Seeking friends and influencers: business and labor groups lobbying in the Chilean congress

Abstract

This study analyzes direct lobbying in the Chilean Congress, contributing to the debate over which legislators are targeted by interest groups. Utilizing a comprehensive dataset constructed from legally mandated records of lobbying meetings, we test theoretical implications predominantly derived from the US context within a different presidential democracy. The focus is on the legislative targets of business and labor groups. The results reveal a marked preference for lobbying allies, aligning with recent theories of information transmission and legislative subsidies. This pattern holds true for both business and labor groups and highlights the significance of ideological alignment for legislative lobbying in Chile. Additionally, the study finds that legislators with influential positions, such as those on key committees or centrally located in the bill collaboration network, are more frequently targeted. This research provides key insights into the dynamics of legislative lobbying in a non-US context, underscoring the generalizability of established theoretical frameworks.

Introduction

Politicians are often lobbied by businesses, labor, and other interest groups seeking to influence policy. Lobbying is a controversial activity. Some see it as a corrupting influence in politics, giving undue influence to a narrow set of powerful actors, while others believe it can improve the quality of policymakers’ decisions by offering information and giving insight into complex matters. Because it often occurs in the shadows, proposals to regulate it tend to involve improving “transparency to ensure that public officials, citizens, and businesses can obtain sufficient information on lobbying activities.” ¹ In the United States, for example, the Lobbying Disclosure Act (LDA) and Foreign Agents Registration Act (FARA) impose various regulations on individuals and entities seeking to lobby public officials. Nevertheless, under the LDA, lobbyists are not required to disclose their meetings with public officials, something that, according to FARA, only needs to be reported if the lobbying is conducted on behalf of foreign principals. ² In contrast, the European Parliament requires rapporteurs, shadow rapporteurs, and committee chairs to report any meetings with interest representatives seeking to influence policy or institutional decision-making. ³ Information on lobbying contacts is crucial to understanding the interactions between interest groups and elected officials.

One facet of the interaction between policymakers and organized groups that has attracted significant scholarly attention is the decision of whom to lobby. ⁴ Understanding which politicians are the focus of lobbying efforts is important for grasping interest group strategies and the potential impact of lobbying efforts on legislative outcomes. As Hojnacki and Kimball note, targeting opposing or undecided legislators may allow organized interests to sway policy in their favor, while lobbying allies is a more subtle process that reinforces existing preferences and encourages participation in the legislative process to advance their objectives. ⁵ This insight is particularly revealing given that the individual contacts sought by these groups often aim to shape the collective decisions of elected assemblies. ⁶

Most empirical studies and theoretical approaches examining the target of lobbying efforts have focused on the United States, but more recently, works have expanded this analysis to the European Union ⁷ and the parliaments of some other high-income nations. ⁸ In this article, we expand the scope of these studies by focusing on Chile, a presidential country in Latin America that has changed its lobbying regulations to require the disclosure of lobbying contacts.

Leveraging newly available data from eight years of lobbying contacts between organized interests and legislators, we examine the targeting strategies of business and labor groups. Our study contributes to the literature on lobbying and legislative politics in two significant ways. First, it evaluates alternative hypotheses on whether organized interests preferentially target legislators viewed as friends or foes. These propositions derive from theories that have primarily been tested with data from the case (United States) that gave rise to such theories in the first place. To our knowledge, this study is the first to systematically examine lobbying targets in a presidential country other than the United States. Although most formal models on this research topic are informed by the US case, they offer broad theoretical insights intended to be generalizable. Evaluating these theories with empirical evidence from other presidential countries is important to ascertain their generalizability.

Furthermore, many empirical studies of lobbying targets by domestic organized interests in the US Congress rely on surveys of lobbyists, many of which suffer from a small number of respondents and potential sampling and response biases that are seldom addressed. Many others examine a single organized interest or are case studies focused on a single event. Our data, however, is based on comprehensive records of meetings legislators had with business and labor organizations over eight years, reported in compliance with Chilean lobbying law. Our results reveal that organized interests most often contact potential friends, although they do not exclusively focus their efforts on such members.

Second, we investigate whether organized interests target influential legislators more frequently. Prior research from the US Congress indicates that organized interest groups commonly target members assigned to jurisdictionally relevant committees. However, we argue that a legislator’s centrality within the chamber’s social network also plays a significant role. Over the last two decades, a growing body of research on legislative social networks has emerged that associates connectedness with influence. ⁹ Our findings show that legislators occupying a more central position are more likely targets of organized interests, as are those assigned to key committees.

The rest of this article is divided into five sections. The next one discusses the longstanding debate regarding the primary targets of interest group lobbying. Subsequently, we turn our attention to the Chilean case. This section justifies the selection of Chile as an illustrative case for examining the empirical implications of this debate. It also provides specific examples of lobbying topics and formulates four hypotheses. Additionally, it addresses the relevance of business and labor interest groups in Chile and their alignment with various ideological blocs. Following this, we discuss the data and methodology employed in the analysis. The subsequent section presents our results. The article ends with a brief conclusion summarizing our findings.

Who is being lobbied

Lobbying serves as a means for organized interests to influence policymaking. ¹⁰ Theories of lobbying tend to differ on the nature of such efforts. For instance, some view it as a method of persuasion, while others consider it a way to subsidize allies; some conceptualize it as a form of exchange, while others describe it as an informational process. These theories also vary in their empirical implications, particularly regarding the critical question of who will most likely be lobbied.

