Bridging the Information Gap: Legislative Member Organizations as Social Networks in the United States and the European Union

Citation preview

Page ix →

ACKNOWLEDGMENTS The idea for this book grew out of our mutual interest in social networks in legislative politics and legislative member organizations (LMOs) as often-overlooked features of the institutional structure of many lawmaking bodies. What started out as an idea for a comparative paper on LMOs quickly became a much greater undertaking, largely as a consequence of the relative dearth of previous research on LMOs. With greater ambition came the need for more data, a greater variety of methodological tools, and careful theorizing. In the process of writing this book, we have become indebted to a great many people. Our sincerest thanks go to all of our respondents in the European Parliament and the U.S. Congress, without whom none of this research would have been possible. We are also grateful to all respondents to our expert survey, who provided the data used in the analysis in chapter 3. Christopher Carman made contributions to this project primarily in its formative stages by helping to conceptualize theory and research design. In addition, he conducted interviews in Brussels and Washington, D.C.; helped to code the qualitative data; and engaged in intercoder reliability checks. He also contributed to the quantitative analyses in chapter 3 (table 3.3) and provided comments on parts of the chapter drafts. We are grateful for the support this project has received from the European Union Center of Excellence at the University of Wisconsin, the European Union Center of Excellence at the University of Pittsburgh, the Steven D. Manners Faculty Development Award via the University Center for Social and Urban Research at the University of Pittsburgh, and the Central Research Development Fund via the Office of Research at the University of Pittsburgh. We thank all participants in a book manuscript workshop held at the University of Pittsburgh on February 25, 2011. We are particularly indebted to our discussants, Marie Hojnacki, John Carey, and John Scholz, who provided invaluable insights and suggestions. We are also grateful to Karen Page x → Lautanen, who helped to organize the conference. We greatly appreciate the input we received from the participants in the Comparative Politics Colloquium in the Political Science Department at the University of Wisconsin–Madison on April 21, 2010, and the Comparative Politics Workshop at Yale University on November 30, 2010. We offer particular thanks to Barry Ames, Sanja Badanjak, David Canon, Skyler Cranmer, Logan Dancey, Christine Degregorio, James Fowler, Scott Gehlbach, Elisabeth Gerber, Justin Gross, Sarah Halpern-Meekin, Kyle Hanniman, Michael Heaney, Yoi Herrera, Jonathan Hurwitz, David Judge, Michael Kaeding, Kristin Kanthak, Justin Kirkland, David Lazer, Christine Mahoney, Melanie Manion, Scott Morgenstern, and Hans Noel for substantive comments and suggestions that have greatly improved this final product. We also thank Mary Stafsholt for her help with our network figures in chapter 2 and Conrad Nicoll for creating the “Where in the World Are LMOs?” map in chapter 3. We have relied on a number of research assistants throughout the data collection process and thank them for their diligence and hard work. At the University of Wisconsin, we are grateful to Sharmila Gosh, Katrina Helmer, Jordan Kook, Luke Petrovich, and Maysen Yen. At the University of Pittsburgh, we thank Kristen Coopie Allen, Anthony Brino, Ian Cook, Samuel Covert, Patrick Dalsass, Daniel Gore, Zach Gozlan, Bruno Heopers, Matthew Higgins, Benjamin Melusky, Brandon Myers, Dillon Narry, Emily Ordway, Suzanna Ploszaj, and Marika Stettner. We also thank Conrad Nicoll for collecting the caucus membership data for the 103rd Congress and Justin Kirkland for supplying cosponsorship data and advice about analysis. Senthil Natarajan at the University of Pittsburgh assisted us in operating the PittGrid virtual supercomputer and advised us about programming. Furthermore, we are grateful to Donald Musa, David Halpern, and the staff at the Qualitative Data Analysis Program (QDAP) at the University Center for Social and Urban Research at the University of Pittsburgh who

transcribed our interviews. For help with our expert survey and information on LMOs in particular national legislatures we thank Aslaug Asgeirsdottir, Sanja Badanjak, Hannah Britton, Mark Di Virgilio, Katja Favretto, Ulrich Hüschen, Jason Koepke, Kostas Kourtikakis, Lisa London, Ewa Mahr, Tricia Olsen, Marios A. Panayides, Michael Schatzberg, Evaldas Sinkevicius, Margit Tavits, Aili Tripp, Kristin Vekasi, and Emilia Zankina. Page xi → We are also grateful to Lyndsay Monaco, who provided Jennifer with much-needed assistance on the home front that allowed her to spend so much time working on this project. We appreciate the support and important substantive input from two anonymous reviewers. We also thank a number of staff at the University of Michigan Press, especially Melody Herr, who has simply been tremendous as our editor. Last but certainly not least, we thank our spouses, Sarah Halpern-Meekin and Robert Victor, without whose understanding, support, and creative ideas this project would not have been possible. We also are grateful to Julia Victor, Adam Victor, and Scrappy for their unconditional love and support. Page xii →

Page 1 →

CHAPTER 1 Bridging the Information Gap: The Social and Political Power of Legislative Member Organizations In the aftermath of the September 11, 2001, terrorist attacks in the United States, a number of lawmakers and commentators put forward a short-lived and ultimately unsuccessful proposal that the U.S. Congress have the option to do its work remotely—a “virtual Congress,” they called it (R. Cohen 2002). The proposal was fueled in part by widespread speculation that the target of one of the hijacked airplanes might have been the White House or the U.S. Capitol; however, a number of members of Congress raised immediate and strong opposition to the suggestion, objecting to the loss of human interaction. Congressman David Dreier (R-CA), then chair of the House Rules Committee, wrote, As an organization, Congress functions in large part because of the regular and personal interactions among Members as they work to build consensus on issues ranging from procedural matters to the budget and appropriations legislation. (2001, 8) Dreier and others recognized the great value in the personal relationships among lawmakers. However, lawmaking has changed dramatically over the past 50 years. Across different legislatures and parliaments, lawmakers find they have fewer opportunities to interact personally with their colleagues because of busy travel schedules and because face-to-face discussions are being replaced by electronic exchanges using smart phones and Internetbased communication. In the United States, for example, many members of Congress are part of the “Tuesday to Thursday” club—that is, they tend to be in Washington for only a few days midweek and to spend the remainder Page 2 → of their time in their districts (Mann and Ornstein 2006, 169). Such developments have raised concerns about a decline in civil interaction in legislative politics (Uslaner 1993) because “personal relationships, face-toface negotiations, building of trust and reciprocity in human behavior on the Hill necessitates being together in one room” (James Thurber quoted in Keller 2001, A1). While personal connections and contacts are often significant in determining the outcome of political events, the creation and maintenance of such relationships and social networks and the specific benefits they provide are not well understood. Lawmakers are notoriously busy people who are always pressed for time and whose attention is constantly sought. They also face countless collective action and coordination problems in their pursuit of policy, power, and electoral victories. What mechanisms do lawmakers have to help overcome these collective action and coordination problems and develop useful networks with fellow lawmakers and relevant outside actors? The obvious answers discussed in the literature are institutional leadership, party organizations, legislative committees, and the seniority system, among other institutions (for overviews, see Shepsle and Bonchek 1997; Stewart 2001). But do these institutions sufficiently satisfy lawmakers’ need to build and maintain the relationships that are imperative to the lawmaking process? In this volume, we argue that legislative member organizations (LMOs) are often-overlooked institutions that help to connect lawmakers in a loose web of relationships that enable vital information to flow efficiently through lawmaking bodies. These LMOs serve a utilitarian purpose, are common throughout the world, and are woefully understudied.

What Is a Legislative Member Organization? All decision-making bodies face a variety of coordination dilemmas. Scholars have written extensively about the formal institutions that legislatures have developed to help decision makers solve their coordination and collective action dilemmas. Legislative parties help large governing groups coordinate strategies and votes (Cox and McCubbins 1993, 2005). Committee systems promote division of labor and help legislators and parliamentarians develop policy expertise (Krehbiel 1991).1 However, numerous parliaments Page 3 → and legislatures have

created LMOs, an additional infrastructure that appears to provide a mechanism for overcoming collective action problems. In the U.S. Congress, for example, the number of LMOs has proliferated in recent years. Examples of LMOs (known in the United States as congressional member organizations or simply caucuses) include the Congressional Black Caucus, the Biomedical Research Caucus, the Congressional Arts Caucus, the Congressional Caucus for Women's Issues, the Older Americans Caucus, the Congressional Hispanic Caucus, and the Rural Health Care Coalition. LMOs in the European Parliament (EP), where they are called intergroups, have become a regular feature of internal politics and are used by parliamentary assistants and members of the European Parliament (MEPs) who seek to share information and shepherd legislation. Examples are the Sky and Space Intergroup, the Intergroup on Ageing and Intergenerational Solidarity, the New Media Intergroup, the Public Services Intergroup, the Tibet Intergroup, the Social Economy Intergroup, and the Intergroup on Extreme Poverty and Human Rights. The presence of these organizations in many legislatures suggests their utility, despite their undefined role in the legislative process and the uncertain benefits they seem to offer. We surmise that if legislators decide to join LMOs, maintain them, and take part in their activities, these organizations likely offer some benefits that the more formal legislative institutions of parties and committees do not. Our comparative research identifies the roles and benefits of LMOs across a variety of legislatures and lays out a generalizable theory of LMOs.

Relationships and Information: The Role of LMOs in Legislative Politics We define LMOs as voluntary organizations within lawmaking bodies that are made up of members who share a common interest in a particular political issue or theme. These voluntary organizations give lawmakers the opportunity to build utilitarian social networks that facilitate the efficient flow of policy-relevant information.2 As such, LMOs help legislators overcome an institutional collective action dilemma, where the outcomes of collective actions may be suboptimal but a lack of coordination within the formal institutions makes the realization of a more widely preferred outcome Page 4 → less likely (Feiock and Scholz 2010). In the case at hand, the institutional collective action dilemma revolves around the high demand and insufficient supply of information in legislative politics. Legislators require information about policy and politics. They want to know their colleagues’ preferences regarding policy. They want to know their constituents’ preferences regarding policy. They want to know how policy will affect their constituents and whether court action could be taken against a policy. There is virtually no limit to information legislators seek in their efforts to pursue policy and political goals. Since their thirst for information knows no bounds, their desire for good sources of information is likewise never satisfied. Legislators need knowledge and information, and no legislature fully addresses the informational needs of its members. LMOs help alleviate the informational deficit inherent in legislative politics.3 LMOs are not lawmakers’ only sources of information, of course. In particular, legislators can and do rely on parties and committees, the principal formal legislative institutions, to gather information. However, parties and committees suffer from a number of weaknesses when it comes to the creation and diffusion of information. Most important, they compel individual legislators to interact primarily with a relatively small, predetermined set of colleagues, thereby impeding the flow of new, innovative policy ideas. LMOs, in contrast, provide opportunities for members of different social groups to form interpersonal ties (see Brass et al. 2004, 808) because LMOs are composed of heterogeneous groups of legislators from different parties and committees. LMOs thus make up social networks that cut across party lines and committee jurisdictions, making them important venues for deliberation and cooperation and especially for the exchange of policy-relevant information. LMO networks give legislative actors the opportunity to become better informed when a particular topic relating to the LMO's issue or cause rises to prominence on the legislative agenda. LMOs allow both substantive information (i.e., policy-oriented information about the content and expected consequences of legislative proposals) and political information (e.g., strategic information about the policy positions of other decision makers) to be efficiently diffused throughout legislative arenas. This information flow is promoted through LMO-based social networks composed of ties that have two key characteristics: they are

weak as well as bridging. Sociologist Mark Granovetter's (1973, 1974) seminal work on the “strength of weak ties” demonstrates theoretically and empirically how Page 5 → effective social coordination takes place in networks with many weakly tied actors rather than actors with many strong ties. The intuition behind the theory is that close friends (those to whom one is connected through strong ties) are likely to have contacts who share characteristics and therefore access to the same information. In contrast, acquaintances (those with whom one shares weak ties) are more likely to have access to information that would otherwise be outside of one's reach. Empirically, Granovetter (1974) demonstrates that job seekers with many weak ties are more likely to receive information about available jobs and are more likely to secure better jobs (in terms of salary and job satisfaction) than those with mostly strong ties. LMO networks are composed primarily of weak ties that bridge structural holes in the legislative network (Burt 1992, 2000, 2004). While many legislators will, of course, have a variety of strong ties to other legislators with whom they can exchange information, we should expect information in those tightly knit networks to often be redundant. In contrast, legislators who can access information from colleagues who are unlike themselves (e.g., from a different party, from a different region, from a different committee, or from the other end of the ideological spectrum) are more likely to have access to valuable information that they would not otherwise receive. Therefore, social institutions that allow legislators to create weak, bridging connections have great informational value. The more weak connections we observe and the more cross-cutting these connections are in joining otherwise disconnected parts of the network, the more efficiently information should flow through a network of lawmakers. This process not only makes individual lawmakers more informed but also increases the overall level of informedness and expertise of the legislature as a whole (Burt 2000; Carpenter, Esterling, and Lazer 1998). That LMOs are voluntary institutions contributes to their weakness and bridging qualities. Legislators are not assigned to LMOs and will not suffer direct negative consequences if they decide to stay outside the LMO system. Consequently, legislators, as exceedingly busy people, are unlikely to spend as much time on LMO activities as on their formal responsibilities in their parties and committees, are not required to interact with their LMO colleagues on a regular basis, and are not compelled to expend time and resources on fostering LMO ties. In this sense, voluntariness contributes to the weakness of most LMO ties.4 The voluntary nature of LMOs also means that lawmakers are free to join and participate based on personal interests Page 6 → and policy priorities rather than ideological preferences or committee jurisdictions.5 This encourages the creation of ties that cut across tightly knit social circles. Hence, voluntary participation contributes to the “bridgingness” of LMO ties—that is, the extent to which LMO ties span structural holes in the legislative network. It is of crucial importance that the costs associated with participation in LMOs are low, because voluntary organizations designed to help overcome institutional collective action dilemmas will only succeed if the transaction costs associated with creating and maintaining them are low. After all, individuals would not choose to join voluntary organizations if participation was exceedingly costly and made them worse off (Feiock and Scholz 2010, 12). The voluntary nature of LMOs has additional advantages. They provide an open issue space because they can be established to cover any topic or cause a group of legislators considers important enough to warrant attention. Furthermore, once the groups are established, leaders and members are free to determine the activities in which LMOs engage, meaning that their activities can readily be adjusted to meet participants’ needs. In this sense, LMOs are more flexible than either parties or committees. Second, as voluntary organizations, LMOs do not play a formal role in the legislative process. For this reason, they do not threaten the formal institutional framework based on legislative parties and committees. LMOs thus exist parallel to parties and committees and provide informational benefits alongside formal legislative institutions. Lawmakers can decide for themselves how much time and effort they are willing to expend on their involvement in LMOs and on maintaining LMO ties, meaning that the transaction costs associated with LMO participation are only as high as individual legislators want them to be. In this sense, LMOs invite their members to free-ride on the

benefits they offer. This raises an important question: Who covers the costs associated with the creation and maintenance of LMOs? These costs are borne in part by individuals or small groups of legislators who have a personal stake in an LMO's cause and who are, for this reason, willing to expend their limited time and resources on running the group. Participation in LMOs is also inexpensive, however, because many LMOs maintain privileged partnerships with particular outside advocates—interest groups, lobbyists, constituent organizations, or corporations—that provide legislative subsidies (Hall and Deardorff 2006) to LMOs by helping recruit members, by providing secretarial Page 7 → support, or by organizing LMO events. Outside advocates thus bear many of the costs of running LMOs. In return, they obtain privileged access to LMO leaders and members, access that has tremendous value to outside advocates, who strive to build long-term, trusted relationships with lawmakers who share the outsiders’ policy priorities. The symbiotic relationships between LMOs and outside advocates are important not only because they help keep legislators’ participation in LMOs cheap but also because they have an important informational component: Outside advocates provide both political and policy information. Moreover, inclusion in LMO-based social networks provides an incentive for outsiders to supply “high-utility information” (by which we mean information that is research-based, reliable, and easily digestible for busy legislators and their staff), for three reasons. First, LMOs give legislators the opportunity to triangulate and verify information they receive from various sources inside and outside the legislature, many of which may be directly involved with the LMO. Second, LMOs create social networks that are internally highly embedded, where everyone is tied to everyone else; this increases the threat that outsiders will suffer reputational costs if they supply low-quality or even misleading information, because news about such transgressions will quickly spread throughout the LMO network. Finally, LMOs grant privileged access to particular outside advocates, and legislators can credibly threaten to revoke this access if the supplied information is redundant, unreliable, or misleading. High-utility information is thus likely to enter the legislature via LMOs, and it is efficiently diffused throughout the weak-tie LMO network. Their voluntary nature entails that LMOs “are what you make of them,”6 complicating efforts to generalize about their utility for individual legislators or the efficacy of different groups. Moreover, LMOs are forums for the exchange of high-utility information rather than arenas of negotiation and bargaining, which means that their impact on legislation is often indirect and diffuse. Indeed, little evidence and few anecdotal examples show that the presence of an LMO helped pass (or defeat) a policy, although our case studies do include some examples of such direct influence. Most of the time, however, LMOs matter because they help legislators make more informed policy decisions by providing access to innovative ideas and policy input, which then make their way into the formal negotiations inside parties and committees or on the floor. In this volume, we demonstrate that although LMOs play an indirect role in influencing policy outcomes, Page 8 → indirect does not equate to insignificant; rather, LMOs play a consequential role in the early stages of policymaking that enriches the information environment in which legislators operate. In the absence of LMOs, legislators have less information about the political and policy-related consequences of their choices. Given the difficulties associated with measuring indirect and diffuse influence, it would be easy to discount or dismiss the role of LMOs in legislative politics. Doing so would be problematic, however, not only because it ignores empirical realities but also because diffuse influence is often exceedingly consequential. By way of analogy, much evidence shows that campaign contributions do not buy the votes of legislators in the U.S. Congress, but observers nonetheless almost universally agree that the activities of lobbyists and interest groups, as the sources of many campaign funds, have some influence in the policy-making process. The voluminous literature exploring the various means by which this influence manifests itself shows that it is important to understand how these external actors affect policy making. The same can be said for LMOs—they may only indirectly affect the outcome of a piece of legislation, and it may be difficult to tie LMO activities to particular legislative outcomes, but it is difficult not to recognize that they play an important role in the legislative process. In summary, LMOs are voluntary institutions inside legislatures that help lawmakers overcome informational collective action dilemmas by encouraging the establishment and maintenance of social relationships between interested actors inside and outside the legislative arena who share a common policy priority. These relationships make up extensive social networks composed, primarily, of weak ties that cut across party and committee lines,

thus facilitating the diffusion of high-utility policy and political information that allows lawmakers to make better policy choices.

