	0.101	0.101	0.102	0.120
Seniority	0.068**	0.070**	0.071**	0.074**
Seniority	0.031	0.031	0.031	0.037
Distance	0.016	0.016	0.015	0.022
Distance	0.018	0.018	0.018	0.022
Government	−0.209**	−0.215**	−0.177**	−0.361***
Government	0.088	0.088	0.088	0.103
Alianza	−0.557***	−0.563***	−0.513***	−0.739***
Alianza	0.106	0.105	0.106	0.125
Other party	0.056	0.036	0.090	–0.177
Other party	0.226	0.226	0.227	0.269
Margin	0.158	0.144	0.190	−0.088
Margin	0.257	0.257	0.255	0.314
Retire	−0.158	−0.157	−0.093	−0.393***
Retire	0.106	0.105	0.106	0.126
Period 2010–2014	0.208**	0.202**	0.299***	−0.104
Period 2010–2014	0.084	0.084	0.084	0.099
Period 2014–2018	0.372***	0.367***	0.454***	0.102
Period 2014–2018	0.087	0.087	0.087	0.102
Constant	3.973***	3.939***	3.621***	2.634***
Constant	0.169	0.171	0.172	0.204
Observations	360	360	360	360
Log likelihood	−1830.4	−1829.9	−1738.1	−1334.8

p < .1, **p < .05, ***p < .01.

The results for non-lawmaking speeches conform to our expectations. The coefficients for female legislator (H4) and distance (H3) are statistically significant in all four models. According to model 2, women legislators give 48.5 speeches, while their male colleagues give 36.7. The effect of distance on the number of non-lawmaking speeches appears in Figure 1. For example, a legislator coming from a district located 600 km from the seat of government in Santiago delivers approximately 44 speeches, which is close to 16 more speeches than a legislator from a district 10 km away.¹⁴

Figure 1.

Distance and number of non-lawmaking speeches.

Results also show that backbenchers are more active than frontbenchers (H1), but the magnitude of this effect is conditional on party size (model 2). Figure 2 illustrates a backbencher’s marginal effects and 95% confidence intervals across different values of party size.¹⁵ For instance, the predicted number of speeches for a backbencher when a party has 30 legislators is about 47 speeches, close to 14 more speeches than a frontbencher in a similar-sized party. As the figure shows, the effect goes up as party size increases but is not statistically different from zero for parties with fewer than 20 legislators.

Figure 2.

Marginal effects for rank-and-file given party size.

The coefficient for seniority is not statistically significant in any model examining non-lawmaking speeches, which is inconsistent with our expectation (H2); this finding does not change if we replace the tenure variable with a dummy for first-time members. Further, the control variables reveal that members of the center-right Alianza give fewer non-lawmaking speeches than do members from the center-left Concertación, and that individuals who will not be running for office at the end of their term give fewer non-lawmaking speeches than those who run again.

The results of the models examining lawmaking speeches appear in Table 2. Consistent with our expectations, more senior legislators (H5) participate more frequently than junior members in these forums. For example, the model predicts that a first-term legislator will deliver just over 67 speeches, while a fourth-term legislator will deliver close to 83 speeches.¹⁶ If we replace the tenure variable with a dummy for senior members (four terms or more), the results show them giving close to 14 more speeches than members with less tenure.

We expected that becoming a committee chair would increase the number of lawmaking speeches delivered (H6). The results show that the coefficient for committee chair is always positive; however, it is only statistically significant at the end of the four-year term (model 8). The results also show that opposition membership increases debate participation (H7).¹⁷ For example, all else being equal, an opposition legislator gives approximately 83 speeches, which is close to 16 more than the number of predicted speeches for a government legislator.

The control variables reveal that members of the center-right Alianza give fewer lawmaking speeches than members from the center-left Concertación (as was the case with non-lawmaking speeches). Interestingly, retiring legislators are as active as other legislators for most of their term; they only become less likely to deliver a speech in the last part of their term, as the elections approach.¹⁸

Findings from lawmaking debates also show no significant difference in the levels of participation between female and male legislators and between members elected from districts closer or farther away from the seat of government. And, as in previous models, the coefficient for the electoral margin lacks statistical significance.