One influential perspective portrays lobbying as a form of exchange that is typically modeled as a vote-buying game. ¹¹ In these models, organized interest groups provide campaign contributions (or bribes) in exchange for favorable legislative votes or other legislative benefits. ¹² For example, Snyder posited that the target of such efforts would be legislators slightly opposed to the proposed change. ¹³ Along similar lines, the implication of Dekel et al.’s model of two opposing lobbyists engaging in a continuing bidding process for legislators’ votes is that a lobbyist would not make an offer to a legislator who favors her position. ¹⁴ While some vote-buying models of competitive lobbying predict that resources would be directed to some legislators favoring the interest group’s position, ¹⁵ the standard expectation is that most lobbying efforts would not be concentrated on anticipated supporters or unequivocal opponents. ¹⁶

Critics of this perspective have underlined the difficulties involved in enforcing such trades, ¹⁷ and supporting evidence from the US case is mixed. ¹⁸ Data on campaign contributions, for instance, show that, for the most part, they are not directed to the hypothesized targets. In Chile, the focus of our study, a series of reforms passed on the heels of some prominent scandals culminated with the prohibition of campaign contributions from private sources other than individuals. This regulatory framework poses significant challenges for interest groups seeking to trade contributions for legislative favors and makes enforcing these exchanges more difficult.

Building upon the concept of competitive lobbying, Austen-Smith and Wright theorized that interest groups not only engage in lobbying uncommitted legislators and those inclined to oppose them but also target some legislators who are, a priori, anticipated to back the group’s stance. ¹⁹ They focus on the first two types to persuade them to support the group’s effort. The rationale behind targeting allies, however, is to counteract the lobbying efforts of opposing groups. One implication of their model is that interest groups are unlikely to lobby legislators with a low probability of altering their positions (i.e., staunch supporters or unequivocal opponents) and would not lobby friendly legislators more than those inclined to oppose them. In this theory, interest groups provide legislators with information rather than campaign contributions.

The idea that information is the primary currency underpinning the relationship between lobbyists and legislators has become pervasive in the relevant literature. ²⁰ Organized interests supply legislators with valuable expertise, policy insights, and perspectives from their constituencies, aiding their decision-making. They present arguments for and against various policies. This view of information sharing as central to the interactions between legislators and lobbying groups extends to studies of the European Union ²¹ and is consistent with reports of lobbying meetings in Chile, as discussed in the next section. Research shows that, despite potential biases from interest groups, lobbying generally leads to more informed legislators. The likelihood of receiving biased information is mitigated by factors like repeated interactions, verifiable information, competition among information providers, legislators’ pre-existing beliefs, and interest groups’ concerns about credibility. ²²

Earlier studies on informational lobbying as a tool for persuasion were criticized for failing to explain an empirical regularity emerging from the US case: interest groups predominantly lobby legislative allies who share their views. ²³ Later models explained the existing pattern by suggesting that allies help garner broader legislative support. ²⁴ For example, Schnakenberg posited that interest groups prefer to lobby legislators who are already allies, providing them with information that can be used to sway opponents. ²⁵ A central implication of his model is that most informational lobbying occurs via intermediaries—allies motivated to disseminate information that could persuade initially resistant legislators. Awad further developed this idea, arguing that intermediaries can tailor information to benefit interest groups. An implication of his model is that moderate allies are more persuasive intermediaries than more extreme ones. ²⁶

An alternative perspective argues that the information provided by interest groups functions as a legislative subsidy rather than as a tool of persuasion. ²⁷ This subsidy helps in formulating and implementing policies, easing legislators’ budget constraints, and achieving shared objectives with their allies. ²⁸ The notion that lobbyists function as “service bureaus” to ex-ante sympathetic legislators goes back to the pioneering work of Bauer et al. ²⁹ Hall and Deardoff theorized that legislative subsidies are targeted at the interest group’s strongest allies. ³⁰ Because informational subsidies are not conceived as tools for persuasion, opponents (“enemies”) are not targeted, and the uncommitted, who may or may not use the subsidies to support the group’s objective, are seldom targeted.

More recently, Ellis and Groll differentiated between two types of legislative subsidies: information acquisition and policy implementation. ³¹ Their model suggests that policy implementation subsidies go to allied legislators lacking resources to enact desired reforms, whereas information acquisition subsidies target legislators whose chances of supporting the group’s policy increase by being lobbied. Ellis and Groll contended that information subsidies can induce legislators to consider previously ignored options and persuade them to support the group’s position.

In summary, the question of whether friends, foes, or those somewhere in-between will be the target of interest groups’ direct lobbying activities has generated substantial scholarly attention. Exchange theories and early informational theories predicted the targeting of uncommitted legislators, moderate opponents, and (possibly) moderate allies. Committed allies or strong foes were portrayed as unlikely targets. More recent theories of information as persuasion predicted allies as the most likely targets because they serve as helpful intermediaries, but moderate allies were depicted as more likely to fulfill this role than extreme ones. ³² In contrast, Hall and Deardorff’s theory of information as a legislative subsidy predicted strong allies as the most likely targets and opponents as unlikely to receive information subsidies. ³³ Some approaches, like counteractive lobbying and vote-buying, focus on legislative votes to derive expectations about whom groups would target in their lobbying efforts. ³⁴ However, other approaches recognize more explicitly that the targeting of legislators can be driven by considerations that extend beyond the realm of vote persuasion.