What We Know about LMOs Previous scholarly research on LMOs has concentrated nearly exclusively on the U.S. Congress. This body of research has, for example, richly described the history of the caucus system, why members of Congress join caucuses, the purposes and functions they serve, and how they interact with other branches of government (e.g., Hammond 1998).7 We also learn from these studies that the membership of caucuses is ideologically diverse Page 9 → and may counterbalance power from the committee system (Ainsworth and Akins 1997; but see Victor and Ringe 2009). Caucuses serve a number of important functions in lawmaking, including helping to establish and maintain government and public agendas (Hammond, Mulhollan, and Stevens 1985), collecting information (Fiellin 1962; Stevens, Miller, and Mann 1974; Stevens, Mulhollan, and Rundquist 1981; Hammond 1998; Hammond, Mulhollan, and Stevens 1985), both challenging and complementing party leadership (Hammond 1991), providing voting cues to members (Levy and Stoudinger 1976), promoting substantive representation without loss to descriptive representation (Canon 1995), and affording a venue for the expression of policy preferences (McCormick and Mitchell 2007). LMOs also create increased opportunities for advocacy and influencing policy, as Bullock and Loomis (1985) and Carman (2005) emphasize for the U.S. Congress and the Scottish Parliament, respectively, and provide access points for outside actors (for the U.S. Congress, see Ainsworth 1997, 523; McCormick and Mitchell 2007; for Westminster, see Judge 1981, 141–42; for the EP, see Butler 2008, 575–79). Finally, LMOs aid representation for constituencies (Loomis 1981; Stevens, Mullhollan, and Rundquist 1981; R. Singh 1996; Miler 2011) and help promote the political and electoral needs of lawmakers (Davidson and Oleszek 1981; Ripley 1983; Caldwell 1989). Overall, we may group the potential roles of LMOs in the legislative process into three categories: information acquisition, provision, and exchange; political coordination (such as agenda setting, search for compromise, and building policy coalitions); and signaling (of policy priorities to constituents). We find and present evidence for all three functions throughout this book, yet we focus primarily on the informational function. This emphasis results from three factors and does not deny or discount the variety of functions LMOs may serve in a given legislature. First, while LMOs’ coordination and signaling functions are more or less prominent in different legislatures, for reasons we explain, their informational role is a universal property: LMOs allow for the efficient exchange of information through the social networks they comprise. Second, the other roles ascribed to LMOs in much of the existing literature—in particular, their coordination function—depend on their capacity to connect political actors who share common policy interests and to allow for the efficient flow of information between those actors. Policy coordination thus presupposes the exchange of both substantive and political information. Finally, our empirical results (in particular, our extensive qualitative data) overwhelmingly support the Page 10 → proposition that LMOs’ major benefit lies in their capacity to provide and diffuse policy-relevant information among legislative actors. Ascribing a secondary role to the signaling and coordination function is, therefore, not simply an a priori theoretical decision; it reflects empirical realities on the ground. These considerations warrant our primary focus on LMOs as information networks. In a first attempt to link the study of social connections in legislative politics with the investigation of the role and structure of LMO-based social networks, Victor and Ringe (2009) argue that the U.S. congressional caucus system provides legislators with the opportunity to establish and maintain social relationships with colleagues who might help them advance their position within the legislature. Building on the insight that not all relationships are created equal and that being associated with some colleagues is more valuable to individual members than others, the authors maintain that legislators engage in the caucus system in an effort to maximize the social utility of their relationships. Members of Congress achieve this goal by associating themselves with those actors who are already powerful within the formal institutional structure because being connected to powerful colleagues such as party or committee leaders or senior legislators is more valuable than being linked to just another rank-and-file member. Accordingly, the authors hypothesize that the caucus system does not, as much of the previous literature on

congressional caucuses maintains, serve as an alternative institutional structure used by formally disadvantaged legislators to counterbalance their structural weaknesses; instead, the caucus system replicates and reinforces the existing distribution of power and influence through the more formal institutions of parties and committees. Unlike previous studies, Victor and Ringe do not focus their analysis on a subset of caucuses but investigate claims using membership information on the entire population of caucuses in the 110th Congress. Their social networks analysis confirms the expectation that formally powerful players such as legislative leaders, senior members, and legislators who are electorally safe are both more connected and more central within the caucus network. The arguments in this book are distinct from Victor and Ringe (2009) in three important ways. First, Victor and Ringe focus on the structural positions of different legislative actors in the caucus network of the U.S. Congress, while we are interested in the ties that make up LMO networks and how the strength of those ties facilitates information exchange. Second, Victor and Ringe focus narrowly on one reason why legislators may join Page 11 → caucuses; they do not consider why legislators may choose to be active in LMOs or test broader hypotheses about the purpose and benefits of the LMOs and the social networks they provide. Finally, Victor and Ringe focus on a single legislature, while our argument and our empirical analyses are explicitly comparative, thus moving us beyond the U.S. Congress. This study therefore makes a significant contribution to the literature on legislative politics as the first comparative analysis of LMOs as voluntary organizations within lawmaking institutions. It does not concentrate on parties and committees, a frequent subject of legislative research, but examines how other arenas of political activity affect legislative processes and outcomes. Moreover, our comparative focus provides new information on an understudied empirical phenomenon across a broad sample of legislatures, especially the two lawmaking bodies we examine in detail. We also add to existing research on the access and influence of outside advocates, such as interest groups and lobbyists, in the legislative decision-making process in that we maintain that LMOs provide an institutionalized means by which external actors can develop relationships with lawmakers and examine how LMOs as points of access enhance the quality of the information that enters the legislative discourse. Moreover, our claim that LMOs provide a means for legislators to establish ties across institutional divides, such as party lines, has contemporary relevance in the U.S. Congress, which has become highly polarized. LMOs offer a means by which legislators can become or remain connected to legislators from the other party, even if only weakly so. The possibility that LMOs serve as an antidote to congressional partisan polarization is a practical implication to which we return in chapter 7. Second, we contribute to the literature on social networks among elite actors, as our argument is explicitly relational and emphasizes the importance of social connections in legislative politics. One important aspect of this focus is our emphasis on LMO ties as cutting across tightly knit social groups composed of either like-minded political actors (e.g., members of the same party) or those who share a common, narrowly defined substantive focus (e.g., members of the same committee). This bridging ability of LMO ties is critical in a legislative context, where homophily—or the tendency of individuals to associate themselves with similar others—implies a distinct strategic disadvantage for political operators, who require extensive information on the positions, strategies, and goals of their colleagues, whether friend or foe. Our research provides an ideal setting in which to Page 12 → utilize social network methods in a way that not only offers insights into the role of LMOs in legislative politics but also highlights an appropriate application of these methods to new and important questions. Third, we contribute to the study of political institutions by focusing on the interaction between formal institutions (i.e., parties and committees) and less formalized, voluntary institutions (i.e., LMOs). As voluntary organizations that play no official role in the legislative process, LMOs help overcome an important institutional collective action dilemma in legislative politics—the great demand and insufficient supply of information. Indeed, LMOs help address informational constraints (inadvertently) imposed by formal legislative institutions. Legislators require a great deal of both political and policy information to make good policy choices, but some of the basic features of both parties and committees stand in the way of efficient information exchange. Because of their voluntary and low-cost nature, LMOs are an inexpensive but effective means of mitigating some of the informational constraints legislators commonly face.

A Comparative Research Design To test our expectations about the existence and roles of LMOs in legislatures, we rely on a comparative research design that has two primary components. First, we collect data on the existence of LMOs in the legislatures of 45 advanced industrial democracies by relying on findings from an expert survey. Second, we provide in-depth case studies of the European Parliament and the U.S. Congress. These cases provide us with complementary and contrasting perspectives that allow us to test the robustness of our theoretical claims. While the two legislatures share similar institutional structures, such as strong committee systems and party-based politics, they vary significantly in the number of parties at play, the role of party leadership in the legislative process, floor procedures, the electoral systems by which members are chosen, and the resources available to legislators in pursing their policy objectives. Our cases allow us to model similarities and differences across the two legislatures and thus to reap the benefits of systematic comparison. We can therefore draw specific expectations about how LMO systems should differ across these cases. For example, given the nature of the U.S. single-member district plurality electoral system, which provides incentives for legislators to cater to the preferences of particular Page 13 → constituencies, we expect Congress to have more LMOs than the EP, for individual U.S. legislators to join more LMOs, and for signaling to be relatively more prevalent in Congress. Our comparative research design also offers intriguing findings about systematic differences between our cases. While respondents in both the EP and Congress highlight the importance and value of cross-partisanship in LMOs, for example, we find the nature of cross-partisanship to differ between the two chambers. In the EP, where multiparty policy coalitions are needed for legislation to pass, intergroups serve as arenas for contestation, bargaining, political coordination, and the exchange of political information across party groups. Crosspartisanship in intergroups is, in other words, explicitly political. In Congress, in contrast, the majority party does not require support from across the aisle to pass legislation (at least in the House of Representatives). In this context, many caucuses seek to protect their bipartisan nature by consciously eschewing divisive issues, making LMOs almost apolitical forums for the exchange of substantive information. Our case studies rely on three types of data. First, we collect complete LMO membership information for at least six years (two or three legislative terms) for each legislature. Analyzed using quantitative methods, these data allow us to predict the factors that determine membership in LMOs. Second, we use this membership information as relational data that are analyzed in a social network analysis framework. This approach allows us to investigate the structure of LMO networks, to establish which legislative actors hold influential positions within the LMO network, and to determine how LMOs relate to more formal legislative institutions such as parties and committees. Social network analysis constitutes a critical part of our study, since our argument emphasizes the importance of LMOs in allowing legislators to establish social connections with colleagues who share their policy priorities and in facilitating information exchange throughout the legislature. Social network analysis has been a mainstay in sociology for many years and in recent decades has become more frequently used in economics, anthropology, physics, computer science, political science, policy analysis, and other fields. Social network analysis focuses on investigating social relationships (ties) between individuals (nodes) and how these connections affect social and political interactions, processes, and outcomes. The basic unit of analysis is thus not individuals but dyads (pairs) of individual actors. As such, social network analysis seeks to identify and demonstrate “aspects of social Page 14 → organization that are not captured by the study of individual attributes or characteristics” (Ward, Stovel, and Sacks 2011, 246). The study of social relations among political actors is not new to political science (e.g., Lazarsfeld, Berelson, and Gaudet 1944; Berelson, Lazarsfeld, and McPhee 1954; Patterson 1959; Galtung 1971; Wallerstein 1974; T. Baumgartner and Burns 1975; Knoke 1976; Laumann and Pappi 1976), but the approach has seen a recent resurgence.8 Until the turn of the century, social network analysis in the social sciences was dominated by sociologists, with Robert Huckfeldt standing out as an early proponent of formal network analysis in the political science field (see Huckfeldt 1979, 1983; Huckfeldt and Sprague 1987; Huckfeldt et al. 1995). Since the late 1970s, Huckfeldt has been investigating the relationships between social networks and political change, vote

choice, party loyalty, race and class politics, participation, and communication. Also notable as earlier examples of the application of a formal network approach are Laumann and Knoke (1987) and Heinz et al. (1990), both of which are studies of organizational networks in the context of interest groups and lobbying. Not until the late 1990s, however, did social network analysis gain an increasingly prominent position in political science (see especially Watts and Strogatz 1998; Barabasi and Albert 1999; but also Caldeira and Patterson 1987). A plethora of social network studies have subsequently taken place across a variety of areas of study, including political communication and participation (e.g., Lake and Huckfeldt 1998; Huckfeldt 2001, 2007; Huckfeldt, Johnson, and Sprague 2004; Mutz 2002a, 2002b, 2006; McClurg 2003, 2006; Klofstad 2011), elections and vote choice (e.g., Houghton 2000; Fowler 2006b; Gimpel, Lee, and Pearson-Werkowitz 2008; Nickerson 2008; Rolfe 2012; Sinclair 2012), party networks (e.g., Heaney and Rojas 2007; Koger, Masket, and Noel 2009), legal precedent and court decisions (e.g., Fowler et al. 2007; Fowler and Jeon 2008), interest group networks (e.g., Carpenter, Esterling, and Lazer 1998, 2003, 2004; Heaney 2006; Grossmann and Dominguez 2009), networks between policy actors (Schneider et al. 2003; Scholz and Wang 2006; Scholz, Berado, and Kile 2008), international relations (e.g., Lazer 1999; Maoz 2010), and legislative politics (e.g., Fowler 2006a, 2006c; Masket 2008; Victor and Ringe 2009; Cho and Fowler 2010; Bratton and Rouse 2011; Kirkland 2011; Ringe, Victor, and Gross 2013). The field has also seen important methodological advances, especially in dealing with missing data in network sampling and in addressing the problem of nonindependence of observations in network data (e.g., Frank 2005; Page 15 → Gile and Handcock 2006; Handcock and Gile 2007; Cranmer and Desmarais 2011; Ringe, Victor, and Gross 2013). Social network theory, we believe, has great appeal when applied to political questions. It is natural to seek to explain politics in terms of the relationships between individuals, and doing so builds organically from game theory, which frequently focuses on the interdependent behavior of political actors (e.g., Knoke and Yang 2008; Jackson 2010). Despite this intuitive appeal, however, advances in technology have only recently made the application of network methods in political science more commonplace. Computers and software have advanced to the point that even laypeople have the power to generate and estimate large networks on home computers. Since network datasets tend to be very large and sampling is not typical or even appropriate in a network setting, the combination of parametric statistics and network data requires vast computer memory and sophisticated software, which have become more readily available and accessible. The third type of data we rely on is qualitative. We conducted 86 in-depth interviews with legislators, members of their staffs, and representatives of outside interest groups to examine the activities of LMOs, how they operate, and how they relate to actors outside of the legislative arena. Given the dearth of previous research on LMOs, especially in a comparative perspective, these interview data are of special importance because they provide firsthand accounts of the roles that LMOs play in legislative processes and how their benefits have contributed to legislative outcomes. In sum, this mixed-method approach provides for significant analytical purchase on the question of why legislatures create LMOs, why members join and choose to be active in these organizations, and what role LMOs play in legislative politics.

Plan of the Book Chapter 2 more fully develops our argument, making the theoretical case for the relational and informational benefits we ascribe to LMOs. In chapter 3, we present the results of an expert survey we conducted to determine where LMOs exist across a sample of 45 advanced industrial democracies. We find LMOs in 25 countries, provide descriptive information about them, and use statistical methods to determine what factors predict the existence of LMOs. We conclude that the electoral system, the number of Page 16 → legislative parties, and the size of the legislature are the principal determinants of the presence or absence of LMOs in a national legislature. Having broadly set the stage, we narrow our focus to LMOs in two legislatures to test our theoretical propositions about the role of LMOs for the establishment of social relationships that cut across party and committee lines and

about the informational function of LMOs. Specifically, we investigate intergroups in the European Parliament and caucuses in the U.S. Congress. Chapter 4 introduces the case studies, outlines our research design, and describes the data on which we rely to test our theoretical propositions. We then trace the development of intergroups in the EP and provide basic descriptive information about these organizations, which have thus far been neglected in the growing literature on EP politics. The chapter also provides similar information for the U.S. Congress, whose caucuses have received more attention in previous research. Unlike existing studies, however, we differentiate between single-party groups that resemble LMOs (and, in some cases, are confusingly also called caucuses), such as the Blue Dog Democrats or the Republican Study Committee, and the bipartisan, voluntary organizations that are the focus of this book, such as the Congressional Human Rights Caucus, the Congressional Coalition on Adoption, or the Congressional Steel Caucus. The former, while important and worthy of scholarly attention, operate more like party factions or voting coalitions and thus fall outside the scope of our theoretical approach to LMOs. Finally, chapter 4 examines what factors drive legislators to join LMOs in both EP and Congress, relying on both qualitative and quantitative methods and data. Chapter 5 focuses on the relational functions of LMOs. We use our interview data to provide evidence for our proposition that LMOs allow legislators to build social relationships with one another. The chapter then moves on to a careful examination of LMO network structures using a variety of social network analysis tools. These analyses confirm one of our key theoretical propositions: LMO ties are bridging ties that connect legislators who would not otherwise be connected to each other. Hence, the structure of LMO networks is such that it ought to facilitate the flow of policy-relevant information throughout the legislature. Having made this case, chapter 6 investigates whether this potential for efficient information flow is realized and confirms that LMOs are important arenas for the exchange of both policy and political information. Information exchange takes place both inside the legislature—between LMO Page 17 → members and their offices—and between insiders and outside advocates associated with particular LMOs. These outside organizations supply legislative subsidies to LMOs and their members by providing policy-relevant information and by bearing many of the costs associated with creating and running LMOs. Finally, chapter 7 reflects on how exactly LMOs might matter in legislative politics. We consider LMOs’ potential direct and indirect effects on legislative processes and outcomes. Having provided evidence from both the EP and Congress that demonstrates how LMOs sometimes influence legislative processes and outcomes directly, however, we conclude that most of the time, the impact of LMOs is indirect and diffuse. LMOs influence the legislative process during its early stages, when legislators are gathering information and communicating with the various stakeholders who share a common interest in an issue or cause. LMOs affect discourse, attention, and priorities and help disseminate otherwise unavailable but policy-relevant information through social networks composed of political actors who share common policy priorities. This is how LMOs matter and why legislators choose to expend valuable time and resources on them.