To review, the analysis reveals substantial support for our expectations. As hypothesized, in the forum dedicated to lawmaking speeches, members of the opposition and more senior members speak more frequently than do junior members and government legislators. Committee chairs give more speeches than others, but only at the end of their terms. Findings also lend support to the view that participation in non-lawmaking speeches is greater among more institutionally disadvantaged members (female legislators, members from districts farther from the seat of government, and backbenchers).

To further evaluate the effect of our key variables in these different models, we proceed to test cross-model differences directly. As Gelman and Stern (2006) noted, a variable may be significant in one model and not in another, but this does not mean that the difference between them is itself significant. For example, we have shown that women deliver more non-lawmaking speeches than men and that there is no statistically significant difference between men and women in terms of lawmaking speeches. However, we have not compared the effect of being a female legislator in the first category (non-lawmaking speeches), with the effect of being a member of the same group in the second category (lawmaking speeches). To evaluate the equality of effects from multiple models, it is often necessary to fit them simultaneously to compute cross-model covariances. Running a simultaneous equation allows us to combine the estimates from the different models (Mize et al., 2019).

Our goal is to answer two questions: whether the coefficients of our key variables are equal across models with different dependent variables (non-lawmaking vs. lawmaking speech counts); and whether the difference between the respective marginal effects is statistically significant (Mize et al., 2019). For continuous variables, we compare the marginal effect of a change equal to one standard deviation. Results appear in Table 3.

Table 3.

Testing equality among coefficients and marginal effects in the two equations.

	Coefficients	Marginal effects
	Chi²	Model 2–Model 6
Authorities	0.29	−2.36
Authorities × party size	1.70	–
Party size	1.48	−0.95
Committee chair	2.38	−9.42
Female	18.86***	18.61**
Seniority	6.95***	−7.16**
Distance	24.44***	6.70**
Government	0.12	6.60
Alianza	1.21	14.66
Other party	1.49	6.98
Margin	0.18	−1.40
Retire	2.72*	−0.87

p < .1, **p < .05, ***p < .01.

Results from testing the equality of coefficients between the two models are shown in the second column of Table 3. They reveal that three of the coefficients differ significantly between the two models. We find considerable disparities (chi-square test) in the coefficients for female legislators, seniority, and district distance. An overall test (log-odds for the two models) also shows that once all variables are taken into account, the two models are significantly different.¹⁹

Comparing the marginal effects across models (third column of Table 3) provides us with more substantive information about differences in speech participation. For example, the marginal effect of being a female legislator in the first category (non-lawmaking) minus the marginal effect of being a female legislator in the second one (lawmaking) is roughly 19 speeches, a statistically significant difference. Regarding the effect of increasing distance by one standard deviation, the marginal effect in the first category minus the marginal effect in the second one equals close to seven speeches, which is also a statistically significant difference. The results also show that the marginal effect of seniority on the number of lawmaking speeches is significantly larger than its effect on the number of non-lawmaking speeches. An increase in tenure equal to one standard deviation leads to a cross-model difference of seven fewer speeches in the non-lawmaking category.

Interestingly, the marginal effects associated with the authorities’ variables are not significantly different between models. As mentioned previously, backbenchers are more likely than frontbenchers to give non-lawmaking speeches, particularly as party size increases. This effect was insignificant when analyzing lawmaking speeches. Comparing the marginal effects of authorities derived from each model reveals only a modest difference (i.e. authorities appear to deliver slightly fewer non-lawmaking speeches) that is not statistically significant.

To summarize, the empirical analysis found relevant differences between the two types of speech forums available to legislators in the Chilean Chamber of Deputies. We hypothesized that members of congress who are comparatively disadvantaged would be more likely than others to speak in non-lawmaking forums. Consistent with our expectations, the results show that in this forum, female legislators participate more often than male legislators; backbenchers participate more often than frontbenchers; and members from districts farther away from the center of government participate more often than those from nearer districts. We also hypothesized that, in the open forum dedicated to lawmaking debates, the participation of members with more expertise―committee chairs and senior members―and members of the opposition is more frequent. Results confirm our expectations regarding senior members and members of the opposition being more active in this forum. Committee chairs appear to deliver more speeches than others only at the end of the four-year legislative term.