In addition to a legislator’s alignment with an interest group, research focusing on the US Congress has identified influential members, such as committee leaders ³⁵ and members of key committees, ³⁶ as prime targets for lobbying efforts. A committee’s specialized jurisdiction and agenda-setting authority endow its members with significant influence to shape outcomes favoring the lobbying group’s interests. Members serving on key committees are in an advantageous position to supply informative cues ³⁷ and influence decisions on issues impacting the policy interests of related groups. ³⁸

Empirical studies on lobbying targets within legislative bodies have primarily centered on the US Congress, where findings suggest that interest groups favor allies and key committee members while also lobbying moderate opponents. ³⁹ Similar patterns have been observed in the European Parliament, Denmark, and the Netherlands. ⁴⁰ The rest of this paper is dedicated to exploring the case of Chile, thus expanding this research to a region where such questions have been largely understudied.

Interest group targeting in Chile’s chamber of deputies

Empirical research on interest group lobbying in Latin America is limited, ⁴¹ with scant studies focusing on the targeting of legislators by interest groups. A notable exception is Benzecry’s qualitative case study of a labor group’s lobbying strategies during the 1988 National Constituency Assembly in Brazil. ⁴² Benzecry found that the group focused its direct lobbying activities on allies, providing them with legislative subsidies such as information and assistance in policy implementation. Conversely, the group indirectly targeted uncommitted foes with tactics like public demonstrations, potentially increasing the cost of opposing its preferred policies.

Our analysis focuses on the Chilean case. Several reasons make Chile an ideal case to advance the study of who gets lobbied. One is the availability of detailed data on legislators’ direct lobbying contacts. This information stems from new regulations enacted in 2014 (Law 20,730), which mandate public officials to record lobbying contacts with groups or individuals, positioning Chile as the first Latin American country to require detailed lobbying disclosures. US lobbying practices, for the most part, do not require disclosing meetings between legislators and lobbyists, leading most research on this area to rely on surveys or lobbying expenditure data. ⁴³ As noted in the introduction, surveys of legislators have non-trivial limitations. Empirical analyses of the European Parliament and other European countries also have predominantly used survey data ⁴⁴ and interviews ⁴⁵ rather than data on direct lobbying contacts. Moreover, Bombardini and Trebbi note that “individual European countries have almost no systems to record the activity of lobbyists.” ⁴⁶

Additionally, the Chilean Congress stands out as one of the most effective assemblies in the region. ⁴⁷ It boasts a robust committee system, ⁴⁸ a reelection-oriented professional membership, and political parties with clearly articulated ideological stances. Chile also features a wide array of interest groups seeking contact with legislators, ⁴⁹ and linkages between prominent interest groups and the country’s parties are well-established. ⁵⁰ Thus, we believe that Chile falls within the domain of the lobbying theories developed with the United States in mind.

Furthermore, the examination of the Chilean case offers a valuable opportunity to assess if the findings from previous studies are applicable beyond their original context. This approach is reinforced by the research design literature, which highlights the significance of testing theories in different cases from those where the initial hypotheses were formulated. ⁵¹

Most formal theories that explore the lobbying targets of special interests present their arguments through mathematical models in a manner that suggests their applicability is not confined to the specific circumstances of the US Congress. This broader applicability is evident in Schnakenberg and Turner’s recent review of this literature. ⁵² Consequently, formal theories of lobbying targets would greatly benefit from empirical validation across diverse settings. For example, formal models of legislative politics and organization, conceived initially with the US Congress in mind, have evolved into foundational elements in comparative legislative studies due to their proven applicability in numerous other contexts. Therefore, research on lobbying targets stands to deepen its insights and broaden its influence by following the precedent set by formal models of legislative politics through examination in other contexts. Our study aims to further this objective.

Meetings between organized interests and Chilean legislators, as revealed in the publicly available documents, often exhibit varying levels of detail regarding the topics addressed. Analysis of these records, however, indicates that a predominant focus of these meetings is the transmission of information. For example, on April 12, 2019, amidst congressional debates on new 5G network regulations, a representative from the company 5G América met with Deputy Giorgio Jackson to discuss the potential of their specialized telecommunications technology. On September 30, 2016, a nurses’ union representative met with Deputy José Pérez Arriagada to provide him with “detailed information” regarding a bill aimed at altering the Sanitary Code, specifically affecting midwives, which was pending a vote. Similarly, during the latter half of 2019, a representative from Agroval, an association of technical and professional training institutions, met with six deputies to inform them about the effects of a controversial employee training bill being debated in Congress. Additionally, on May 6, 2015, during deliberations on a healthcare law regulating treatments for high-cost illnesses and access to expensive medications, an Abbvie Laboratories representative briefed Deputy Manuel Monsalve on the company’s business operations.