Page 18 →

CHAPTER 2 Solutions to Informational Collective Action Dilemmas: Theorizing the Benefits of Legislative Member Organizations Across the clusters in an organization or market, creativity is a diffusion process of repeated discovery in which a good idea is carried across structural holes to be discovered in one cluster of people, rediscovered in another, then rediscovered in still others—and each discovery is no less an experience of creativity for people encountering the good idea. Thus, value accumulates as an idea moves through the social structure; each transmission from one group to another has the potential to add value. (Burt 2004, 389) In this chapter we lay out our argument about the role and value of legislative member organizations (LMOs) in legislative politics. We propose that LMOs allow legislators to build relationships with each other and that the social networks composed of these relationships provide for an efficient exchange of high-quality information inside the chamber. LMOs thus help mitigate what we identify to be an important institutional collective action dilemma (Feiock and Scholz 2010): the great demand and inadequate supply of policy-relevant information in legislative politics. The voluntary nature of LMOs plays a crucial role in this regard. It entails a number of advantages over the formal legislative institutions of parties and committees. First, LMO participation is cheap; lawmakers can determine their own level of involvement without threat of sanction. Hence, they can reap the informational benefits of LMOs while incurring only those transaction costs considered appropriate. Second, as voluntary institutions without a formal role in the legislative process, LMOs do not threaten the stability of the formal institutional framework. Accordingly, parties and committees have little incentive to compete with LMOs for Page 19 → legislators’ time and efforts or to undermine the LMO system as a whole. Rather, parties and committees can reap the benefits that LMOs provide. Third, LMOs are flexible institutions that can readily be adjusted to meet their members’ needs, primarily because the issue scope of LMOs is open: LMOs can be set up to address any cause or issue lawmakers consider a policy priority. Finally, voluntary membership in LMOs entails that they are heterogeneous institutions composed of members from multiple parties and committees who share common policy priorities and substantive interests. Ties between LMO members, therefore, combine to form extensive social networks that cut across party lines and committee jurisdictions. These social networks offer important opportunities for discussion, collaboration, and cosponsorship but most importantly for the exchange of policy-relevant information. Building on classic insights from social network analysis, we emphasize the importance of weak ties between legislators that span structural holes in the legislative network for the diffusion of information throughout the legislature. According to Granovetter (1973), individuals benefit more from being affiliated with acquaintances (through weak ties) than with close friends (through strong ties), as the closely tied social circles that are prevalent in most social contexts are connected through weak ties. Without such weak ties, these social circles would be isolated from one another, and information would not be diffused efficiently beyond tightly knit social groups. Weak, cross-cutting ties therefore provide individuals with access to otherwise unattainable information and make all members of the network better informed (Burt 1992, 2000, 2004). Yet weak ties are significant not only because they tend to be bridging. They also contribute to the low transaction costs associated with LMO participation since weak ties are inexpensive to create and maintain. Another factor that contributes to the low-cost nature of LMOs is that they are often associated with actors and organizations outside of the legislature itself. These outside organizations carry many of the costs associated with running LMOs, thus allowing legislators to free-ride on their benefits. Ties with outsiders also have an informational component, because interest groups provide both substantive and political information to lawmakers, and LMOs provide an incentive structure that increases the quality and utility of this information: Outsiders associated with

particular LMOs enjoy institutionalized, privileged access to groups of legislators who share their concern for a given issue or cause, which lawmakers can threaten to revoke if Page 20 → the information provided is not of sufficient quality to help them make more informed policy choices. In sum, LMOs are inexpensive, voluntary institutions that encourage the inflow of high-utility information from outside the legislature and facilitate the efficient diffusion of this information across social and institutional boundaries inside the chamber. These features enable lawmakers to overcome informational constraints in their legislative activities.

Previous Research on Legislative Member Organizations The existing literature on LMOs is limited, focuses primarily on the congressional caucus system, and includes little work conducted over the past decade (e.g., Fiellin 1962; Loomis 1981; Stevens, Mullhollan, and Rundquist 1981; Hammond, Mulhollan, and Stevens 1983, 1985; Hammond 1991, 1998; Vega 1993; Ainsworth and Akins 1997; Victor and Ringe 2009; but see Carman 2005 for Scotland; Judge 1981, 1990 on Westminster).1 Broadly, research on LMOs highlights three principal functions: information acquisition and exchange, political coordination, and symbolic and substantive representation. In the most exhaustive investigation of caucuses in the U.S. Congress, for example, Hammond emphasizes that caucuses in the U.S. Congress may be partisan or bipartisan, bicameral or unicameral. They provide information to their members, affect agendas, draft bills and amendments, develop legislative strategy, build supporting coalitions, form voting blocs, and even launch congressional leadership careers. They precipitate, and participate in, battles on the floor…. Their influence may be through research, information exchange, agenda setting, policy formulation, and coalition building. (1998, 7–8) While not ignorant of representational aspects of caucus membership, Hammond focuses primarily on the informational and coordination functions of LMOs. Their representational role is most carefully considered by Miler (2011), who argues that caucus membership is an important means of constituency representation. Miler demonstrates that legislators’ decisions about caucus membership reflect their constituents’ interests; members of Congress thus use caucus membership to signal their awareness of and prioritization of issues relevant to their constituents. Page 21 → In this book, we focus on LMOs’ informational role because it is a more universal function than either the signaling of preferences and priorities or political coordination. Even a cursory look at LMOs in legislatures other than the U.S. Congress suggests that signaling and coordination are likely more important in some lawmaking institutions than others. The Israeli Knesset, for example, has a large number of LMOs (called members’ lobbies, or shdula),2 yet the incentives for Knesset members to engage in signaling of policy priorities to constituents are far less pronounced in Israel's closed-list proportional representation system than in the U.S. Congress, whose members are elected by plurality in single-member districts. Voters do not elect individual Knesset candidates and cannot influence the ranking of candidates on party lists. Similarly, the U.K. Parliament has several hundred allparty groups (APGs),3 yet there is much less need for political coordination in Westminster than in some other legislatures, given the primacy of single-party cabinets, the dominance of the legislative majority party, a pronounced government-opposition dynamic, and high levels of party discipline.4 In this system, political coordination in LMOs—such as agenda setting, the establishment of policy consensus, and coalition building—will be less meaningful and consequential than in legislatures where cross-party cooperation is imperative and policy coalitions are fluid and continuously renegotiated. In contrast, members of all lawmaking institutions have great informational needs that LMOs may help fulfill, implying that LMOs’ informational role is likely more universal than either coordination or signaling.5 This informational value of LMOs has been recognized in existing research on caucuses in the U.S. Congress (Fiellin 1962; Stevens, Miller, and Mann 1974; Stevens, Mulhollan, and Rundquist 1981; Vega 1993; Hammond 1998).

Hammond, for example, finds that 78 percent of the caucuses in the U.S. Congress included in her sample serve as forums for information exchange (1998, 26), while Ainsworth and Akins (1997) argue that caucuses augment Congress's institutional structure by acting as informational counterparts to committees. Despite this recognition, the existing literature falls short in a number of important ways. First, attempts to highlight the various roles that LMOs play in allowing for the coordination of legislative action outside the formal party and committee structure (e.g., Fiellin 1962; Stevens, Miller, and Mann 1974; Loomis 1981; Stevens, Mullhollan, and Rundquist 1981; Hammond, Mulhollan, and Stevens 1983, 1985; Miller 1990; Hammond 1991, 1998; Page 22 → Vega 1993) draw attention away from the importance of LMOs as circuits of information. We argue, however, that the primary and most universal function of LMOs is informational. Second, it is unclear from the existing literature how LMOs fulfill their informational role in legislative politics. We explicitly theorize this informational function and emphasize in particular how LMOs allow legislators and their staff to build social relationships with other legislative actors. The particular structures of the social networks made up of these relationships, in turn, provide for the efficient exchange of policy-relevant information between legislative offices. While some commentators recognize the LMOs’ importance for the establishment and maintenance of social relationships inside legislatures, not much is made of this acknowledgment, especially with regard to the link between LMOs as social networks and the exchange of information. We address this shortcoming by investigating the nature of social relationships based on LMO membership using the most appropriate methodology—that is, social network analysis. Third, while the extant literature acknowledges that LMOs allow the establishment of contacts between actors inside the legislature and lobbyists and interest groups on the outside (Judge 1981, 141–42; Stevens, Mullhollan, and Rundquist 1981; Hammond, Mulhollan, and Stevens 1983, 1985; Ainsworth 1997, 523), the importance of this purpose of LMOs has not been given the emphasis it warrants. Serving as access points for outsiders is among the most important functions fulfilled by LMOs. Indeed, LMOs serve not only as facilitators of information exchange inside the legislature, as commonly suggested, but also as critical venues of information exchange with actors outside of the legislature. Finally, no comparative research has been conducted on LMOs and their functions across different legislatures. While it is illuminating to learn about caucuses in the U.S. Congress or all-party groups in Westminster (Judge 1981), the similarities and differences of LMO structures and activities must be investigated in a cross-national perspective if we want to gain generalizable insight into the role of these groups in legislative politics. This exercise not only allows us to identify important commonalities across cases, such as the more universal role of LMOs as circuits of information exchange, but also to determine under what conditions LMOs fulfill other functions. As our examples illustrate, there are a priori reasons to expect LMOs to serve as signaling devices in some Page 23 → legislatures yet to be more valuable as arenas for political coordination in others. Our comparative research design provides significant analytical leverage in this regard and allows us to empirically test the conditions under which LMOs play the various roles ascribed to them in previous research. The theoretical propositions and methods of social network analysis offer an exceedingly useful framework for our investigation, allowing us to identify the relationships between actors and how these individual-level relationships provide an aggregate-level structure that facilitates information flow in legislative politics. Social network analysis focuses on investigating the connections (ties) that exist between individual actors (nodes), and examines how these connections affect social and political interactions, processes, and outcomes. In a general sense, ties may be based on friendship or kinship, communications, financial relations and exchanges, shared membership in associations, joint participation in events, the exchange of goods, or any other type of relationship that researchers may be able to measure (Heaney and McClurg 2009). The idea that lawmakers’ social networks are a meaningful component of legislative politics is hardly new (e.g., Routt 1938; Eulau 1962; Bogue and Marlaire 1975) and was pursued for many years by Samuel Patterson and his collaborators (Patterson 1959; Caldeira and Patterson 1987, 1988; Caldeira, Clark, and Patterson 1993; L. W. Arnold, Deen, and Patterson 2000; see also Peoples 2008). However, the number of applications of social network

analysis in legislative studies has increased significantly in the more recent past, mostly focusing on the U.S. Congress,6 and has shown interpersonal relationships to be politically consequential. A particularly popular measure of social connectedness has been cosponsorship among legislators (Burkett and Skvoretz 2001; Crisp, Kanthak, and Leijonhufvud 2004; Fowler 2006a, 2006c; Gross and Shalizi 2007; Bräuniger and Brunner 2009; Cho and Fowler 2010; Bratton and Rouse 2011; Kirkland 2011), but numerous other network indicators have also been used, such as committee assignments (Porter et al. 2005, 2007), campaign contributions (Peoples 2010; Victor and Koger 2011), connections between legislative staff (Baughman 2006; Ringe, Victor, and Gross 2013), shared work space and spatial proximity (Masket 2008; L. Cohen and Malloy 2010; Rogowski and Sinclair 2011), and the focus of our study, shared LMO membership (Victor and Ringe 2009). Page 24 →

The Informational Benefits of LMOs: Solving Informational Collective Action Dilemmas in Legislative Politics We see LMOs as a solution to an institutional collective action dilemma, where the outcomes of collective actions may be suboptimal but a lack of coordination within the formal institutions makes the achievement of a better and more widely preferred outcome less likely (Feiock and Scholz 2010). Voluntary institutions, according to Feiock and Scholz, often operate alongside formal ones to solve such institutional collective action problems. In fact, voluntary institutions may be more efficient than formal institutions in solving some collective dilemmas because their decision and transaction costs are low. Collective action dilemmas can be solved by a variety of institutions that could be thought of as existing on a continuum from a consolidated authority, which is highly efficient but also strict and hierarchical, to entirely selforganizing institutions, which are fragmented but also flexible and low-cost (Feiock and Scholz 2010). LMOs are located closer to the self-organizing institutions end of this continuum, but they are not ad hoc, informal networks. While their level of formality varies across legislatures—for example, as a function of the rules governing their existence and the resources available to them (see chapter 3)—LMOs are institutionalized enough that actors inside and outside the legislature can have some expectation of what LMOs do, how often they meet, and what to expect from them. The institutional collective action dilemma that we maintain LMOs help mitigate revolves around the great demand and inadequate supply of policy-relevant information in legislative politics. LMOs alleviate this dilemma by allowing legislators to build relationships with each other and, through these relationships, by providing for the efficient diffusion of information. Both of these functions are recognized in the existing work on LMOs (see especially Fiellin 1962; Stevens, Miller, and Mann 1974; Ainsworth and Akins 1997), but their significance in the literature as a whole is muddled by other roles and activities ascribed to LMOs, most importantly political coordination. Political coordination certainly takes place in some LMOs, yet it is pursued by only a small subset of LMO members (generally the leaders of the LMOs and a small group of particularly active members), while most members benefit primarily from LMOs’ relational and informational functions. Page 25 → Information is a highly valued good in legislative politics. The two principal types of information that matter are political information and policy information.7 Political information has been described as “political intelligence” (Schlozman and Tierney 1986, 299–300; Wright 1996, 82–87) about the positions of key decision makers or potential policy coalitions as well as about constituency preferences. This type of information is important for determining the probability of success of a given policy initiative and to gauge the level of public support. Policy information, in contrast, is about the content and expected consequences of legislation. It is most valuable if it describes carefully researched causal relationships between the content of a policy proposal and its likely implications once implemented (Esterling 2004). Relationships and information flow are very closely associated, of course, since relationships are critical to

exchanging the necessary political and policy information lawmakers require when engaging in their legislative work. Relationships are particularly important because information is a scarce good, and the collection and dissemination of information is costly. Legislators face great transaction costs in gathering, verifying, summarizing, reproducing, and communicating information. No legislature fully meets the informational needs of its members, although there are factors that provide for variation in the extent to which legislators face an informational deficit in their work. Policy choices on legislation tend to be more complex if, for example, they affect a large, diverse number of constituents. Hence, such decisions call for greater expertise to determine their effects after implementation. Some legislators also have more informational resources available through such means as personal staff, committee staff, legislative research services, or informational resources provided by their parties. These resources may be important not only to gain access to or produce information but also to allow legislators to efficiently process what may be a large amount of incoming information. In general, however, there is no such thing as too much information for legislators trying to make fully informed decisions on policies across a wide range of issue areas. The main institutional features of legislative politics, parties and committees, go a long way toward addressing the informational deficit legislators face but do not solve it. In fact, some of their inherent characteristics impede the optimal exchange of policy-relevant information. From an informational point of view, differences among parties, committees, and LMOs relate to rules of participation, social network structures, the sets of participants involved, and issue scope, as illustrated in table 2.1. Page 26 → Rules of participation concern whether lawmakers are obligated to take part in the institution's activities or may choose not to participate. Social network structure can be either exogenous or endogenous from the point of view of individual lawmakers. Network structures are exogenous if they are fixed by factors outside of the control of individual network participants, which is the case if participation in the network is obligatory. They are endogenous if they are the result of choices made by network participants themselves, which happens if participation is voluntary. The set of network participants may be homogeneous or heterogeneous, given the personal attributes of the lawmakers who participate in the institution. Finally, the scope of the issues addressed in a given institutional context is either open, when there are no limits to what issues may be selected for consideration, or closed, if such limits are set a priori. This discussion simplifies reality in two ways for the sake of theoretical clarity. First, we keep the four characteristics that describe the informational utility of parties, committees, and LMOs analytically separate, when in fact they are closely interrelated. For example, obligatory participation in an institution implies that network structures are exogenous from the point of view of individual participants. We only relax these analytic distinctions in our discussion of LMOs below to illustrate how key features of LMOs jointly provide for important informational benefits in legislative politics. Second, we treat the differences between obligatory and voluntary participation, between exogenous and endogenous network structures, between ideologically homogeneous and heterogeneous sets of participants, and between open and closed issue scope as dichotomous. Doing so provides for a clearer analytical discussion, although in the real world of legislative politics, these differences describe tendencies toward one or the other end of a continuum. Parties, for example, are not entirely ideologically homogeneous organizations, the issue scope of committees is not wholly determined by predetermined jurisdictions, and basic rules about membership in LMOs make their network structure at least partly exogenous. Page 27 → We first examine these characteristics in the context of parties and committees, as formal legislative institutions. Participation in both parties and committees is obligatory in the sense that legislators who refuse to engage in these activities will suffer negative professional consequences because their career success depends on participation. In addition, formal rules may even require that lawmakers must serve on a given number of