Conclusions

Our study of legislative speech began with a discussion of electoral incentives and the rules that structure the participation of members of congress. Researchers have argued that personalized voting systems, such as the one in place in Chile, lead to more open rules that limit party leaders’ control over who takes the floor. In this context, electoral imperatives encourage legislators to use speeches for position taking, which can benefit the party as a whole. We notice, however, that the type of speech, more specifically, whether it was or was not an element of a lawmaking debate, also matters. Parties have fewer incentives to screen participants in settings dedicated to non-lawmaking speeches and greater incentives to develop norms that encourage expertise in settings dedicated to lawmaking debates.²⁰

In Chile, the rules structuring speechmaking during bill debates are rather open, and parties cannot formally screen who takes the floor. However, we noted that they are not particularly suitable for constituency-oriented position-taking. In contrast, speech participation during non-lawmaking stages is in the hands of party leaders. This forum is ideal for standing up for district needs. We argue that these differences are crucial to understanding which legislators are more likely to give speeches in each forum.

We found that it is in lawmaking speeches where parties, despite their lack of formal control, have developed norms that favor the participation of more senior members. In those stages where non-lawmaking speeches are delivered and parties have the power to screen members, comparatively disadvantaged members who appear more likely to benefit from connecting with constituents and cultivating a personal vote—backbenchers, members from more remote districts, and women—are more likely to speak. We agree that speech rules are mostly endogenous to the interests of the players, but also argue that in contexts wherein formal rules might create some inefficient collective behavior, agents can develop norms that foster coordination and mutual gains.

Footnotes

Funding

Juan Pablo Micozzi received financial support from the Asociación Mexicana de Cultura A.C. for research related to this article.

ORCID iD

Eduardo Alemán

Notes

Author biographies

Eduardo Alemán is an Associate Professor of Political Science at the University of Houston, where he teaches courses in comparative politics, political institutions, and Latin American politics. He is the author of over two dozen articles in peer-reviewed journals and coeditor of the book Legislative Institutions and Lawmaking in Latin America (Oxford University Press, 2016).

Juan Pablo Micozzi is an Associate Professor in Political Science at ITAM, in Mexico City. His research interests include legislative politics and political representation. He has published articles in such journals as the American Journal of Political Science and The Journal of Politics.

References

Alemán

Eduardo

(2006) Policy Gatekeepers in Latin American Legislatures. Latin American Politics and Society 48(3): 125–155.

Alemán

Eduardo

Micozzi

Juan Pablo

Ramírez

Margarita M

(2018) The Hidden Electoral Connection: Analysing parliamentary questions in the Chilean Congress. The Journal of Legislative Studies 24(2): 227–244.

Alemán

Eduardo

Ramirez

Margarita M

Slapin

Jonathan B

(2017) Party Strategies, Constituency Links, and Legislative Speech. Legislative Studies Quarterly 42(4): 637–659.

André

Audrey

Depauw

Sam

(2014) District Magnitude and the Personal Vote. Electoral Studies 35: 102–114.

Angell

Alan

(2003) Party Change in Chile in Comparative Perspective. Revista de Ciencia Política 23(2): 88–108.

Bäck

Hanna

Baumann

Markus

Debus

Marc

Müller

Jochen

(2019) The Unequal Distribution of Speaking Time in Parliamentary-Party Groups. Legislative Studies Quarterly 44(1): 163–193.

Bäck

Hanna

Debus

Marc

Müller

Jochen

(2014) Who Takes the Parliamentary Floor? the role of gender in speechmaking in the Swedish Riksdag. Political Research Quarterly 67(3): 504–518.

Baldez

Lisa

Carey

John M

(1999) Presidential Agenda Control and Spending Policy: Lessons from General Pinochet’s constitution. American Journal of Political Science 43(1): 29–55.

Brouard

Sylvain

Costa

Olivier

Kerrouche

Éric

Schnatterer

Tinette

(2013) Why Do French MPs Focus More on Constituency Work than on Parliamentary Work? Journal of Legislative Studies 19(2): 141–159.

10.

Díaz Rioseco

Diego

Giannini

Pilar

Luna

Juan Pablo

Núñez

Rodrigo

(2006) El Secreto de mi Éxito. Seis Caminos para Llegar y Permanecer en Valparaíso. Revista de Revista de Ciencia Política 26(1): 169–190.

11.

Dockendorff

Andrés

(2018) Legislative Behaviour in Chile: Three essays examining the effects of presidentialism and party politics on Chilean members of Congress. PhD Thesis, University of Essex, UK.

12.