Some meetings are held to learn about legislators’ positions on particular issues. On May 8, 2018, a representative from Hábitat, a private pension fund administrator, met with then Deputy Gabriel Boric to discern his views on a potential pension reform. Likewise, on May 5, 2015, a representative of the mining company Anglo American Chile consulted with Deputy Daniella Cicardini, seeking to learn her perspective on challenges in the mining sector. Notably, both deputies were newly elected legislators at the time of these meetings. This dynamic, where information flows from legislators to interest groups, was highlighted by Heberlig in his study of union lobbyists targeting first-term members of the US Congress. ⁵³

At other times, organized interests engage with legislators to request their intervention with government agencies and bureaucrats to advocate on policy and regulatory matters on their behalf, as Hall and Miler noted for the US Congress. ⁵⁴ For example, on April 6, 2018, members of a business group representing small and medium fishing companies met with Deputy Leonidas Romero Sáez, requesting him to join them in a meeting with the Undersecretary of Fisheries to advocate for changes to a fishing law. On August 2, 2017, representatives for employees in the Public Defender’s unit met with Deputy José Ortiz to solicit his help in their dealings with the Ministry of Justice over ongoing arrangements to transition employees from precarious, less secure positions to more formal, stable employment statuses. Between March and May of 2017, members of a union representing preschool teachers met with three far-left deputies, asking them to intervene on their behalf with the Ministry of Education on the implementation of a regulatory decree that they argued was negatively affecting the hiring of teachers.

Sometimes, the discussion of legislation is explicitly stated. For instance, on January 11, 2016, a Tabbaco Farmers representative informed Deputy Hugo Gutiérrez about the potential repercussions of a proposed bill on over a thousand families engaged in this sector, seeking to persuade him to vote against the bill. Similarly, during the latter half of 2019, a state preschool workers’ union representative met with three different deputies to voice concerns about early education bills proposed by the government and to request the legislators’ support in opposing these initiatives. In another instance, on April 12, 2016, a representative from Chile’s public employees’ union ANEF met with Deputy Gabriel Boric amid an ongoing strike in the Atacama region to persuade him to reject a recent government bill intended to resolve the labor dispute.

Upon reviewing the reported topics from the meetings between legislators and businesses, which are briefly described and differ in specifics, we observed that approximately 20% mention a bill or law. The proportion of meetings between legislators and unions where a bill or law is mentioned in the topics is about 29%. Notably, only a limited number of bills are referenced across multiple meetings, and these rarely involve meetings conducted by the same entity.

One notable example revolves around the pension reform that President Sebastián Piñera proposed to Congress in 2018. The bill had the support of parties on the right of the political spectrum and was backed by businesses, especially private pension fund administrator companies. Conversely, parties on the left and labor unions, like the country’s leading workers’ federation, opposed it. Based on the information disclosed, we identified 23 meetings with legislators about this reform since its announcement in October of that year—15 of which involved five businesses and eight involved five labor unions. Two-thirds of the legislators targeted by businesses belonged to right-wing parties, and the left-wing legislators contacted by businesses were members of two key committees with jurisdictions over the reform (Finance and Labor). Also, three-fourths of the legislators contacted by the unions belonged to leftist parties that opposed the reform.

While a significant portion of meetings between organized interests and legislators involve businesses and labor groups, only a few individual firms and unions have numerous meetings with legislators. Specifically, during the first legislative period studied, 472 different businesses engaged in meetings with legislators, with this number increasing to 619 in the next period. In both periods, the median number of meetings was one per firm, and the 75th percentile was three meetings in the first period and two in the second. A small fraction of businesses—5.5% in the first period and 3.7% in the second—conducted ten or more meetings with legislators. Regarding labor organizations, 323 met with legislators in the first period and 328 in the second. The median number of meetings held was one across both periods, and the 75th percentile was two meetings. Notably, 3.7% of unions in the first period and 4.6% during the second period held ten or more meetings with legislators.

Hypotheses

Leveraging data on lobbying contacts from 2014 to 2022, this study aims to evaluate a series of hypotheses derived from the theoretical literature on lobbying targets, as discussed in the preceding section. The analysis focuses on the type of legislator business and labor groups target. The first hypothesis is grounded in the proposition that allied legislators are more frequently targeted than adversaries. This aligns with recent theories of information transmission ⁵⁵ and legislative subsidies. ⁵⁶

H1. Allied legislators are contacted more often than opponents.

The second hypothesis posits that legislators identified as steadfast allies or unequivocal adversaries are engaged less frequently by organized interests compared to their more moderate counterparts. This assumption is predicated on the belief that these legislators are considered unswayable and, as a result, targeting them would be an inefficient use of resources. Instead, moderates, perceived as more susceptible to persuasion, will be the primary targets of lobbying efforts. This hypothesis aligns with the theory of counteractive lobbying and views on lobbying as a form of exchange.

H2. Strong supporters or unequivocal adversaries are contacted less frequently than their moderate counterparts.

Furthermore, the analysis also enables us to comment on other expectations derived from the specialized literature. These include examining whether interest groups favor moderate allies over staunch ones ⁵⁷ and whether they avoid engaging with foes altogether ⁵⁸ or choose to meet with friends and foes. ⁵⁹

Two additional hypotheses focus on the likelihood of influential legislators being targeted. The first posits that interest groups are more inclined to contact legislators who are members of key committees. These legislators have the authority, opportunity, and motivation to shape legislation in areas pertinent to the interest group.

H3. Membership in key committees is associated with more frequent contact with interest groups.