committees, for example, or take part in certain party activities. Obligatory participation implies that the network structure of parties and committees is exogenous to individual network participants; as long as members take part in party and committee activities, they are socially tied to one another. Indeed, legislative rules may again prescribe the number of committees of which each legislator must be a member, and members may not even have influence over their own committee assignments if their parties control the distribution of committee seats. Parties and committees differ, however, regarding their sets of participants. The set of participants in parties is ideologically homogeneous, since parties are groups of political actors who share a general sense of what political outcomes are desirable and band together on this basis to collectively pursue their goals. In contrast, committee networks are composed of ideologically heterogeneous sets of participants, because different political parties are represented in committee. Issue scope also differs for parties and committees. It is open in the case of parties, since the absence of predefined, formal jurisdictional limits means that parties and their members may choose to take up any issue for consideration. Committees, in contrast, generally have predetermined jurisdictions, so issue scope is closed. What implications do these characteristics have from an informational point of view? Both parties and committees clearly serve as important information providers for individual legislators. Parties employ specialized staff (inside the legislature and/or within the party organization itself), maintain relationships with experts outside of the party organization, and when in government, rely on expertise grounded in ministries, agencies, and other government departments. The creation, processing, and exchange of information are also critical functions of legislative committees; indeed, for some commentators they are the primary raison d'être of legislative committee systems (Krehbiel 1991; see also Fenno 1966, 1973; Cooper 1970; Deering and Smith 1997). Legislative committees allow lawmakers to specialize and to become policy experts in particular fields. A system of committees therefore provides a wealth of high-quality information for the Page 28 → chamber by dividing its members into substantive jurisdictions and encouraging specialization within those jurisdictions. These characteristics of parties and committees, however, entail a number of disadvantages from an informational point of view. Obligatory participation and exogenous network structures constrain information exchange in that individual legislators interact with a predetermined group of colleagues when participating in party and committee activities. This group of social contacts tends to be fairly stable over time, at least within one legislative session. Legislators are, therefore, part of tightly knit social networks, which impedes the efficient flow of information: Information exchanged in such networks tends to be redundant, because most network members have access to the same (or at least similar) information, as we discuss in detail below. The reality that party networks are composed of ideologically homogeneous sets of participants further impedes information exchange, because ideological commitments may get in the way of finding “best” policy solution. Information provided by other legislators from the same party, party staff, or other internal resources may thus fail to contribute to finding solutions that most effectively promote the achievement of a particular policy objective. In this regard, committees differ from parties because their heterogeneous makeup alleviates the negative repercussions of obligatory membership and exogenous network structures. After all, committee members confront competing ideas and policy solutions following from the different preferences of committee members, which is not the case for parties. One crucial characteristic of committees greatly constrains the efficient flow of information, however—a closed issue scope through strictly confined jurisdictions. Committee jurisdictions can stand in the way of finding the best policy solutions because information exchanged in a given committee may ignore the full implications of policy decisions in other policy realms. Jurisdictions, in this sense, may be too narrow to address complex issues that touch on more than one policy area. They may also be too broad, however, if the issue at hand is too small to be efficiently addressed within a committee that covers an expansive policy realm. This problem does not arise in the same way in the context of parties, where issue scope is open and issues therefore can be addressed in a more flexible and encompassing fashion. In sum, core features of both parties and committees impede the efficient flow of information in legislative politics. Both suffer from the problems associated with obligatory participation and exogenous network Page 29 → structures. In the case of parties, this problem is alleviated by the openness of the issue scope, but it is in turn

aggravated by the ideological homogeneity of network participants. The reverse is true for committees: The set of participants is ideologically heterogeneous, which is an informational advantage. Issue scope, however, is closed, which impedes information exchange. For these reasons, parties and committees, as the principal formal legislative institutions, do not on their own solve the informational deficit in legislative politics. LMOs are one way in which legislators alleviate their informational collective action dilemma. LMOs differ from parties and committees in that participation in them is voluntary, which also entails that the structure of LMO networks is endogenous from the perspective of individual members. LMOs are ideologically heterogeneous, as they tend to be composed of members from multiple political parties. In addition, members of LMOs tend to sit on multiple committees, which further contributes to their heterogeneous makeup. Finally, issue scope is open, since legislators are free to create LMOs on any issue they consider a policy priority. All these characteristics are advantageous from an informational point of view, especially since LMOs do not exist in a vacuum: They are voluntary institutions that exist parallel to the formal institutional structure—in particular, parties and committees. Lawmakers can thus reap the informational benefits of parties and committees as well as LMOs. The key characteristic underlying the informational benefits of LMOs is that they are voluntary institutions. Legislators do not have to join any groups, they do not have to participate in their meetings, and they do not have to interact with their members. If they so choose, lawmakers can stay entirely outside of the LMO system. Unless a legislator takes on a leadership position in an LMO, nothing is mandatory. This feature is a disadvantage for LMOs, because they are incapable of imposing meaningful sanctions and are thus a poor mechanism for creating voting coalitions. LMOs are inefficient in this sense, as tends to be the case for the voluntary institutions described by Feiock and Scholz (2010). However, the voluntary nature of LMOs is also a key advantage. First, LMO membership is cheap—or, more precisely, transaction costs associated with participation in LMOs are as low as any given legislator wants them to be. Legislators can opt in and out of the LMO system as they see fit. They can weigh the benefits associated with participating against not participating and decide on a case-by-case basis whether to become Page 30 → involved. Because lawmakers can control their own level of involvement, LMOs are what members make of them: They can choose whether to use them as a resource for their work. As a result, LMOs may lie dormant or act only in the most passive way for extended periods until the issues they address rise in importance and salience. But when that happens, LMOs serve as a ready-made vehicle through which interested legislators may gather relevant information and develop politically viable courses of action to address the significant problems of the day. In addition, members can free-ride LMO benefits. Without contributing to LMO events or organization, a member can reap informational benefits by receiving emails, newsletters, or other informational resources distributed by the LMO. A second advantage of the voluntary nature of LMOs is that it allows them to coexist alongside parties and committees because all legislative actors understand that the formal responsibilities in parties and committees have greater priority than LMO activities. Legislators will not be blamed for expending more resources on party and committee work than on LMO activities; it is, in fact, expected that such will be the case. Of course, the reality is that LMOs do not play a formal role in the legislative process and that they lack the capacity to create legislative coalitions capable of passing bills, which would put them in competition with parties and committees. These characteristics are particularly important because they allow LMOs to exist and contribute to lawmakers’ legislative work without threatening the stability of the formal institutional system (Feiock and Scholz 2010).8 The voluntary nature of LMOs also relates to their open issue scope. LMOs do not have predetermined jurisdictions; in fact, they often involve issues or causes that “do not fit comfortably into the committee or party system” (Ainsworth and Akins 1997, 410). Lawmakers consequently can pursue their policy interests and priorities or those of their constituents regardless of their formal responsibilities in parties and committees. Legislators are free to join any LMO, while they are generally constrained in their ability to choose a given committee assignment, for example. LMOs thus provide a degree of flexibility when it comes to issue selection that is greater than in the case of parties and committees, which are more hierarchical in their organization and more constrained by legislative rules, norms, conventions, or expectations.

Voluntary participation and open issue scope, in turn, mean that LMOs bring together legislators with shared policy interests and priorities across formal institutional boundaries imposed by party affiliation and committee Page 31 → assignments. LMOs thus create heterogeneous networks of legislators with shared substantive interests and priorities. However, shared interests and priorities do not necessarily mean shared preferences and policy positions. Here it is instructive to differentiate between preferences regarding policies and preferences regarding outcomes. Policies are “the objects of legislative choice” (Krehbiel 1991, 66). Policy preferences thus relate to decision makers’ beliefs about the appropriate policy actions to bring about a particular goal. Outcome preferences are the preferences of legislators regarding the “the effects of policies upon their enactment and implementation” (Krehbiel 1991, 66); hence, outcome preferences relate to the goals policies are supposed to achieve rather than the means (policies) of achieving those goals. On the basis of this differentiation, we argue that members of a given LMO likely share common policy priorities and outcome preferences, but they often have divergent policy preferences. For example, members of the U.S. Congress may join the Congressional Caucus on Parkinson's Disease because they consider that the treatment of Parkinson's disease to be an important cause worthy of attention and support and because they share a common outcome preference—finding a cure for the disease. However, some members of the caucus may favor the use of stem cell research in this effort, while others strongly oppose it, meaning that they have divergent policy preferences.9 Indeed, it is quite likely that members of the LMO do not share common policy preferences because many LMOs are composed of actors from multiple parties.10 The implication is that LMOs allow legislators to establish relationships with colleagues with whom they often disagree but with whom they share common substantive interests, priorities, and preferences regarding outcomes. LMOs are not confrontational arenas of politics, however, in the same way that parties and committees are. Disagreements are common, but LMOs’ lack of formal legislative authority means that legislators interact with colleagues in a context that is less competitive, antagonistic, and constraining than would be the case in committees or on the floor. This, again, makes sense, because legislators would not choose to join voluntarily if participation made them worse off (Feiock and Scholz 2010, 12). Finally, voluntary participation, open issue scope, and heterogeneous sets of participants mean that LMO network structures are endogenous from the point of view of individual participants. These endogenous network structures are at the core of our argument about the informational utility of LMOs. Page 32 →

Information Diffusion through LMO Networks Our emphasis in this book is on LMO-based social networks as circuits of information exchange. We are interested in investigating the social structures that allow for the flow of information among individual actors who are linked to each other via joint membership in LMOs. Two characteristics of LMO network structures facilitate the efficient diffusion of information throughout the legislature:11 LMOs are primarily composed of weak ties that connect legislators who would otherwise remain isolated from each other. In other words, LMO ties are weak, bridging ties. While any tie between two nodes in a social network has the potential to allow for the passing of information, some ties are likely to allow for the more efficient flow of information than others. A useful distinction in this regard exists between strong ties and weak ties. Strong ties reflect close and frequent social contacts, such as those between friends or between close and trusted colleagues, while weak ties represent more casual social contacts, linking acquaintances whose connections are more tenuous and context dependent. In his influential article, Mark Granovetter (1973) emphasizes the “strength of weak ties” and maintains that they are particularly important for the diffusion of information throughout a social network. According to Granovetter, individuals benefit more from being affiliated with acquaintances (through weak ties) than with close friends (through strong ties) because tightly bound social circles are connected to each other not through strong ties but weak ones. Without such weak ties, social circles would be isolated from one another, which means that weak ties bridge “structural holes” in the network of legislators (Burt 1992, 2000, 2004), or the spaces between the different tightly knit regions of the network that would be left empty in the absence of the weak LMO ties.12 Friedkin (1980) and Feld (1981) offer explanations for why weak ties are more likely than strong ties to be

bridging. Their explanations are based on the same idea highlighted by Granovetter (1973, 1362; see also Burt 1992)—that is, the tendency of two individuals who are tied to a third to also be tied to one another, a phenomenon that Friedkin refers to as triadic closure and Feld calls transitivity. Friedkin maintains that ties between two nodes tend to be weak because “strong ties encourage triadic closure, which eliminates local bridges. Other things being equal, weak local bridges will tend to be maintained over time, while strong local bridges will tend to be eliminated” (1980, 417). Indeed, he finds that local bridges and inter-group ties are Page 33 → disproportionately weak ties and that two individuals who are strongly tied to the same actor via strong ties are more likely to be connected to each other. Along similar lines, Feld (1981, 1016) argues that the connections that are typically found between closely tied social clusters are based on foci—that is, social, psychological, legal, or physical entities around which joint activities of individuals are organized (e.g., workplaces, voluntary associations, hangouts, families, and so forth). These foci can be more or less constraining, meaning that they are more or less likely to force members to interact frequently and intensely. This implies that weak ties are generated by foci that are not very constraining. The less constraining an underlying focus, however, the more bridging a tie, because more constraining foci are more likely to entail transitivity. Therefore, weak ties are more likely to be bridging (Feld 1981, 1023–24). None of this is to say that strong ties cannot be bridging, but they tend to breed local cohesion and thereby overall social fragmentation (Granovetter 1973, 1378). Weak ties, in comparison, are more likely to provide the channels through which socially distant ideas, influences, or information can reach individual members of the network. Without weak ties, individuals are more likely to be cut off from knowledge outside of their most immediate social circle (Granovetter 1973, 1370–71). In the context of LMO networks, weak ties are not significant only because they are more likely to be crosscutting, however. They also matter because weak ties are less expensive to create and maintain than strong ties. This characteristic is important, given that voluntary institutions persist and flourish alongside formal ones only if the transaction costs associated with participation are low. This is the case for LMOs, where actors can choose to participate and interact with others on as regular a basis as they please. These interactions may not even be faceto-face, since lawmakers often have access to information exchanged in the LMO, such as position papers or policy briefs, without even attending meetings. Moreover, LMOs help legislators identify colleagues who can provide information on a given cause on an ad hoc basis, because LMO membership and especially LMO leadership signals which colleagues prioritize a cause and might therefore be able to provide information. For these reasons, weak, bridging ties are particularly important for the diffusion of information in a legislature. Consider the hypothetical network presented in figure 2.1, which illustrates our argument.13 The figure shows five LMO members and one LMO leader, plus twelve additional legislators, labeled A–L, who are not members of the LMO. Black lines indicate strong ties between legislators, while the lighter gray lines indicate weak ties. LMO Members 3–5 are part of small unique networks that may, for example, be conceived of as groups of legislators who are close to each other by virtue of regular and fairly intensive interaction in parties, committees, or ideologically based subgroups (e.g., factions within a given legislative party). Accordingly, the members of these three networks are all connected to each other via strong ties, which makes for a set of tightly linked tangential social networks. Page 34 → In contrast, most of the ties between LMO members are weak, indicated by the lighter gray lines; only three ties (between the LMO Leader and LMO Members 1 and 2) are strong ties. This strong subnetwork represents the few highly active members who are strongly linked to one another. All remaining internal LMO ties are weak and cut across the three tangential networks, thus connecting a more heterogeneous set of actors. This is an example of the more general principle in social networks that strong ties tend to be embedded in tightly linked regions of a larger network, while weak ties tend to cross between these regions (Easley and Kleinberg 2010, 7). Page 35 → We can observe here the key intuition behind the ideas of the strength of weak ties: If it were not for the weak ties between members of the LMO, the three tangential networks that revolve around LMO Members 3–5 would be

isolated from each other. The weak LMO-based ties provide for the possibility that information will reach all tangential social circles comprised of strong ties. This diffusion of information is not assured, of course, because not all information will be distributed equally throughout the LMO network. The structure of the LMO network, however, provides the potential of information flow into regions of the network that would otherwise be left isolated. While LMOs can thus collectively serve as conduits for information, they also give special prominence to LMO Members 3–5, as they connect the LMO network with their respective tangential networks. If LMO Member 3 did not exist, legislators A, B, C, and D would be isolated from the LMO-based network and would not benefit from the information that flows through the network. LMO Member 3's membership in the LMO constitutes the only link between these four legislators and the LMO network. LMO Member 3's social ties thus bridge a structural hole in the network. This position gives LMO Member 3 special prominence, which comes with a series of advantages. These advantages include earlier access to a broader diversity and higher volume of information, a competitive advantage in detecting and developing good ideas and rewarding opportunities, and increased attractiveness to other people as a contact in their own networks (Burt 1982, 1992, 2000, 2002, 2004). This position also provides LMO Member 3 with the opportunity to act as a broker in the communication between others and the flow of information through the legislative network (Burt 1992, 34–36; Burt 2000, 354–55; see also Aldrich 1999; Thornton 1999). In figure 2.1, LMO Member 3 is the only node within that immediate social circle with access to information diffused through the LMO network. As such, LMO Member 3 can regulate access to the tangential network composed of LMO Member 3 and legislators A, B, C, and D. This influence also runs the other way, as LMO Member 3 can choose to pass along or not pass along information from her tightly knit social circle to the LMO network. These observations imply that legislators may join and be active in LMOs in an effort to attain these pivotal positions between different subsets of the legislative social network, which can be used to advance personal interests and objectives. Such opportunism could come at the expense of the legislative organization as a whole, which would benefit from an Page 36 → unencumbered information flow across structural holes in the network. Much research has shown, however, that this tension between private and collective interests is not as stark as it might seem, because information and control benefits to individuals aggregate to the collective level.14 Burt (2000, 359–62) discusses numerous organizational studies highlighting this point (e.g., Krackhardt and Stern 1988; Ancona and Caldwell 1992; Rosenthal 1996; M. T. Hansen 1999; Leana and Van Buren 1999; M. T. Hansen, Podolny, and Pfeffer 2000). Innovation and creative policy solutions often arise from the synthesis of multiple ideas that may be known in distinct subparts of the network but that would not meet in the absence of actors and ties that bridge structural holds (Burt 2004, 350). Hence, individual-level brokerage generates constituency for new ideas synthesized from the diverse information clusters to which bridging actors have access (Burt 2000, 362), which in turn raises the aggregate level of “informedness” inside the network (Carpenter, Esterling, and Lazer 1998). This allows a more sophisticated exchange between network members, which, in a political context such as a legislature, can improve the quality of policy outcomes. We focus here on these collective gains to be reaped through the social structure provided by LMO networks.15 So far, we have emphasized the importance and utility of weak, bridging ties in the diffusion of information in social networks in general and in a legislative context in particular. This emphasis, however, masks a broader debate in the literature on the benefits of social network structure to information exchange and knowledge creation.16 This debate revolves around the trade-off between the value of weak, bridging ties that provide access to a diverse set of actors and nonredundant information and the cooperative benefits of highly embedded, strong ties that improve the quality of information exchange by increasing network participants’ confidence in the integrity of the social exchanges that take place inside a network.17 At the core of the latter position is the argument that greater network closure promotes reciprocity and trust among network members (e.g., Granovetter 1985, 1992; Coleman 1988, 1990; Uzzi 1996; Portes 1998; Uzzi and Lancaster 2003; McFadyen, Semadeni, and Cannella 2009). In other words, high levels of network density (the