Gamboa

Ricardo

Toro

Sergio Y

(2018) The Electoral Connection in Presidential Systems: Non-legislative actions inside the Chilean Congress. The Journal of Legislative Studies 24(3): 249–271.

13.

Gelman

Andrew

Stern

Hal

(2006) The Difference between Significant and Not Significant is Not Itself Statistically Significant. American Statistician 60(4): 328–331.

14.

Giannetti

Daniela

Pedrazzani

Andrea

(2016) Rules and Speeches: How parliamentary rules affect legislators’ speechmaking behavior. Legislative Studies Quarterly 41(3): 771–800.

15.

Heitshusen

Valerie

Young

Garry

Wood

David M

(2005) Electoral Context and MP Constituency Focus in Australia, Canada, Ireland, New Zealand, and the United Kingdom. American Journal of Political Science 49(1): 32–45.

16.

Kellerman

Michael

(2012) Estimating Ideal Points in the British House of Commons Using Early Day Motions. American Journal of Political Science 56(3): 757–771.

17.

Luna

Juan P

(2008) Partidos políticos y sociedad en Chile. Trayectoria y mutaciones recientes. In Arturo

Fontaine

Cristián

Larroulet

Jorge

Navarrete

Ignacio

Walker

(eds) Reforma de los Partidos Políticos en Chile. Santiago: UNDP, CEP, Libertad y Desarrollo, and Proyectamérica y Cieplan, 75–124.

18.

Maltzman

Forrest

Sigelman

Lee

(1996) The Politics of Talk: Unconstrained floor time in the US House of Representatives. The Journal of Politics 58(3): 819–830.

19.

Martin

Lanny W

Vanberg

Georg

(2008) Coalition Government and Political Communication. Political Research Quarterly 61(3): 502–516.

20.

Martin

Shane

(2011) Using Parliamentary Questions to Measure Constituency Focus: An application to the Irish case. Political Studies 59(2): 472–488.

21.

Mize

Trenton D

Doan

Long

J Scott

(2019) A General Framework for Comparing Prediction and Marginal Effects Across Models. Sociological Methodology 49(1): 152–189.

22.

Morgenstern

Scott

Polga-Hecimovich

John

Siavelis

Peter M

(2014) Ni Chicha ni Limoná: Party nationalization in pre- and post-authoritarian Chile. Party Politics 20(5): 751–766.

23.

Navia

Patricio

(2008) Legislative Candidate Selection in Chile. In Siavelis

Peter M

Morgenstern

Scott

(eds) Pathways to Power in Latin America. University Park, PA: Pennsylvania State University Press, 92–118.

24.

Palanza

Valeria

Scartascini

Carlos

Tommasi

Mariano

(2016) Congressional Institutionalization: A cross-national comparison. Legislative Studies Quarterly 41(1): 7–34.

25.

Pearson

Kathryn

Dancey

Logan

(2011) Elevating Women’s Voices in Congress: Speech participation in the house of representatives. Political Research Quarterly 64(4): 910–923.

26.

Pedrazzani

Andrea

(2017) Wasting or Saving Time? How government and opposition parties use time during legislative debates. Evidence from the Italian case. The Journal of Legislative Studies 23(3): 439–464.

27.

Proksch

Sven-Oliver

Slapin

Jonathan B

(2012) Institutional Foundations of Legislative Speech. American Journal of Political Science 56(3): 520–537.

28.

Proksch

Sven-Oliver

Slapin

Jonathan B

(2015) The Politics of Parliamentary Debate. New York: Cambridge University Press.

29.

Rocca

Michael S

(2007) Nonlegislative Debate in the US House of Representatives. American Politics Research 35(4): 489–505.

30.

Rosas

Guillermo

(2005) The Ideological Organization of Latin American Legislative Parties: An empirical analysis of elite policy preferences. Comparative Political Studies 38(7): 824–849.

31.

Siavelis

Peter M

(2002) The Hidden Logic of Candidate Selection for Chilean Parliamentary Elections. Comparative Politics 34(4): 410–438.

32.

Toro Maureira

Sergio

(2007) Conducta legislative ante las iniciativas del ejecutivo: unidad de los bloques políticos en Chile. Revista de Ciencia Política 27(1): 23–41.

33.

United Nations Development Programme (UNDP/PNUD) (2018) Desigualdad Regional en Chile: Ingresos, Salud y Educación en Perspectiva Territorial. Santiago, Chile: PNUD.