Lastly, we argue that influence is also associated with a legislator’s position in the network of linkages that develops among the congressional membership. Connecting with other members is a pathway to bridge-building and enhancing knowledge of other actors, policies, and legislative strategies. Research in social networks consistently underscores that an individual’s influence is rooted not only in context or resources but also in an actor’s placement within the network of social connections. ⁶⁰ Occupying a more central role within a network typically indicates greater influence and deeper knowledge of the network’s dynamics, which aids in identifying and recruiting allies. ⁶¹

Legislative studies have emphasized the importance of bill collaboration networks. ⁶² Such connections are thought to be indicative of influence and facilitate the transmission of information. ⁶³ Better-connected legislators have been shown to be more effective in passing legislation ⁶⁴ and securing federal grant money. ⁶⁵ However, unlike research centered on bill initiation and passage, current empirical studies on congressional lobbying have not yet explored the effect of a legislator’s social network position.

Consequently, we consider a legislator’s centrality in the bill coauthorship network as a reliable indicator of influence. Better-connected legislators are more likely to act as effective intermediaries, disseminating information that aligns with an interest group’s agenda. ⁶⁶ This perspective forms the basis of our fourth hypothesis.

H4. Centrality in the bill coauthorship network is associated with more frequent contact with interest groups.

Business and labor

This study concentrates on the legislative targets of business and labor groups. These groups are typically better organized and possess more resources than other entities, such as cause-oriented groups. They are also more extensively researched, which aids in identifying legislative allies and opponents. As de Figueiredo and Cameron observe, these interest groups often adopt ideological stances and are commonly allied with major political parties. ⁶⁷ They give the United States as an example, where labor groups generally align with the Democratic Party and businesses with the Republican Party.

In Chile, there is a consensus among scholars that business sector representatives are closely aligned with right-leaning parties. ⁶⁸ Conversely, labor groups maintain strong connections with left-leaning parties. ⁶⁹ This alignment of business (labor) and right (left) parties is not unique to Chile and has been observed in lobbying studies of other contexts. Otjes and Rasmussen identified a similar trend in Denmark, and Wessels noted it within the European Parliament. ⁷⁰ Similar patterns have been noted in Mexico ⁷¹ and Brazil. ⁷²

Chile’s business sector has had a significant influence on policy decisions, a power attributed to its influential cross-sectoral association, the Confederation of Production and Commerce (CPC), and its ties to right-wing political parties. ⁷³ The country’s two major rightist parties, Independent Democratic Union (UDI) and National Renewal (RN), coalesced into one of the two coalitions that dominated political competition following the transition to democracy. Giraudy’s analysis of party press releases revealed UDI’s forceful pro-business advocacy. ⁷⁴ She further illustrates that both RN and UDI had robust pro-business platforms and a significant proportion of leaders from business groups. ⁷⁵ During the period studied, the coalition’s junior party—initially Amplitud and later Evópoli—had fewer longstanding business connections but clearly supported policies favoring private enterprise and free market competition.

In contrast, the Chilean labor movement has historically been aligned with the political left. The country’s main labor federation, the Central Union of Workers of Chile (CUT), participated in the “No” campaign organized for the 1988 plebiscite that denied General Pinochet another eight years in the presidency. The CUT also supported the center-left coalition that emerged victorious in the subsequent elections that began the new democratic era. During the period studied here, the CUT was led by Bárbara Figueroa Sandoval, a member of the teacher’s union (CPC) and the Communist Party (PCCh). She ran for Congress in 2009, endorsed by a coalition of center-left parties and the PCCh. Since 1996, the leadership of the CUT has alternated between members of the Socialist Party (PS) and the PCCh.

The CUT includes numerous public employee unions belonging to the National Association of State Employees (ANEF), an influential umbrella organization. The ANEF’s leadership, under Raúl de La Puente (PS) from 1996 to 2016, under Carlos Insunza (PCCh) from 2016 to 2018, and under José Pérez Debelli (PS) thereafter, underscores this alignment. Additionally, the Federation of Copper Workers (FTC), the largest union in the country and a prominent member of the CUT, was led by Raimundo Espinoza, a PS member, for 25 years. Chile’s two other union federations besides the CUT are the Autonomous Central Union (CAT) and the National Union of Workers (UNT). During the period studied here, CAT was led by Óscar Olivos, a former militant of the PS, and the UNT was led by Segundo Steilen, a member of the Christian Democratic Party (DC), the more moderate party of the center-left coalition.

While the PCCh has consistently adopted pro-union stances, the center-left parties embraced a pro-market model for the first 16 years since the return to democracy ⁷⁶ and often clashed with the more leftist unions. Despite their inclination to foster cooperative relationships with businesses, these parties remained the primary congressional allies for unions. When the center-left parties joined the PCCh to form the Nueva Mayoría coalition in 2013, they moved leftward, aligning more closely with union positions. The CUT endorsed the Nueva Mayoría in the 2013 elections, with its leader, Bárbara Figueroa, emphasizing the coalition’s commitment to strengthening unions and collective bargaining.

Another political group consistently adopting pro-labor stances is the Frente Amplio (FA), a coalition of small leftist parties. Democratic Revolution (RD), the main founding party in this coalition, appointed union leader Margarita Portuguez as its president in 2021, signaling its pro-labor orientation. Although the FA has garnered support from some prominent union leaders, it has yet to develop more formal linkages to the unions, as the traditional leftist parties have done.

Having delineated the political affiliations of the two interest groups that form the core of this analysis, we now move to the subsequent section, where the data, variables, and statistical model employed in the research are described.