number of observed ties in a network as a proportion of all possible ties), triadic closure (when three actors are all connected to one another), or embeddedness (when the two endpoints of a tie have a large number of common neighbors) benefit information exchange because they put the interaction between any two Page 37 → network participants on display in a social sense (Easley and Kleinberg 2010): When one party “misbehaves” (e.g., by providing misleading information), there is a greater likelihood for reputational costs and perhaps even sanctions to be imposed by friends or colleagues.18 Just as substantive information flows through the network, so does information about the misconduct of one of its members. Network closure thus increases participants’ confidence in the information exchanges that take place, a benefit that may be amplified by increasing tie strength between any two actors.19 This line of argument seems to contradict our proposition that weak, bridging LMO ties improve the flow of information in legislative politics. It is, therefore, important to consider the contexts within which LMOs provide the benefits of weak, bridging ties on the one hand and strong, embedded ties on the other. The argument about the informational value of weak ties that bridge structural holes between tightly knit regions of a network applies to the legislative network as a whole—that is, the social network composed of the entire population of legislators and the complete set of social ties that connect them. For the purpose of this discussion, we are referring to the network composed of all 18 actors in figure 2.1. In this context, weak LMO ties facilitate the flow of information by cutting across party and committee lines. Nested within this legislative network are two smaller ones: the social network composed exclusively of LMO members (LMO Leader and LMO Members 1–5) and the network composed of the LMO Leader and LMO Members 1 and 2—the three actors who are strongly tied to one another. The benefits of LMO ties inside those subnetworks differ from their utility in the broader legislative context. Most important, both of these networks are characterized by high levels of network closure, because everyone is connected to everyone else. Consider, for example, the number of mutual neighbors shared by LMO Members 3 and 4 in figure 2.1, focusing only on the six-actor LMO network (thus ignoring the tangential networks).20 We can see that LMO Member 3 and LMO Member 4 share four direct neighbors: the LMO Leader as well as LMO Members 1, 2, and 5. The tie connecting LMO Members 3 and 4 thus has an embeddedness of four, which is the highest possible value because there are only four neighbors to be shared. If we repeated this exercise for all members of the LMO, we would find that we were dealing with a network entirely composed of ties that are maximally embedded. This increases network participants’ confidence in the information exchanged. Page 38 → Figure 2.1 illustrates the great degree of closure in the LMO network, yet it also shows the majority of ties between LMO members to be weak ties (represented by thin gray lines). While the weakness of those ties may not promote additional reciprocity and trust between network members,21 it nevertheless constitutes a second notable benefit for members of the LMO network: Weak LMO ties allow for the diffusion of policy-relevant information between legislators who care about a common cause, but they are inexpensive to establish and maintain. After all, the level of involvement of most legislators in LMOs is too limited and irregular to provide for the creation and maintenance of strong ties, since participation in LMOs is voluntary and competes with other, obligatory legislative activities like committee meetings. Therefore, LMOs provide extensive sets of weak ties that are maintained through occasional participation in LMO events, attendance at social gatherings, or simply the reception of information through an LMO's email list.22 The establishment and maintenance of these weak ties does not require extensive and continuous time or effort, but such ties still provide valuable access to information about an issue or cause that all LMO members deem important. Finally, consider the second subnetwork in figure 2.1, the small network composed of the LMO Leader and LMO Members 1 and 2. This network of strongly tied LMO members represents the small group of LMO participants who are particularly active and/or serve in leadership positions; we might thus refer to it as the LMO leadership network. Like the LMO network, the LMO leadership network is characterized by high levels of network closure, since all members have direct ties to one another, yet it differs in that these ties are strong ties (as the thick black lines indicate). These strong ties are the result of the greater frequency and intensity of social interactions among

members of the LMO leadership network. Hence, in the LMO leadership network, both network closure and strong ties enhance reciprocity and trust in social exchanges. While the costs of maintaining strong ties might outweigh their benefits for regular LMO members, LMO leaders and highly active members are willing to carry these costs for three primary reasons. First, their voluntary membership in the leadership network implies that they care more than the average member about the LMO's cause; hence, they are more willing to expend time and effort on LMO activities, including the maintenance of strong social ties. Second, most political coordination (as opposed to information exchange and signaling) in LMOs happens inside the LMO Page 39 → leadership network. This cooperation and coordination necessitates and therefore increases the value of strong ties (as well as produces them). Third, the value of strong ties is particularly pronounced in the LMO leadership network because the great majority of LMOs purposely set up their leadership to include members of more than one party—legislators who are at least nominally political opponents. In this context, where political actors with divergent preferences seek to coordinate their activities, the need for reciprocity and trust is particularly pronounced. In sum, the distribution of LMO ties captures the informational benefits of weak (cheap), bridging ties in the legislative network; of weak (cheap), embedded ties in the LMO network; and of strong, embedded ties in the LMO leadership network. LMOs ties are informationally valuable because they allow lawmakers to reap the benefits of network closure and tie strength within the LMO and LMO leadership networks, yet they bridge structural holes in the legislative network to encourage the flow of novel and diverse information.23 This flow of diverse information, while important in a variety of social contexts, is particularly valuable in legislative politics. Legislators, unlike many other types of political actors, have strong incentives to seek information not only from sources with whom they are Page 40 → predisposed to agree but also from sources who do not necessarily share their preferences. Research on social networks shows that people tend to choose likeminded political conversation partners (Bauer, de Sola Pool, and Dexter 1963; Milbrath 1963; Huckfeldt and Sprague 1987; Mutz 2006; but see Huckfeldt, Johnson, and Sprague 2004) and that voluntary associations in particular have overwhelmingly homogeneous membership (Feld 1982; McPherson 1983; McPherson and SmithLovin 1986, 1987; Marsden 1990; McPherson and Ranger-Moore 1991; McPherson, Popielarz, and Drobnic 1992; Popielarz and McPherson 1995). On the basis of these studies, we might also expect legislators to avoid associating with political adversaries—for example, because they can rationally reduce the costs of obtaining information by developing relationships only with those who are well-informed and like-minded (Downs 1957) or because of the disabling consequences of disagreement in discussion groups (Lazarsfeld, Berelson, and Gaudet 1944; Berelson, Lazarsfeld, and McPhee 1954; Asch 1956; Mutz and Martin 2001; Mutz 2002b). Yet legislators differ from other political actors, such as lobbyists or voters, in that they cannot afford to suppress all dissonanceproducing information by choosing to interact only with those with whom they are predisposed to agree (Ringe, Victor, and Gross 2013). In fact, having information that is “wholly and completely an extension of individually based political preferences” (Huckfeldt and Sprague 1987, 1199) would put legislators at a great strategic disadvantage compared to their colleagues. To succeed in their strategic interaction, lawmakers must maximize their information on the content and expected consequences of the policy proposals as well as on the positions, strategies, and goals of their counterparts and political opponents. LMO-based policy networks and the weak, bridging ties of which they tend to be composed provide an important structure for the exchange of such crosscutting information.

An Invitation to Free-Ride: LMO Participation by Outside Advocates This discussion raises two important questions. First, what are the sources for the information that is exchanged efficiently through LMO networks? And second, if LMOs are low-cost institutions from the point of view of individual participants, who carries the cost of creating and maintaining them (Feiock and Scholz 2010)? Partial answers to these questions lie within the legislatures themselves, in that individual (groups of) legislators both assume the cost of running LMOs and serve as sources of information for their colleagues. Individual legislators and their offices or small groups of lawmakers often choose to expend some of their limited

resources on creating and maintaining LMOs. They may do so for personal reasons. For example, legislators who have loved ones affected by Alzheimer's disease might be prompted to form an LMO devoted to finding a cure, or lawmakers may be union members and seek to provide a forum for labor issues. Alternatively, legislators may have reason to establish an LMO because it reflects a key concern of their constituencies or because they did not get the committee assignments of their choice and use the LMO to pursue particular personal interests. Any such reason may serve as a strong enough incentive for legislators to take on the task and the costs of establishing and running LMOs. LMO leaders also serve as key providers of information within the LMO—for example, because they possess specialized expertise on the issue or cause their group promotes or because they are willing to have their staffers Page 41 → research a topic and draft a policy report. Leaders may also serve in a coordinating function, calling on other group members to provide expertise during informational events held by the LMO. Internal expertise, however, is unlikely to fully accommodate the informational needs of legislators for a number of reasons related to the voluntary nature of LMOs. First, the benefits associated with LMO participation must outweigh the costs. If the task of generating and providing internal expertise became too cumbersome, rank-andfile LMO members would have little reason to join and participate. The second, related reason is that many members participate only sporadically in LMO activities. As a result, LMOs are unlikely to foster the same level of individual-level expertise as legislative committees, for example, where participation is obligatory. After all, just because legislators prioritize a cause does not mean that they know all about it; indeed, lawmakers may join LMOs for the express purpose of learning about an issue. Finally, if all relevant information were readily available internally, there would be little need to establish the group, as lawmakers would find alternative ways to gather the information they seek. It is similarly unlikely that actors inside the legislature carry all costs associated with creating and maintaining LMOs, given the limited resources available to them. Running an LMO tends to be costly in terms of time and effort, and the payoffs associated with LMO activities are lower and more diffuse than participation in the formal legislative work of parties and committees. For some legislators, these costs may be worth carrying. It is not likely, however, that an expansive and active LMO system would be carried solely by such volunteerism. Who, then, helps cover the full costs of establishing and maintaining LMOs and provides much of the information that is efficiently diffused through LMO-based social networks? The answer lies in the extensive, institutionalized, and privileged ties many LMOs establish with actors outside the legislative realm. LMOs and their leaders regularly interact with “outside advocates”—that is, political actors and groups outside their chamber who share the LMOs’ substantive focus. One of the key goals of these outside advocates is to attract the attention of one or more policymakers and to appeal to potential supporters inside the legislature to advance their cause. LMO membership is one important way to recognize these potential supporters, since becoming a member of a given LMO indicates that a legislator considers the LMO's issue or cause to be of some importance. LMO Page 42 → members are, therefore, potential “inside advocates” who seek to promote the LMO's cause by trying to raise awareness, set the agenda, or influence policy proposals within the chamber. The prospect of working closely with inside advocates on a common issue provides a strong incentive for outside advocacy groups to help create and maintain LMOs, because close allies inside the legislature can be invaluable in efforts to influence policy, since it is harder for undecided legislators to turn away colleagues engaged in inside advocacy (Ainsworth 1997, 523). Outside advocates thus help create and maintain LMOs—for example, by gathering initial support for the establishment of a group, by keeping membership lists, by organizing events, and by supplying information relevant to legislators’ policy-making activities that is then diffused through the LMO network.24 The engagement of outside advocates constitutes an example of what Hall and Deardorff (2006, 72) call lobbying as “an attempt to subsidize the legislative resources of members.” This conceptualization implies that outsiders help keep LMOs cheap. LMOs are set up in such a way that they not only allow but encourage free riding: All legislators can come and go, participate or not, without facing sanctions for their choices. LMO leaders and their offices and/or outside advocates readily assume the costs for everyone who chooses to be involved.

Our argument about the involvement of outside advocates in LMOs deviates from Hall and Deardorff's (2006) in two notable ways, however. First, Hall and Deardorff conceptualize lobbying as an effort to provide “costly policy information, political intelligence, and labor” to “already likeminded legislators” (69). In contrast, we maintain that legislative subsidy in the context of LMOs is not limited to supporting political allies because LMO members share policy priorities, not preferences. Hence, legislative subsidies for LMOs likely benefit at least some lawmakers whose preferences do not match those of the outside advocates. Those members are potential supporters, however, which leads to the second way in which our argument differs from Hall and Deardorff's. They maintain that direct lobbying “typically is not a strategy for changing legislators’ preferences over policy” (72). Yet given that at least some level of preference divergence exists among LMO members, the provision of information as a legislative subsidy cannot be separated from the provision of information as a potential act of persuasion. Because even lobbying as legislative subsidy may involve this element of persuasion, legislators cannot always be sure that the information provided by outside advocates is of high utility, since the preferences and goals of Page 43 → outside advocates may not match those of the legislators. As a result, the information outside advocates reveal is both strategic and not inherently preference-neutral (Crombez 2002). In fact, we can take it as a given that outside advocates will provide information in line with their own policy preference, all else being equal. Of course, legislators are well aware of this fact (J. M. Hansen 1991) and will take steps to counteract perceived biases in the information they are offered. One way of achieving this objective is to diversify the supply of information and to shape biases in a way that prevents legislators from becoming entirely dependent on a single interest. This approach allows groups to counteract their opponents’ efforts at misrepresentation (Austen-Smith and Wright 1992, 1994; Beyers 2004; Esterling 2004). The downside to this strategy is that it forces legislators to process and evaluate a potentially huge amount of information. As Esterling points out with regard to the U.S. Congress (2004, 230), “In relation to interest groups the problem for Congress is not dependence but complexity.” This complexity may undermine one of the key characteristics of LMOs as voluntary institutions that help mitigate legislators’ informational collective action dilemma: to help gather information in an efficient and low-cost manner. If the principal way of avoiding biased information is to diversify input, the costs associated with LMO participation may quickly outweigh the benefits, giving legislators little incentive to join. LMOs help members deal with this dilemma by simultaneously providing three incentives for outsiders to provide high-utility information, or high-quality information that is research-based (see Esterling 2004), reliable (see Coen 2007, 339), and presented in an easily digestible format such as a policy brief, a position paper, or a presentation by an outside expert (see Hall and Deardorff 2006, 74). First, at least some LMOs stimulate openness by bringing together a variety of interests in deliberative forums, such as panel discussions that pit diverse interests against each other. This approach allows legislators to reduce uncertainty about the likely impact of legislation upon implementation by triangulating publicly revealed information. Second, individual LMOs feature a high level of internal embeddedness, where any two members of a singleLMO network have a high number of common neighbors. Since outside advocates are connected to these networks, they are tied into these highly embedded social structures, thereby increasing the likelihood that they will be found out and suffer reputational costs if they provide low-utility or misleading information (Granovetter Page 44 → 1985, 1992; Tullock 1985; Coleman 1988, 1990; Greif 1989; Uzzi 1996), especially because at least some legislators and other participants in LMOs are also policy experts who might recognize that the information provided lacks proper foundation. Finally, LMOs provide an incentive structure for the supply of high-utility information in that they are a valuable access point for outside advocates who seek to promote their policy agendas and advance the interests of their clients. As a consequence of this reality, legislators can credibly threaten to cut off those outsiders who provide redundant information in line with commonly known positions (Gilligan and Krehbiel 1987, 1989; Lupia and McCubbins 1998) or who supply inconsistent, unsubstantiated, or even misleading information (Schlozman and Tierney 1986; J. M. Hansen 1991; Ainsworth 1993; Esterling 2004). After all, policymakers ultimately get to decide which lobbyists are included in or excluded from the legislative process (Maloney, Jordan, and

McLaughlin 1994; Broscheid and Coen 2003, 2007; Grant 2004, 2005). LMOs not only provide access points for outsiders but also grant privileged access to a single (or a limited number of) outside advocate(s). This notion of privileged access is significant for two reasons. First, outside advocates ultimately aspire to such privileged relationships; they want to build long-term relationships with lawmakers and become trusted sources of advice and policy input (Berry 1989; Victor and Koger 2011). In fact, Bernhagen and Bräuninger (2005) demonstrate that under certain conditions, business interests are prepared to abandon their short-term policy goals in favor of maintaining their reputations (Marshall 2010). Second, access to policymakers is most valuable to outside advocates if few others have it, because the more outsiders are granted access, the more diluted the benefits associated with it become and the smaller the relative influence of a single advocate (Broscheid and Coen 2003, 2007). Accordingly, losing access as punishment for providing low-utility information is especially costly in the case of LMOs, because it means losing a particularly valuable relationship with legislators and their offices (see Broscheid and Coen 2003, 2007). Moreover, losing access is particularly costly since an LMO represents a group of lawmakers whose participation signals that they consider the cause to be a policy priority. This loss of access also brings the risk of empowering direct competitors in the marketplace for outside advocates, since the legislators will have to look for new providers of the informational goods they require to make informed policy choices. Accordingly, outside advocates who have gained privileged access via LMOs are induced to consistently provide high-utility information. The costs of not doing so are exceedingly high. Page 45 → The logic of our argument about LMOs as institutionalized points of access to a small number of outside advocates is captured in figure 2.2, which shows the legislative network from above but adds one more actor, an outside advocate. This network shows a strong tie between the outside advocate and the LMO Leader as well as LMO Members 1 and 2. These strong ties indicate the privileged access enjoyed by the outsider; as such, they provide the incentive for the outside advocate to supply high-utility information.25 The strong ties also suggest that the interactions between the LMO leaders and the outside advocate are iterative, further amplifying this effect (see Page 46 → Schlozman and Tierney 1986, 97; J. M. Hansen 1987, 1991; Kingdon 1989, 82; Esterling 2004, 83–85). Providing low-utility information at time t may be punished with the withdrawal of nonpolicy benefits at time t + 1. In sum, the structure of this LMO-based network—in particular, the distribution of strong and weak ties in it—increases the likelihood that high-utility information enters the legislature and is diffused throughout the LMO network. While the strong ties between the outside advocate and the LMO leadership ensure that the information supplied by the outside advocate and introduced into the LMO network is of high utility, the weak ties in the LMO network facilitate the diffusion of this information. Ties established through shared LMO membership and the social networks it creates thus provide an opportunity structure for the efficient flow of policy-relevant information.

Conclusion LMOs connect legislators as well as legislators and outside advocates who share a common policy interest. The social networks made up of LMO ties facilitate the flow of policy relevant information throughout the legislature. The structure of LMO networks and the weak, bridging ties of which they tend to be comprised provide lawmakers with access to information that would otherwise be difficult to obtain. These weak ties are particularly important for the flow of information in a legislature because the cross-cutting membership of LMOs transcends the boundaries imposed by partisanship, opposing ideology, and committee jurisdiction, thus helping to bridge structural holes in the legislative social network. LMOs are voluntary institutions that impose few costs on legislators who choose to participate. Indeed, LMOs invite legislators to free-ride on the informational benefits the organizations provide, as a small number of lawmakers expend time and resources on LMO maintenance—and many of the costs associated with creating and maintaining LMOs are carried by outsiders with institutionalized, privileged access to LMOs. In return for this

privilege, outside advocates not only help run LMOs but also provide information in an environment where information is scarce. Moreover, for fear of losing their privileged position in the competitive marketplace of interest group representation, outside advocates have strong incentives to Page 47 → supply high-utility information that is carefully researched, reliably of high substantive quality, and processed for easy consumption. In this sense, the LMO-based networks that include both legislative insiders and outsiders are more structured than the more open “issue networks” of actors involved in the deliberation, negotiation, and decision making in a particular policy area described in some of the Americanist literature on lobbying (Heclo 1978; Sabatier 1988; Jenkins-Smith and Sabatier 1994). In particular, the set of participants in LMOs is less fluid than is the case in issue groups, which more easily allow for participants to move in and out of the policies and issues defining the network (see also Gais, Peterson, and Walker 1984; Browne and Paik 1993). Indeed, LMOs allow lawmakers to deal with the great and increasing number of outside actors by structuring access to the legislature and its members (Ainsworth 1997, 517; see also Ainsworth 1993). LMOs help maximize the utility of any given connection or tie with outside advocates, in turn enhancing the utility of the information these outsiders provide. In environments where “facts compete” (F. Baumgartner et al. 2009, 66), this dynamic reduces informational uncertainty, alleviates the risk that lawmakers will become associated with poor public policy, and enhances lawmakers’ inside advocacy efforts. LMOs are, of course, not the only way in which legislators structure their institutional environment in an effort to build relationships and improve the quality of the information they receive. Parties, staff, and especially the committee system can play similar roles inside the legislature. However, the characteristics of parties and committees when it comes to rules of participation, membership, issue scope, and network structure can be informationally constraining. In contrast, the voluntary nature of LMOs, their heterogeneous membership, and their open issue scope and network structures help alleviate the informational collective action dilemma in legislative politics, where great demand for information is met with inadequate supply. In sum, LMO networks and the distribution of strong and weak ties in them increase the likelihood that highutility information is diffused throughout the legislature, thereby allowing legislators to make more informed decisions. The theoretical framework laid out in this chapter leads to a number of expectations about what we ought to find in our empirical analyses. These expectations relate to the nature and distribution of social ties in LMO networks on the one hand and to the benefits of LMOs to Page 48 → lawmakers on the other. We surmise that LMO network structures should be conducive to the efficient exchange of information. Most important, we expect that social ties established through joint membership in LMOs will cut across groups of legislators that are connected through various institutional memberships, particularly parties and committees. LMO ties thus are bridging ties. Furthermore, we expect our qualitative data to highlight both the relational and informational benefits of LMOs: LMOs give legislators opportunities to build social relationships with one another and facilitate the flow of high-utility policy and political information inside the legislature as well as between insiders and privileged outsider advocates. We expect these relational and informational benefits to be legislators’ primary motivations for participating in LMOs, but we also consider evidence for the signaling and representational functions of LMOs and their importance for political coordination. Indeed, we would be surprised not to find some evidence for all three functions across our cases, since they are not mutually exclusive. While we expect to find support for our informational theory of LMOs across the two case studies, variation should occur with regard to both coordination and signaling. We expect to find more evidence for political coordination as the need for cross-partisan cooperation in the legislature increases. Hence, we should find some evidence for bipartisan coordination in the U.S. Congress, where party cohesion is less pronounced than in some other legislatures (especially in parliamentary regimes) and the logic of logrolling necessitates a degree of cross-partisanship. The need for coordination and cooperation is greater in the European Parliament, however, for two reasons. First, the distribution of seats is such that no single party group can unilaterally pass legislation. Furthermore, only a single two-party coalition, between the two large center-right and center-left party groups—the European People's Party (EPP) and the Progressive Alliance of Socialists and Democrats (S&D, formerly the Party of European Socialists [PES])—can muster a majority large enough to pass laws. Hence, unless the two largest party groups collude, at

least three parties are necessary for legislation to pass for sure. The second reason for greater levels of crosspartisanship and political coordination across party lines in the European Parliament is a distinct norm of consensus that compels members to seek the broadest possible coalition of supporters for a given piece of legislation. While this norm is neither a formal requirement nor always observed, it is an important codeterminant of the behavior of members (see Benedetto 2006; Ringe 2010). In general, both cross-partisan information exchange Page 49 → and political coordination ought to receive greater emphasis in the European Union than in the United States. We also expect to find, however, that the less formalized nature of LMOs will be significant for successful political coordination in both chambers and that coordination will be more important for LMO leaders than for occasional participants, who will primarily reap the informational benefits of LMOs. We also expect to find variance in the extent to which LMOs serve as a means to signal policy priorities and preferences to constituents. Here, the electoral system will make the difference, because constituency service is more important in electoral systems that emphasize close links between legislators and constituents. Accordingly, we expect our data to show constituency concerns to be more prevalent motives for legislators to join and participate in LMOs in the U.S. Congress than in the European Parliament, where lawmakers are elected using proportional representation and party lists and the electoral connection to voters is far less developed. To summarize, we expect to find that the primary benefits of LMOs for most members are relational and informational and that although political coordination and constituency service matter, they are of secondary importance for most lawmakers; ties between members of LMOs tend to be weak and bridging; LMOs give their members access to high-utility policy and political information; LMOs and their members develop privileged relationships with particular outside advocates, which facilitate the exchange of information with them; outsiders also provide legislative subsidies to LMOs and their members that make participation in LMOs cheap for most legislators; systematic differences emerge from our two case studies, and the most important of those differences is that signaling and constituency service matter more in the U.S. Congress than in the European Parliament because of stronger legislator-constituency links, while the greater need for cross-partisanship in the European Parliament entails greater emphasis on cross-partisan information exchange and political coordination. The reminder of the book is dedicated to testing these expectations.

Page 50 →

CHAPTER 3 Where in the World Are LMOs? Legislative member organizations (LMOs) exist in a great number of legislatures. They are a more widespread phenomenon than most people—including most legislative scholars—realize. In this chapter, we seek to answer two elementary but exceedingly important questions: First, where do LMOs exist, and second, what factors affect the likelihood that LMOs will emerge and flourish in a lawmaking body? These questions are not easily addressed, as we discovered during the year we spent attempting to gather comparative information on LMOs. There is no centralized database or other resource that provides information on the existence of LMOs, much less the rules that govern them, the legislators who join them, or the activities in which they engage. Disparate standards in the quantity and quality of information provided by parliamentary websites further complicate the search for information on LMOs. Finally, in some countries with LMOs, the groups are so informal that even experts on the relevant national legislatures can provide very little information on them. In an effort to overcome these difficulties, we used several techniques to compile comparative data on LMOs in a broad sample of advanced industrial democracies. At the heart of this effort is an expert survey that seeks to ascertain whether LMOs exist in a given national legislature. We also tried to gather as much information on LMOs in various countries as possible, relying on the responses to our survey as well as additional information gathered directly from legislatures and from a handful of academic and popular news sources. This chapter has three key purposes. First, we take the initial step of presenting basic descriptive information about the presence of LMOs across a broad sample of legislatures. Next, we describe the general conditions under which LMOs exist in various legislatures, using this information to set the stage for our case studies. Finally, this chapter provides a Page 51 → springboard for scholars interested in expanding this research and pursuing more detailed information about LMOs in various countries. We begin by describing our data collection efforts and then present a quantitative analysis to identify what factors make it more or less likely that LMOs exist in a given national legislature. This analysis reveals the importance of three key variables: the electoral system, the size of the legislature, and the effective number of legislative parties. Specifically, majoritarian electoral institutions, a higher number of seats, and a greater number of parties increase the probability that LMOs exist. Finally, using these indicators as the basis for categorization, we provide descriptive information about LMOs in 25 legislatures.

Our Sample of Countries To conduct our cross-national, broad-based examination of LMOs, we first had to collect information on the existence of LMOs in a sample of advanced industrialized, liberal democracies.1 We included in our sample countries that are members of the Organisation for Economic Co-Operation and Development (OECD), the European Union (EU), or the Group of 20 (G-20) and that have “institutionalized democracy” scores of 8, 9, or 10 in the 2009 Polity IV Dataset.2 These selection criteria result in a geographically and institutionally diverse sample of legislatures on six continents. The 45 countries in our analysis are listed in table 3.1.

Data Collection and Classification LMOs are underresearched legislative institutions, which makes collecting systematic data on them from a large number of countries exceedingly difficult. Ours is also the first comparative examination of LMOs, which means that there is little previous research on which we can build. There exists no centralized or universal source that provides information on the existence, makeup, activities, or purposes of LMOs. Moreover, considerable variance exists across legislatures in the formality and quality of LMOs. We used a variety of methods to ascertain whether LMOs exist in various legislatures, including searching legislatures’ websites, calling legislative clerks, and

asking academic specialists; however, we learned that Page 52 → no single source was sufficient for us to gain confidence that our findings were accurate. We therefore used multiple methods to gather the data we needed. We first sought to establish where LMOs exist and where they do not by conducting an email expert survey.3 Our initial list of contacts was drawn from The Handbook of National Legislatures (Fish and Kroenig 2009). We then used a snowballing technique, extending our initial sample following recommendations from our first-round respondents and other personal contacts. In total, we contacted 379 individuals, of whom 318 are academics, 32 are parliamentary officials, and 29 are “others” (for example, nonacademics who work for public policy organizations, government officials who work for institutions other than the legislature, and the like). We received 102 usable responses from 76 academics, 15 parliamentary officials, and 11 others,4 for a response rate of 27 percent. The average number of individuals we contacted per country was 8, with a range from 1 (Luxembourg and Malta) to 16 (Australia and Canada). Page 53 → To take full advantage of respondents’ expertise, we singled out those who indicated either that they could provide more detailed information about LMOs in their countries or that they had additional knowledge beyond what was asked in our initial set of questions. We contacted these individuals with a series of further questions.5 In this second survey round, we followed up with 21 respondents (16 academics, 1 parliamentary official, and 4 others). We received 13 responses from 10 academics and 3 others, for a response rate of 62 percent.6 We supplemented our survey with academic and popular news sources for the nine countries where we received contradictory responses or where information was scarce: Australia, Belgium, Brazil, Chile, Germany, India, Indonesia, Italy, and New Zealand. We also contacted 44 of the legislatures in our sample directly (via email and /or phone) to gather information about the existence and nature of LMOs.7 Finally, we relied on the European Centre for Parliamentary Research and Documentation (ECPRD) Request 1075, “Legal Rules Concerning the Founding of Deputies’ Groupings,” which provides some information on associations of parliamentarians as well as the relevant legal provisions, parliamentary practices, and functioning across a range of European democracies.8 The most important finding from our survey is that LMOs exist in the national legislatures of 25 of our 45 countries.9 However, a great deal of variance exists in the level of institutionalization of these groups in different legislatures. LMOs in the United Kingdom, for example, are highly regulated in comparison to those in Italy, New Zealand, and South Korea. We therefore differentiate between institutionalized and noninstitutionalized LMOs. The former describes systems where there are explicit parliamentary rules governing LMOs, where LMOs are registered by the parliament,10 or where at least some LMOs charge membership fees. These three characteristics do not apply in noninstitutionalized LMO systems.11 Table 3.1 provides a list of the countries in our sample with LMOs, differentiating between institutionalized and noninstitutionalized ones. Figure 3.1 shows where LMOs do and do not exist.

What Predicts the Presence of LMOs? A first step in our analysis is to predict what factors make it more or less likely that LMOs exist in a given legislature. To do so, we conduct a binary logit regression analysis, where the dependent variable is whether LMOs exist. The independent variables are Page 54 → size of legislature number of legislative parties parliamentary, presidential, or hybrid system federal or unitary country bicameral or unicameral legislature

ethnic fractionalization electoral system (majoritarian, proportional representation, or modified) personal voting (whether electoral institutions create incentives for politicians to cultivate a “personal vote” [Johnson and Wallack 2007]) Table 3.2 provides descriptive statistics and sources for our predictors. Our expectations about where LMOs exist follow from the three primary functions of LMOs identified in the preceding chapter. Broadly, we expect that legislatures that have a greater need for information exchange and coordination across homogeneous social groups are more likely to have LMOs, because LMOs provide opportunities to develop and maintain cross-cutting relationships both within the legislature and with outside advocates, and these relationship may form the basis for both information exchange and political coordination. Page 55 → Page 56 → The most important determinants of LMOs are observed at the level of the legislature. We therefore consider the factors inside a legislature that increase the need for information exchange and political coordination across groups of lawmakers, and we expect that LMOs are more likely to exist in large legislatures with a greater number of legislative parties. More general institutional and societal factors also might affect the likelihood that LMOs exist in a legislature. For example, LMOs may be more likely to exist in presidential systems because the separation of powers creates incentives for information exchange and political coordination between legislative actors; in federal countries, where there is a potential need for information and coordination to address the needs of constituents at different levels of governance; and in bicameral legislatures, where the consent of both chambers may be needed to pass legislation. Finally, greater societal heterogeneity might increase the probability that LMOs are formed and maintained, since more diverse interests need to be accounted for in the policy-making process. We also expect that the electoral system influences the likelihood of the existence of LMOs in a legislature. This expectation has less to do with the need for information exchange and political coordination than with incentives to signal policy preferences and priorities to constituents. Accordingly, we expect to find LMOs in countries with majoritarian electoral institutions and to a lesser extent in countries with modified electoral systems. They are least likely to emerge in countries that use pure proportional representation.12 Moreover, we expect personal voting, where voters have the option of electing individuals rather than parties, to affect the likelihood of observing LMOs in a legislature, since personal voting strengthens the link between voters and representatives. We expect this variable to be positively associated with the presence of LMOs. Table 3.3 presents the results of our analysis. Model 1 includes all of our predictors and indicates that three factors affect the likelihood that LMOs exist in a given national legislature: the size of the legislature, the effective Page 57 → number of legislative parties, and the electoral system. Both legislature size and the number of parties have positive, statistically significant coefficients, indicating that the existence of LMOs is in part explained by institutional factors at the level of the legislature itself that produce a greater need for information exchange and political coordination. Also significant are the binary variables for proportional representation and modified electoral systems. Their negative signs indicate that legislatures where members are elected using those systems are less likely to have LMOs than those with majoritarian electoral institutions. Additional analyses (not shown here) do not reveal any statistically significant difference between proportional and modified electoral systems in predicting the existence of LMOs. These findings suggest that a greater need for signaling preferences and priorities to constituents increases the likelihood of the existence of LMOs. However, this effect is limited to majoritarian electoral institutions, as the personal voting variable does not achieve statistical significance. LMOs, in other words, appear to serve as signaling devices only in legislatures whose members are elected using majoritarian electoral systems. The remaining structural variables (parliamentarism, presidentialism, federalism, bicameralism, and ethnic

fractionalization) do not achieve statistical significance. We drop these predictors in Model 2 to check the robustness of our results, given that we are working with only 45 observations.13 The findings hold, except that the effective number of parties variable is now only borderline statistically significant (z = 1.71).14 Page 58 → To facilitate the interpretation of our results, figure 3.2 presents the predicted probabilities on the basis of Model 2. The predicted probability of finding LMOs in a legislature equals 1 when legislators are elected using majoritarian electoral systems; indeed, all six countries in our sample with majoritarian electoral systems have LMOs. In contrast, the predicted value for both proportional representation and mixed systems is effectively 0 when the other variables are held at their means. The predicted value for the smallest legislature in our sample (59 members) is 0.74, a number that increases to 0.99 for the largest legislature (650 members). Finally, a legislature Page 59 → with the minimum effective number of parties (1.94) has a predicted probability of 0.81, while the largest number of parties (9.32) results in a predicted value of 0.99.