Data and Methods

This study utilizes official records of lobbying meetings undertaken by members of Chile’s Chamber of Deputies. This information is compiled and made public by the Council for Transparency, an autonomous agency created by law to further the goals of government transparency and access to state information. The Lobby Law (law 20,730) requires public officials, not only members of Congress, to disclose all meetings with organized interests, including those with lobbyists and individuals promoting, defending, or representing any particular interest with the objective of influencing decisions made by public officials in exercising their functions. ⁷⁷ The type of public officials covered by this law is vast and extends to various levels of the bureaucracy, armed forces, prosecutor’s office, central bank, and local officials, such as mayors. Information compiled by this office is organized in such a way that one can easily obtain records for each member of Congress. These records allow us to identify and categorize meetings conducted by legislators, the date of such encounters, and the identity of the participating entity. Information on the topic of the meeting is also included but filled with different degrees of specificity, as noted in the previous section.

We focus on four types of groups: big business, other business, labor, and professional. Big business consists of enterprises that are part of the influential CPC. This group includes leading firms in banking and finance, mining, construction, agriculture, commerce, and industry. The second category encompasses other private businesses, including a wide array of small businesses. The third category consists of labor unions. Lastly, the fourth category includes professional interest groups, such as engineers, medical technicians, nurses, and chemists. These groups do not have the longstanding linkages with left-of-center parties that labor unions have, and their objectives center on protecting professional interests rather than collective bargaining. Nevertheless, they are included in our analysis to explore the extent to which their lobbying targets align with those of labor.

During the eight-year span under study, there were 5,129 reported meetings with these four types of organized groups. Of these, 18.6% were with big business organizations, 41.4% with other businesses, 31.3% with labor unions, and 8.8% with professional interest groups. These meetings represent 72% of the total meetings held by members of the Chamber of Deputies with organized interests. The other 28% of meetings between organized interests and deputies refer mostly to so-called cause interest groups, such as those promoting a single issue (e.g., advocating for the rights of people with health conditions or impairments or promoting specific cultural activities) and to a lesser degree, civic groups (e.g., neighborhood associations). ⁷⁸

Our dependent variable is a count Y _ik describing the total number of meetings each legislator i had with representatives of each of the four interest groups k over two four-year legislative terms, 2014–2018 and 2018–2022. Rather than running four separate regressions where the outcome is the number of meetings a legislator had with each interest group type, we stack the data together. Thus, each legislator in a congressional period is measured four times. The overall mean number of meetings per legislator is 4.5. The mean number of meetings legislators had with big business is 3.4, with other business is 7.5, with labor is 5.6, and with professional groups is 1.6.

Two of the three hypotheses (H2–H3) apply to all types of interest groups. Running an unrestricted model, estimated jointly by member i and group k, is more efficient than running four separate regressions, allowing us to get better estimates of the effects of committee membership, legislator centrality, and extremism (and also the control variables). The expectation derived from the first hypothesis, however, is that the slope for ideology would be positive when meeting business groups (the number of meetings increasing as we move rightwards) and negative for meetings with labor (the number of meetings decreasing as we move rightwards). Thus, we interact the ideology variable with dummies indicating interest group categories (other business, labor, and professional, with big business being the baseline). ⁷⁹ Restricted models, one per interest group category, are reported in the appendix and lead to similar conclusions.

Given that our dependent variable is an overdispersed count outcome, we use a negative binomial model with robust standard errors. In addition, because we have repeated observations of the outcome per subject, which are not independent, we treat subjects as a random factor and run a multilevel model with legislator random effects. Likelihood-ratio tests show enough variability between legislators to favor the inclusion of random effects over a standard negative binomial.

The key independent variable used to gauge whether legislators and interest groups are friends or foes is ideology. It measures individual ideal points estimated from roll call votes. In Chile’s fragmented party system, parties occupy various ideological positions. Ideal point estimations capture legislators’ positions well along one dimension of conflict, which analysts have identified as the ideological left-to-right continuum. ⁸⁰ Moreover, the ordering of party positions along this dimension coincides with expert assessments of parties’ ideological positions and party orderings derived from surveys of legislators and the general population. ⁸¹

We use the WNominate technique to calculate ideal points, ⁸² though almost identical results are achievable with other methods. We bridged roll call data from both legislative periods to ensure the comparability of ideal points. The ideal point scores range from 1 (far right) to -1 (far left). ⁸³ The ordering of parties along this dimension is consistent with previous works on the ideological positions of Chilean parties.

To examine our second hypothesis, we introduce a variable labeled centrism/extremism, which measures the absolute distance from the ideological midpoint between the two blocs. This midpoint is between the leftmost legislator in the rightist coalition and the rightmost legislator in the leftist coalition, as measured by the ideal points. Thus, moderate legislators—less likely to be identified as staunch allies or opponents—have low values.

In addition, we account for membership in key committees by incorporating three categorical variables. The first indicates whether a legislator was assigned to the Finance Committee. This is the money committee, and all bills that include expenditures must pass through it. The second variable reflects membership in the committees overseeing the country’s crucial natural resources, specifically mining and fishery. The third variable indicates membership in the labor and public works committees. ⁸⁴

To capture influence based on a legislator’s position within the coauthorship network, we calculate two classic centrality measures derived from the social networks literature: closeness centrality and betweenness centrality. ⁸⁵ The first captures how close a legislator is to others in the network. It measures the shortest distance between an individual and all other network members. Betweenness centrality captures the extent to which a legislator lies on the shortest path between others. It is commonly interpreted as indicative of an individual’s influence over the flow of information—the extent to which an actor acts as a bridge between different parts of the network. ⁸⁶ It measures the proportion of all shortest paths between two actors in the network that pass through a given legislator. Chile’s coauthorship network is dense, with significant clustering, and both measures are highly correlated (0.78), so we include them alternatively in our analysis.