Categorizing LMOs Our statistical analyses tell us a great deal. Multivariate models enable us to highlight three key predictors of the existence of LMOs in our sample of advanced liberal democracies: the size of the legislature (institutions with a higher number of seats are more likely to have LMOs); the effective number of parties (legislatures with more parties are more likely to have LMOs); and the electoral system (countries with majoritarian electoral systems are more likely to have LMOs). Yet while our statistical models help us to identify the significant predictors of LMOs, they do not provide a feel for the nuances in LMOs across the parliaments. For that, we must turn to a more descriptive analysis of our cases. Category 1: Majoritarian Electoral System LMOs exist in all six countries in our sample with majoritarian electoral institutions—Australia, Canada, France, India, the United Kingdom, and the United States. Furthermore, all except Australia have institutionalized LMO systems. Australia is also the only one of the six countries that has a smaller than average legislature, while India is the one country that has a higher than average effective number of parties. Our first category of countries with LMOs thus features those with majoritarian electoral systems. These cases are listed in the right column of table 3.4, which illustrates our categorization. France, the United Kingdom, and the United States have a very large number of LMOs. In France, for example, the National Assembly lists 117 groupes d'études (study groups) on its website, including groups on tourism (tourisme), the defense industry (industrie de defense), the integration of disabled and vulnerable people (intégration des personnes fragilisées et handicapées), construction and housing (construction et logement), medicine and health products (médicament et produits de santé), and olive growing (oléiculture). Eighteen of these groups have only a single member, while the largest—the Study Group on the Tibet Question—includes 172 deputies; the average number of Page 60 → members is 48 (in a legislature of 577 members). When describing the activities of groupes d'études, respondents to our survey suggest activities similar to those we observe in our case study chapters—that is, studying the issues to which the groups are devoted; trying to advance their cause; staging events, colloquiums, and hearings on various issues of interest (which happen more frequently when a bill related to a group's cause is under consideration); holding meetings with experts and lobbyists; and publishing policy reports. Groups may also facilitate the creation of compromise or consensus on a given issue and thus help the adoption of parliamentary bills. One respondent reports that groupes d'études play a role that is “quite important” in the French National Assembly. Page 61 → Aside from caucuses in the U.S. Congress, perhaps the most well-known LMOs are all-party groups (APGs) in the United Kingdom. The Westminster website describes APGs as “informal, cross-party, interest groups that have no official status within Parliament and are not accorded any powers or funding by it” and highlights the fact that

some groups “have existed for many decades whereas others come and go in response to issues of the day.”15 The number of APGs has increased significantly over time, as a comparison of the number of subject groups indicates (subject groups are APGs devoted to a particular issue or cause as opposed to country groups that relate to a particular country or region): the number of subject groups has grown from 80 in 1986 to 142 in 1996 (Ball 2011b) to 444 at the end of the 2005–10 Parliament. Among those 444 APGs are groups on such diverse issues as breast cancer, corporate governance, electoral reform, headache disorders, photography, agriculture and food, tribal peoples, jazz appreciation, housing justice, and street children. LMOs also exist in Canada, where they do not have a single label but are known as caucuses, all-party parliamentary caucuses or groups, or parliamentary coalitions or committees. Parliament's website lists the AllParty Parliamentary Caucus on Ending Malaria, the All-Party Parliamentary Group for the Prevention of Genocide and Other Crimes against Humanity, the BRIC Trade Development Caucus, the Canadian Association of Parliamentarians for Population and Development, the Committee against the Commercial Sexual Exploitation of Children and Youth, Parliamentarians for Global Action, the Parliamentary Friends of Burma, the Parliamentary Friends of Falun Gong, the Parliamentary Friends of Tibet, and the Save Darfur Parliamentary Coalition.16 Other groups are not included on the list, however. For example, we learned about the existence of the Pro-Life Caucus, the All-Party International Conservation Caucus, the HIV/AIDS and TB Parliamentary Caucus, the Parliamentarians for Nuclear Non-Proliferation and Disarmament, and the Parliamentary Outdoors Caucus. According to one of our survey respondents, LMOs in Canada “have little or no public profile” and their “influence is very hard to track.” The same holds true for LMOs in India and Australia. The most widely known LMO in Australia is the Parliamentary Christian Fellowship, while other groups with less of a public profile include the Cross-Party Parliamentary Group for Drug Law Reform, the Cross-Party Parliamentary Group on Amnesty, the Cross-Party Parliamentary Group on Primary Care and Public Health, and the Australian Parliamentary Group on Population and Development. Page 62 → Parliamentary Forums, as LMOs are called in India, seek to provide “members with information and knowledge on specific issues of national concern and [to assist] them to adopt a result-oriented approach towards related issues.”17 According to the Centre for Legislative Research and Advocacy,18 Parliamentary Forums can effectively function as an informed action group in the policy making process. As representatives of people and as policy and decision makers overseeing government functions, Parliamentarians are in a unique position to engage with policy making. Theme based Forums provide Parliamentarians, cutting across party lines, a common platform to come together and discuss the different views and concerns and engage with experts and civil society groups and consequently bring convergent knowledge to influence the policy making process. The Forum Secretariat can provide access to adequate information on issues and necessary assistance to carry out important parliamentary functions, which will greatly benefit the Members of Parliament. [The objectives of Parliamentary Forums are] to create a platform for MPs to exchange views and information with civil society groups and experts [and] to provide assistance to Forum Members in raising issues and concerns in the Parliament. While Parliament's website lists some forums and their members (Water Conservation and Management; Youth; Children; Population and Public Health; and Global Warming and Climate Change), the Indian Medical Parliamentarians’ Forum, the Parliamentarians’ Group on TB, and the Parliamentarians’ Group on the Millennium Development Goals are not mentioned.19 Category 2: Proportional Representation or Mixed System, Large Legislature, Large Effective Number of Parties

While all countries with majoritarian electoral institutions have LMOs, that statement does not hold true for countries with either proportional representation or mixed systems: Only 19 of the 39 countries with such electoral systems have LMOs. Nine of 20 countries with proportional representation systems (Brazil, Finland, Indonesia, Israel, New Zealand, Norway, Poland, South Africa, Sweden) have LMOs, as do 10 of 19 countries with modified electoral systems (Belgium, Chile, Estonia, Germany, Hungary, Italy, Japan, Lithuania, South Korea, and Switzerland). Page 63 → The two major determinants of whether countries with nonmajoritarian electoral institutions have LMOs are the size of the legislature and the effective number of parties. As both factors increase, they make the existence of LMOs more likely. Table 3.4 provides an overview of countries with small legislatures (fewer seats than the average chamber in our sample, 287 seats), a low effective number of legislative parties (less than the average number in our sample, 4.06). LMOs exist in all four countries with legislatures that are larger than average and that have more than the mean effective number of legislative parties. Two of the four have LMOs that are institutionalized. LMOs in Brazil are quite prominent and called frentes parlamentares (parliamentary fronts); they are sometimes referred to informally as bancadas (blocs). The parliament's website lists 99 frentes parlamentares, with groups including: for sports (Frente Parlamentar do Esporte); against crack (Frente Parlamentar contra o Crack); in defense of the domestic wholesale markets (Frente Parlamentar em Defesa das Centrais de Abastecimento Interno); in defense of community universities (Frente Parlamentar em Defesa das Universidades Comunitárias); in defense of the Brazilian nuclear program (Frente Parlamentar Mista em Defesa do Programa Nuclear Brasileiro, or PNB); for tobacco control (Frente Parlamentar pelo Controle do Tabaco); for fighting corruption (Frente Parlamentar de Combate à Corrupção); in defense of the Brazilian pharmaceutical industry (Frente Parlamentar em Defesa da Indústria Farmacêutica Brasileira); and in defense of the rights of the child and adolescent (Frente Parlamentar em Defesa dos Direitos da Criança e do Adolescente). 20 Frentes parlamentares have been a feature of legislative politics in Brazil for many decades, and some groups from past legislative terms are quite well known, such as the Evangelical Parliamentary Front (Frente Parlamentar Evangélica or Bancada Evangélica) and the Parliamentary Front for Agrarian Support (Frente Parlamentar de Apoio à Agropecuária or Bancada Ruralista), a powerful collection of landowners in the late 1980s and early 1990s. LMOs in Brazil work to facilitate coordination and information exchange, both within the legislature and in collaboration with outside groups that participate in meetings and provide input and advice. These groups ultimately seek to set the legislative agenda, lobby the executive branch for resources to support their cause, and directly influence legislation. LMOs in the Sweden are less institutionalized than is the case in Brazil, but they nevertheless have to follow certain internal rules if they wish to use Page 64 → the facilities and equipment of the Riksdag (ECPRD 2008). Examples of LMOs, or sammanslutningar (associations; some groups are also known as networks or alliances), include the Swedish Support Committee for Human Rights in Turkey and the Swedish-Tibetan Parliamentarian Alliance. While several groups have more of a social aspect (for example, the Beatles Society), others are explicitly policy-driven (such as the Children's Group, where questions and solutions regarding child care and education are discussed). Germany and Indonesia are the two countries with large legislatures and a high number of parties where LMOs are not institutionalized. In Germany, LMOs are not a major part of the legislative process, but a handful of parliamentary groups (Parlamentarische Gruppen, sometimes referred to as Parlamentsgruppen or Parlamentariergruppen) do exist. Among the more prominent examples are parliamentary groups on aviation and space (Parlamentsgruppe Luft- und Raumfahrt), rail traffic (Parlamentsgruppe Schienenverkehr), free flowing rivers (Parlamentarische Gruppe Frei fließende Flüsse), inland water transportation (Parlamentarische Gruppe Binnenschifffahrt), Europe-Union (Europa-Union), and classic languages (Parlamentariergruppe Klassische Sprachen).

Indonesia's LMOs are variously called caucuses, groups, or committees. Examples include the Women's Caucus; the Palestine Issues Group; the Committee for Parliamentarians from Member Countries of the Organisation of Islamic Conference; the Committee on the Millennium Development Goals; and the Working Group on the Environment. There are also a few regionally defined LMOs, such as the Sumatra Caucus. Finally, a caucus of members from “conflict regions” has been established, bringing together deputies from provinces in eastern Indonesia that have been affected by internal civil conflict, among them Maluku, North Maluku, Papua, and West Irian Jaya (Sherlock 2005, 37).

Category 3: Proportional Representation or Mixed System, Large Legislature, Small Effective Number of Parties Six of the 11 countries (55 percent) with nonmajoritarian electoral systems, large legislatures, and a smaller than average effective number of parties have LMOs. In two of the six, Japan and Poland, LMOs are institutionalized. In Japan, LMOs are called giin renmei, or Diet members’ leagues. There are hundreds of giin renmei, including groups on Amnesty International, viral hepatitis prevention and treatment, malpractice prevention, disarmament Page 65 → and nonproliferation, military pensions, road safety, parliamentary reform, promoting bicycle use, sports, tax reform, flood control, martial arts, and the flower industry. While some groups exist in name only, others are quite active and capable of influencing legislation. Some giin renmei, for example, helped push for electoral reform in Japan in the 1990s (Wolfe 1995, especially 1064–70). LMOs in Poland are called zespoły (groups).21 There are currently 57 LMOs in the Polish Sejm, including groups on energy (Parlamentarny Zespół ds. Energetyki), the tradition of independence (Parlamentarny Zespół ds. Tradycji Niepodległościowych), cooperation with NGOs (Parlamentarny Zespół ds. Współpracy z Organizacjami Pozarządowymi), history lovers (Parlamentarny Zespół Miłośników Historii), enterprise (Parlamentarny Zespół Przedsiębiorczości), free Cuba (Parlamentarny Zespół “Wolna Kuba”), metropolitan Warsaw (Zespół Parlamentarny na rzecz Metropolii Warszawskiej), and autism (Parlamentarna Grupa ds. Autyzmu). The Parliamentary Women's Group (Parlamentarna Grupa Kobiet) has a particularly long tradition (ECPRD 2008), while the “Parliamentary Group to Investigate the Crash of TU-154 MZ on 10 April 2010” (Parlamentarny Zespół ds. zbadania przyczyn katastrofy TU-154 M z 10 kwietnia 2010 r.) is the largest group, counting 157 of the legislature's 460 deputies.22 The average group has 27 members. Hungary, Italy, South Africa, and South Korea have LMOs that are not institutionalized. Hungary has a total of 18 LMOs, called csoportok (Groups). Among them are a veterinarians’ group, a teachers’ group, the sober life group, a 1956-ers’ circle, a children and youth group, and a group of MPs who lost their liberties for political reasons in the past. Most of these groups were formed immediately after the fall of communism. Only one or two additional groups were set up during subsequent legislative terms, indicating a members’ declining inclination to form crossparty groups based on profession, geographical area, or other considerations. Nevertheless, their political and professional influence can be quite significant according to one of our survey respondents. In Italy, the parliament features a number of very informal groups, usually labeled the Parliamentary Friends of…(Parlamentari Amici del…). Examples include the friends of the mountains; of tennis; of the boat/sailing; of philately; of contemporary art; of hunting and fishing; and of rugby. There is also a “Federalist Intergroup for the European Constitution” and an organ donors group. South Africa has recently seen the creation of a small Page 66 → number of LMOs—the Multiparty Women's Caucus, the Inter-Faith Council, and the Sports and Recreation Council. Finally, South Korea has a number of “clubs” of legislators on such issues as national parks, the educational environment, bicycles, natural science and engineering, and the Korean alphabet.23 Category 4: Proportional Representation or Mixed System, Small Legislature, Large Effective Number of Parties Half of the countries (8 of 16) that have nonmajoritarian electoral institutions, smaller than average legislatures, and a larger than average effective number of parties feature LMO systems; 5 of these countries (62.5 percent) have institutionalized systems. Estonia, for example, has 30 LMOs, called ühendused (deputies’ associations or

unions).24 These groups have attracted between 6 and 44 of the legislature's 101 members, with an average of 14. The largest LMO is the Children's Rights Support Group (laste õiguste toetusrühm), and others include the National Heritage Association (muinsuskaitse ühendus); the Hunters Association (jahimeeste toetusrühm); the Civil Society Support Group (kodukandi toetusrühm); the Space Research Support Group (kosmoseuuringute toetusrühm); the National Minorities Association (vähemusrahvuste saadikute ühendus); and the Association of Islands (saarte ühendus). Finland, too, has about 30 LMOs, or kerhot (clubs) (ECPRD 2008), ranging from social groups (the Parliamentary Motorcycle Club, Sports Club, Karaoke Club, Fine Arts Club, Chess Club, and Bible Group) to explicitly political ones (the Parliamentary Group for Animal Protection, the Parliamentary Human Rights Group, the Parliamentary Group for the Well-Being of Older People, and the Parliamentary Group on Poverty). Kerhot pursue either specific or general political goals and work to enhance cooperation and knowledge exchange between parliamentarians as well as between parliamentarians and outside experts. Israel's Knesset has 88 LMOs, called shdula (members’ lobbies). Several lobbies have only a single member, while the largest, the Lobby for Local Government, includes 51 of the Knesset's 120 deputies; the average lobby size is 5. Examples of lobbies are the Agricultural Lobby; the Lobby for Green Businesses; the Lobby for the Promotion of Football in Israel; the Christian Allies Caucus; the Lobby for the Struggle against Racism; the Page 67 → Lobby for Zionist Religious Education; the Lobby for Saving the Dead Sea; and the Lobby for Independent Farms. In the Lithuanian Seimas, LMOs are called parlamentinės grupės (parliamentary groups).25 Examples are the Parliamentary Group of Dzūkija Friends (Dzūkijos bičiulių parlamentinė grupė); Parliamentary Prayer Group (maldos parlamentinė grupė); Parliamentary Group of the Friends of Šiauliai Area (Šiaulių krašto bičiulių parlamentinė grupė); Parliamentary Group for Democratic Belarus (parlamentinė grupė „Už demokratinę Baltarusiją”); Parliamentary Group for Human Dignity (parlamentinė grupė „Už žmogaus orumą”); Group for Society Development and Reproductive Health And Rights (Visuomenės plėtros, reprodukcinės sveikatos ir teisių parlamentinė grupė), and the Temporary Group of Young Parliamentarians (laikinoji jaunųjų parlamentarų grupė). The Temporary Group of Young Parliamentarians has only 8 members, but the Parliamentary Group for Democratic Belarus has 42 members, or about 30 percent of all Seimas deputies; other groups fall in between.26 Their topics and activities are also quite diverse. Some take part in conferences, seminars, and public debates, while others maintain ties with NGOs in Lithuania and abroad. The Parliamentary Group for Democratic Belarus communicates with the Belarusian opposition, for example, while the Group on Lisbon monitors the implementation of the European Union's Lisbon Strategy. It analyzes the activities of the EU, the Lithuanian government, and the Seimas; organizes public discussions; and brings together parties interested in the Lisbon Strategy (ECPRD 2008). Some LMOs in Switzerland, where they are called parliamentary groups,27 are extremely active and organize talks, trips, and various other activities that highlight their interests (ECPRD 2008), including professional networking and the preparation of parliamentary motions. As of March 2010, the Federal Assembly had 58 parliamentary groups (not counting parliamentary friendship or liaison groups with particular geographic regions), among them groups for the disabled (Parlamentarische Gruppe für Behindertenfragen; intergroupe parlementaire pour les questions relatives aux personnes handicapées); biodiversity (Parlamentarische Gruppe Biodiversität und Artenschutz; intergroupe parlementaire Biodiversité et gestion des espèces); Digital Sustainability (Parlamentarische Gruppe Digitale Nachhaltigkeit; intergroupe parlementaire Informatique durable); Drug Policy (Parlamentarische Gruppe Drogenpolitik; intergroupe Page 68 → parlementaire Politique des drogues); Renewable Energy (Parlamentarische Gruppe Erneuerbare Energien; Intergroupe parlementaire Énergies renouvelables); Women (Parlamentarische Frauengruppe; intergroupe parlementaire Femmes); Children and Youth (Parlamentarische Gruppe Kinder und Jugend; intergroupe parlementaire Enfants et adolescents); and Fire Services (Parlamentarische Kerngruppe Feuerwehr; intergroupe parlementaire Sapeurs-pompiers).28 In Belgium, Chile, and Norway, LMOs are not institutionalized. In Belgium, LMOs are labeled interparliamentary groups, and examples include the Interparliamentary Group for Tibet, the Interparliamentary Group on Breast