Lastly, we add a series of control variables. The first captures whether a legislator is in her first term in office. There are reasons to expect new members to be less influential than more senior ones and, therefore, less likely to be targeted. However, as noted before, interest groups may seek contacts with new legislators to gauge their positions. The second control variable measures the rurality of the legislator’s district. The density of labor unions and businesses tends to be lower in more rural districts, which may impact the degree to which interest groups target legislators. The third control indicates whether the legislator is a woman. This variable seeks to capture whether gender effects influence interest group targeting, particularly given that women, for a long time, represented only a small share of the chamber’s membership. Lastly, we add a categorical variable for the first of our two periods, 2014-2018. Summary statistics for all variables are presented in a table in the appendix.

Results

Table 1 presents our main results. Coefficients are highlighted in bold, with the corresponding standard errors presented directly beneath them. The last rows show the variance component, the number of cases, the number of unique subjects (random intercepts per legislator), and the Akaike information criteria (AIC). The table presents three alternative model specifications (labeled M1 to M3) using the same dependent variable. The second (M2) includes the same independent variables as the first (M1) except for the centrality measure (closeness is included in M1 and betweenness in M2). The third specification (M3) includes the centrism/extremism variable previously described instead of the ideology variable.

Table 1. Results

Note: * p < 0.1, ** p < 0.05, *** p < 0.01

To evaluate the implications of the first hypothesis (H1), we estimate the effect of a legislator’s ideological position on the number of meetings she has with the four types of interest groups. As previously indicated, the effect of ideology on the number of meetings depends on the type of group involved, which is why we include an interaction between these variables. The predicted number of meetings, based on M1, is illustrated in Figure 1. The top panels display the predicted number of meetings with the two business categories, while the bottom panels show the predicted number of meetings with labor and professional groups. Ideal point estimates are plotted on the horizontal axes. Within this figure, black lines denote the point predictions, and these are surrounded by gray shades, which represent the 95% confidence intervals.

Figure 1. Predicted number of meetings by legislators’ ideal points.

The results depicted in Figure 1 align with H1. They show that legislators’ meetings with business groups increase as their ideological stance shifts rightward. Conversely, their meetings with labor and professional groups increase as their ideological stance moves leftward. For example, moving from the left bloc’s median ideological position to the right bloc’s median ideological position increases the predicted number of meetings with big business groups from 2.5 to about 4.5. In other words, as a legislator’s ideology shifts from a typical leftist position to a typical rightist position, the number of meetings with big business entities increases by 80%, indicating a significant effect. A similar shift in ideology changes the predicted number of meetings with other business groups from 6.2 to 9.3, which amounts to an increase of 50%.

Conversely, moving from the median ideological position of the left bloc to that of the right bloc results in a 37% decrease in the number of meetings between a legislator and labor groups, dropping from 6.7 to 4.2. A similar shift in the ideology variable leads to a less substantial impact on the number of meetings with professional groups, reducing their frequency from 1.8 to 1.3, a decrease of approximately 28%. These shifts in ideology result in statistically significant changes at p < .01 for meetings with the first three categories of interest groups.

The hypothesis (H2) that less persuadable legislators, represented by more extreme ideologies, are less likely to be targeted compared to more persuadable moderates is tested in M3. Contrary to H2, the coefficient for centrism/extremism, which measures the absolute distance from the midpoint between the left and right blocs, is positive and lacks statistical significance. The result does not support the notion that organized interests more frequently target moderates over legislators adopting more extreme positions, challenging the implications derived from counteracting lobbying and lobbying as a form of exchange that consider more extreme members less persuadable and, therefore, less likely to be targeted.

The results also show that serving on key committees increases the frequency of contacts with organized interests, consistent with the third hypothesis (H3). The coefficients are statistically significant across the different models shown in Table 1. For instance, belonging to the Finance Committee increases the number of meetings by about 31% (from 4.4 if the legislator is not in a key committee to 5.7). Similarly, increases in the number of meetings associated with serving on committees related to natural resources (by 25%) or labor and public works (by 21%), though slightly lower, are still significant.

In addition, the results also lend support to the fourth hypothesis (H4). Both the closeness centrality and the betweenness centrality variables exhibit positive and statistically significant coefficients. Specifically, moving from one standard deviation below the mean to one above the mean in closeness centrality is associated with a 62% rise in the predicted number of meetings (increasing from approximately 3.5 to 5.6, based on M1). Similarly, an equivalent increase in betweenness centrality leads to an increase of 45% in the predicted meeting count (estimated using M2).

Regarding the control variables, two findings stand out. The first is that women are significantly less likely to be lobbied. That this is the case after controlling for other relevant factors raises significant questions about the interest groups’ perceptions of women’s roles in Congress. While exploring such questions is beyond the scope of our analysis, it is a revealing finding nonetheless. The second one is that legislators belonging to rural districts are less likely to be targeted. Moving from one standard deviation below the mean district rurality to one above the mean increases the predicted number of meetings by 19% (estimated using M1).

In summary, our empirical findings demonstrate that interest groups lobby allied legislators significantly more often than foes. This pattern aligns with recent theories of information transmission ⁸⁷ and legislative subsidies ⁸⁸ and mirrors similar observations in the analyses of the US Congress. However, this finding diverges from the standard expectation set by theories of exchange or counteracting lobbying.

Lastly, we run a series of additional models as a robustness check, with the full results shown in the online appendix. Two of these models replace the ideology variable with categorical variables representing different partisan groups.

In the first, we include a dummy variable indicating whether the legislator belongs to a party considered to be an ally of the organized interest. More specifically, this variable takes the value of 1 if the legislator belongs to the right bloc when the number of meetings refers to contacts with business groups and if she belongs to the left bloc when the number of meetings refers to contacts with labor and professional groups. Otherwise, this variable takes the value of 0. The results of this model are consistent with the previously presented findings, showing that organized interests more frequently target legislators from allied blocs.

In the second model, we replace the previous dummy with two categorical variables, dividing allies into moderates and other allies. ⁸⁹ The findings from this second model reveal that moderate allies are less likely to be targeted than other allies. In other words, business groups appear to seek out the traditional right-wing parties more than more moderate rightist legislators, and labor and professional groups appear less likely to go after the more moderate Christian Democratic legislators than the rest of the left bloc. This finding aligns with Hall and Deardorff’s model of legislative subsidies, which saw organized interests as prone to service primarily staunch allies. ⁹⁰ It also seems incongruent with Awad’s argument, which expected moderate allies to be more likely to be targeted than more extreme ones because they are likely to be better intermediaries (i.e., more successful at persuading potential opponents). ⁹¹

Lastly, the online appendix presents the outcomes of four negative binomial regressions that separate the number of meetings by interest type, employing distinct dependent variables instead of combining these data. The results confirm the findings that interest groups meet significantly more frequently with allies than foes (H1) and that legislators occupying a more central position in the chamber’s social network are targeted more often than those in peripheral positions (H4). Concerning membership in key committees, being part of the Finance Committee significantly increases the frequency of meetings with both business groups and labor. Membership in the committees overseeing the country’s natural resources significantly increases the frequency of meetings with both business groups, while serving in the labor or public works committees significantly increases the frequency of meetings with labor and the other business category. Where the coefficients indicating membership in these key committees do not reach statistical significance, they remain consistently positive. Moreover, tests confirm the joint significance of the coefficients for key committee membership across these four models, corroborating the findings of the models presented in Table 1 and supporting the third hypothesis (H3).

Conclusions

With a novel analysis of direct lobbying efforts in Chile, this paper contributes to the ongoing debate regarding the types of legislators targeted by interest groups. The legal mandate for legislators to record lobbying meetings has facilitated the construction of a comprehensive dataset, enabling the testing of various hypotheses. Evaluating whether insights derived from the US context—the case from which several theoretical propositions were initially induced—hold true in other countries that meet the models’ scope conditions is important to gauge the robustness of these theories.

In Chile, as in other countries, contacts with legislators entail information sharing. Interest groups convey private information, sometimes about the policy implications of legislative decisions for themselves or specific constituencies, and other times sharing expert assessments. The empirical analysis concentrated on the targets of business and labor groups. The results showed that allies were the primary lobbying targets. They revealed that interest groups predominantly target ideological allies, whether defined by individual positions or partisan affiliation. This finding underscores the importance of ideological alignment in understanding the dynamics of legislative lobbying undertaken by business and labor groups.

Regarding the targets of organized interests, the results are consistent with informational theories developed to explain lobbying activities in the US Congress, but diverge from the implications derived from those characterizing lobbying as a form of exchange. They are also inconsistent with the notion that organized interests equally target allies and adversaries.

The findings resonate with Hall and Deardoff’s theory of legislative subsidies in that organized interests predominantly target their strongest allies rather than moderates, who could be perceived as more persuadable. However, their expectation that foes will not be lobbied is not met. The evidence from Chile suggests that persuasion is also a strategic goal, aligning with the informational arguments presented by Awad and Ellis and Groll. ⁹² In short, the analysis of Chile offers a distinct perspective that allows us to discern between competing theories previously developed from studying the US case. This comparative approach illuminates which theoretical expectations hold in a different presidential democracy where lobbying is prevalent and which do not.

The analysis also shows that more influential legislators—those in key committees and centrally positioned in bill collaboration networks—are more likely to be targeted by interest groups. This result underscores the strategic value of committee assignments and social connections in the legislative arena.

Lastly, the analysis presented in this article suggests some fruitful avenues for future research. For instance, new studies, such as Junk et al., highlighted the significant underrepresentation of women among lobbyists. ⁹³ Our findings complement theirs, suggesting that interest groups are less likely to target women legislators. The generalizability of this observation and its broader implications warrant further investigation. Exploring the lobbying strategies of business and labor in presidential democracies where parties lack clearly identifiable ideological stances and well-established links with organized interests presents another valuable direction for future research. Several Latin American countries fit these characteristics. Investigating these contexts could reveal decisive factors when partisan or ideological cues offer limited guidance, furthering our understanding of the strategic approaches pursued by business and labor.