Cancer, the Interparliamentary Group for Peace in the Western Sahara, and the Interparliamentary Group “Handicap.” Chile has bancadas on a variety of issues such as regionalism (Bancada Regionalista), fisheries (Pesquera), agriculture (Agricola), small and medium-sized enterprises (Bancada PYMES), progressives (Bancada Progresista), and the environment (Bancada Verde) (see Navarrete Yánez 2010 for more details). Finally, Norway's parliament features a number of informal grupper (groups), such as the Friends of Israel, Friends of Palestine, Support for Children, and the Tibet Group. There is also a sports club, a group advocating Nynorsk (one of the country's two official written languages), and a Christian group. Category 5: Proportional Representation or Mixed System, Small Legislature, Small Effective Number of Parties The last category in table 3.4 is legislatures that have a smaller than average number of members, that have a small effective number of parties, and where members are elected using either a proportional representation or modified electoral systems. Of the eight countries that fall into this category, only New Zealand has LMOs. These groups, known as (cross-party) parliamentary (lobby) groups or caucuses, are neither institutionalized nor very influential. Among the relatively small number of groups are the Cross-Party Parliamentary Group on Burma, the Parliamentary Internet Caucus, and the Parliamentary Lobby Group for Tibet. One country with nonmajoritarian electoral institutions, a small legislature, and a small number of parties is Bulgaria. While we should not expect its national legislature to have LMOs, efforts to encourage the establishment of such groups have occurred, though they failed. In fact, in 2006, the U.S. government, through the U.S. Agency for International Development (USAID), Page 69 → supported and funded the creation of a local government caucus with the goal of initiating, discussing, and facilitating the passing of a constitutional amendment in support of financial decentralization to local government, according to one agency official.29 In the context of this initiative, 11 members of the Bulgarian parliament went on a USAID-funded study tour of the U.S. Congress in July 2006. Their visit included “meetings, reviews, and an advanced discussion on the existing models of multiparty groups (caucuses).” Furthermore, participants were “introduced to U.S. examples of informal organizations of members of the U.S. Congress established to discuss issues of public interest and conduct legislative research and policy planning for their members” (USAID 2006).30 The local government caucus was created and achieved its ultimate policy goal of passing the constitutional amendment granting greater levels of decentralization. USAID hoped that this success would not only consolidate this particular caucus but also spur the creation of additional groups. The agency saw three major benefits in LMOs:31 the establishment of policy coalitions comprised of members of parliament from different parties, as well as interested outside actors; access to information;32 and the advancement of key policies. However, not only did no other caucuses form in Bulgaria, even the local government caucus disbanded after achieving its core objective. Four key reasons account for the caucus's failure.33 First, Bulgaria's political elite lacked an agenda, both in a general sense and with regard to specific policy issues. As a result, there were no clearly defined political goals to pursue in the context of LMOs. Second, Bulgarian politics are highly partisan, to the extent that lawmakers were not prepared to join cross-party groups for fear that doing so would negatively affect their relationship with party leaders.34 Third, there was no outside support for LMOs in Bulgaria. USAID had provided support for the local government caucus, but once the agency pulled out, no one stepped in to fill this role because of the country's generally weak civil society. Finally, because Bulgaria does not have a majoritarian electoral system, the link between constituents and representatives is exceedingly weak and constituency service is almost irrelevant. Hence, legislators had no electoral incentive to engage in LMOs. Given the combination of these factors, it is not surprising that LMOs floundered in Bulgaria, but the case nevertheless provides a number of interesting insights. First, USAID saw the creation of LMOs as a means to achieve the same benefits that we ascribe to these groups, the most important of Page 70 → which is access to information through the establishment of relationships that connect actors across political parties as well as legislators and interested outsiders. Second, actors outside the legislature were important for the creation and

maintenance of LMOs, which supports our intuition that the external dimension of LMOs is critical in both theoretical and substantive terms. Third, the example confirms that the variables we identify as relevant to the creation and maintenance of LMOs (electoral system, legislature size, number of parties) do not tell a complete story. Future research on LMOs clearly must take into account additional factors (such as the intensity of partisanship and perhaps even the political culture in a legislature).

Comparative Characteristics of LMOs Some legislatures register their LMOs and have rules in place to regulate their operations. Westminster's APGs, for example, are subject to the most detailed set of rules in our sample of countries.35 Other legislatures place no constraints on LMOs. LMOs in Hungary, Italy, and South Korea, for example, are completely informal and are not regulated by the parliaments’ rules of procedure. Similarly, in Indonesia, New Zealand, and Belgium, the parliaments do not catalog or keep track of LMOs. Even some legislatures with a higher level of institutionalization have only minimal regulations in place. The Israeli Knesset emphasizes that it “does not interfere” in the activities of members’ lobbies, and Knesset members “are permitted to form lobbies without any explicit permission.”36 In Poland, the only provision concerning LMOs is that zespoły leaders must inform the marshal of the Sejm about their groups’ composition and internal statutes.37 Similarly, ühendused in the Estonian parliament are required to submit their names, objectives, and the names of the members and their party affiliations to the parliament.38 Other legislatures also register their LMOs but do not interfere with groups that choose to operate without formal recognition. Parliamentary groups in Switzerland, for example, are supposed to notify Parliamentary Services of their formation and of their members, but respondents to our survey indicate that informal groups also exist. Delegates may choose to form such groups because parliamentary groups are required to be open to all members, 39 and some group may not wish to permit all legislators to join. A similar dynamic may explain the existence of informal LMOs in Page 71 → Brazil, where frentes parlamentares must be comprised of least one-third of members of the Federal Legislature.40 Groups that do not meet this requirement may thus remain informal, even if their activities resemble those of registered frentes parlamentares. Remaining informal may have costs, however, because in some legislatures (e.g., in Switzerland and France) only registered LMOs have access to parliamentary facilities, such as meeting rooms.41 Yet such restrictions may only have a limited practical impact, because individual members or small groups can simply reserve conference rooms (e.g., in Italy and Canada). Nonetheless, a few legislatures have regulations about the use of parliamentary infrastructure and facilities. In Brazil, the rules governing LMOs explicitly state that rooms in the legislature can be used to hold meetings only if they do not interfere with the chamber's formal operation.42 In the Israeli Knesset, lobbies may hold meetings in members’ offices or in the offices of the parliamentary groups but may not use committee rooms to prevent “any misconception that the lobby meeting is official Knesset business.”43 Aside from granting access to physical space, very few legislatures provide any kind of direct support to their LMOs. Exceptions are groupes d'études in France and kerhot in Finland. In the French National Assembly, LMOs have access to small financial subsidies for specific purposes or projects, such as meetings, travel, publishing reports, or employing part-time staff. They are also assigned administrative staff. The Finish Eduskunta goes even further and offers “allowances” to registered kerhot that have written rules and elected officials. To be eligible for allowances, groups must present a plan of action, a draft budget, and a copy of their present bank account to the Accounts Office of the Administrative Department. Groups that receive funding are required to present a financial statement and an annual report of their activity to the Accounts Office, and their spending is strictly controlled. In 2007, a total of 140,000 Euros was distributed across various groups, with annual support ranging between 500 and 65,000 Euros. The largest sum was received by the largest kerho, the Sports Club, and its many subdivisions. Most groups do not ask for allowances and do not receive any support from the Eduskunta (ECPRD 2008).

For the most part, LMOs do not incur great financial costs for their operations as long as they have access to meeting rooms. Yet to cover basic expenses, some groups in Brazil, Japan, and the United Kingdom charge membership fees. In Westminster, there is a five pound annual limit on membership fees for Members of Parliament, while no such limit applies to Page 72 → individuals or organizations from outside Parliament. Outside financial or material support must, however, be registered with Parliament. Financial support includes any monetary donation, including membership fees, while material support comes in the form of sponsoring of events (for example, covering the cost of receptions or dinners), nonmonetary gifts (subsidizing travel expenses, printing costs, tickets for cultural or sporting events), and secretarial or research support. Both types of support must be disclosed within 28 days of their receipt if their total value equals or exceeds 1,500 pounds in one calendar year. When a parliamentarian gains personally from a benefit received in connection with an APG, these benefits must also be disclosed in accordance with rules governing the Register of Members’ Financial Interests (“House of Commons” 2010). The extent of outside support for APGs is quite large: more than 1.6 million pounds were donated by outsiders in 2010; 280 groups received support from corporations, lobbyists, or interest groups; and at least 45 groups were run by professional public affairs companies (Ball 2011b).44 Most LMOs in most legislatures are cross-partisan, even if no formal rules specify that they must be, as is the case in Finland. Similarly, Switzerland's rule that parliamentary groups must be open to all deputies means that registered groups have a de facto cross-partisan makeup. In Brazil, cross-partisanship in formal frentes parlamentares is ensured by the requirement that registered groups comprise at least one-third of members of the Federal Legislature since no single party holds that many seats. Even Norway, Italy, and other countries with noninstitutionalized LMO systems appear to feature only cross-party groups. In contrast, Israel and Japan do not require cross-partisanship and some single-party groups exist. In France, some groupes d'études have only a single member, making them by definition single-party. Larger groups, however, are bound by rules about the partisan distribution of vice presidencies based on the size of both the study group and the party.45 Westminster again has clearly specified rules about cross-partisanship in LMOs. Most important, all registered groups must be open to members of more than one political party. Moreover, the vast majority of groups in Westminster are not only cross-party but genuinely all-party groups (“House of Commons” 2010). In addition to simply registering with the Office of the Parliamentary Commissioner for Standards, LMOs may also seek inclusion on the “Approved List.” To receive approval, however, groups must present the names of 20 qualifying members—10 from the government Page 73 → party or parties, and 10 from parties that are not in government, with at least 6 of them from the main opposition party. Despite these additional rules, most groups apply for inclusion on the Approved List because it provides important privileges, most notably the right to advertise their meetings on the All-Party Notices (a list compiled and issued by the government whips every week when the House of Commons is sitting) and priority when booking rooms in Parliament. In most countries, LMOs are quite explicitly groups of legislators that are formally closed to outsiders. In other words, interest groups, businesses, or nongovernmental organizations and their representatives cannot become members even if they regularly participate in LMO activities. Exceptions exist in Finland, where civil servants may become members of kerhot, and in Westminster, where anyone, including ministers and outsiders, may join at the discretion of the group (although at least one APG officer must be a member of the House of Commons).46 Some groups even grant voting rights to outsiders who are not members of the House of Commons or the House of Lords (in which case they are called associate parliamentary groups instead of all-party parliamentary groups [“House of Commons” 2010]).

Conclusion While exceedingly fruitful in terms of providing systematic, comparative, descriptive data on LMOs, our survey did not provide a great amount of detail on the activities and functions of LMOs. In terms of the former, our respondents indicate that LMOs organize meetings, seminars, and colloquiums (France, Norway, Switzerland); issue reports, press releases, and opinions (Estonia, France); and engage in occasional travel (France, Switzerland). Respondents list among LMOs’ functions socializing and networking (Canada, Finland,

Switzerland), information gathering and education (Brazil, France, India), and cooperation and knowledge exchange between legislators and outside experts, scholars, lobbyists, and NGOs (Finland, France, India, Italy, Norway). Furthermore, respondents emphasize LMOs’ attempts to set the agenda (Brazil, India), to engage in policy coordination and consensus building (Brazil, France, Switzerland), and more generally to pursue political goals through specific policy actions (Brazil, Finland, France, India, Israel, Italy, Japan, Switzerland). In the following three chapters, we delve more deeply into questions Page 74 → that relate to the activities, functions, and benefits of LMOs in an effort to flesh out what our survey respondents are suggesting. To achieve this objective, we present two nuanced case studies that investigate LMOs in the European Parliament and the U.S. Congress. In so doing, we move beyond the more exploratory context of this chapter and present generalizable insights about LMOs as circuits of information exchange, arenas of political coordination, and opportunities to signal preferences and priorities to constituents. Based on the findings of this chapter, we expect that our evidence will reveal the informational value of LMOs for legislators in both the European Parliament and Congress, that coordination will be more prominent in the European Parliament than in Congress (because crosspartisan coalitions are required for the passage of legislation), and that signaling is more common in Congress than in the European Parliament (because the majoritarian electoral system encourages the use of caucuses as a means to signal preferences and priorities to constituents). Our survey of LMOs in 45 advanced industrial democracies provides a plethora of information that has not previously been collected in a systematic, comparative framework. Much remains to be discovered about LMOs and their roles in legislative politics, however. The following chapters take a series of first steps in this direction.

Page 75 →

CHAPTER 4 Building the Case: Legislative Member Organizations in the European Parliament and the U.S. Congress This chapter introduces two legislatures whose legislative member organizations (LMOs) we analyze as case studies to investigate their role in legislative politics. Our two cases are intergroups in the European Parliament and congressional member organizations (CMOs), usually referred to as caucuses, in the U.S. Congress.1 We use these cases in three empirical chapters to evaluate in detail our theoretical propositions about the value of LMOs for the establishment of relationships between legislators and the diffusion of policy-relevant information in legislative politics. Our two cases lend themselves to a comparative research design because they share core institutional structures, such as party-based politics and active standing committee systems, but provide for variation when it comes to a number of features of interest. Indeed, we observe this variation in particular with two of the three variables identified in the cross-national analysis in chapter 3 as explaining the presence of LMOs in industrialized liberal democracies: the electoral system by which representatives are selected and the number of parties. Members of Congress are selected using a single-member district plurality electoral system; as a result and as one would expect, there are only two legislative parties. Members of the European Parliament, in contrast, are elected using proportional representation systems across the 27 EU member states2 and sit in seven transnational party groups. Some variation exists with regard to the third explanatory variable (the size of the legislature) since the EP (with its 736 members)3 is quite a bit larger than Congress (with 435 members). Both chambers are larger than the average legislature we consider in chapter 3, however. Nonetheless, these systematic differences allow us not only to confirm that our Page 76 → key propositions about the role of LMOs in legislative politics hold across legislatures but also to fine-tune our findings and provide nuanced insights into the existence and operation of LMOs. To briefly restate our argument, we propose that LMOs allow legislators to establish internal information networks based on policy priorities that transcend the boundaries imposed by partisanship, opposing ideology, and committee jurisdictions. These networks are composed mainly of weak, bridging ties between legislators and their offices (Granovetter 1973, 1974). Information flows more efficiently through these networks because much of the information circulating in a social network composed of strong ties tends to be redundant as a consequence of the frequent interactions between network members. LMO ties transcend closely tied social circles that would otherwise be isolated from each other because they span structural holes in the legislative network by cutting across party and committee lines (Burt 1992, 2000, 2004). Hence, LMO ties help to efficiently diffuse socially distant ideas and information that are relevant to legislators’ policy choices, thereby helping to mitigate the informational collective action dilemma prevalent in legislative politics, where a great demand for political and policy information is met with an insufficient supply. One of the key characteristics of LMOs that allows them to play this role is their voluntary nature—legislators can join (or depart) at will. This reality entails that LMO networks are endogenous from the point of view of individual legislators, since ties are formed as a result of their voluntary participation in one or more LMOs. It also means that the issue scope of LMOs is open because legislators can set up LMOs on any issue they consider to be a policy priority. Finally, the voluntary nature of LMOs makes participation cheap for two reasons. First, voluntary membership means that most LMO ties are weak ties, as participation in LMOs for most legislators is too limited to provide for the establishment of strong ones. Weak ties are cheap, however, because creating and maintaining them does not require much effort. Second, LMOs are low-cost for most participants because voluntary participation means that members are invited to free-ride on the informational benefits LMOs offer. The costs associated with establishing and running LMOs are borne by legislators who have a strong enough personal stake in the LMO's issue to voluntarily expend some of their limited resources and/or by outside advocacy

organizations that support an LMO's cause. The relationships between LMOs and outside advocates are quite extensive, Page 77 → and we highlight two aspects of outsiders’ involvement in LMO systems: the privileged relationships they maintain with LMOs and their leaders, which consist of outsiders providing legislative subsidies to lawmakers who share their policy priorities, and the nature of information exchange between insiders and outsiders. The two are closely linked in the sense that the privileged access LMOs grant to particular outside advocates provides an incentive structure that increases the quality of information LMOs receive from outside groups, for three reasons. First, LMOs allow legislators to triangulate information they receive from various sources; second, LMOs feature a high level of internal social embeddedness, where information about the transgressions of a member (such as an outside advocate providing misleading information) is likely to become common knowledge; and third, legislators can credibly threaten to revoke the privileged access they grant to outside advocates who provide redundant, unreliable, or misleading information. For these reasons, outside advocates face strong incentives to provide what we call high-utility information—information that is research-based, reliable, and presented in an easily digestible format. In this sense, LMO ties between insiders and outsiders encourage the flow of high-utility information into the legislature; the LMO network, in turn, efficiently diffuses this information inside the legislature and allows lawmakers to make better policy choices. Both case studies provide strong evidence in support of our theoretical account of LMOs as social networks that facilitate the exchange of high-utility information between legislative offices. We show, first, that LMO network structures offer the potential to facilitate the flow of information inside the legislature, most importantly across party and committee lines. Second, our qualitative data demonstrate the actual informational value of LMOs for the overwhelming majority of our respondents in the EP and Congress. We also carefully consider the extent to which LMOs fulfill the other two functions highlighted throughout this book, political coordination and the signaling of preferences and priorities. Our findings are again in line with our expectations: There is more evidence for coordination across parties and committees in the EP than in Congress because cross-partisan cooperation is necessary for passing legislation. Meanwhile, LMOs fulfill more of a signaling function in Congress than in the EP because the majoritarian electoral system in the United States provides incentives to signal preferences and priorities to constituents. Page 78 →

The Data We rely on two principal sets of data for each case study. The first is membership information in LMOs covering two legislative terms in the EP (the EP's 6th term [2004–9], or EP6, and its 7th term [2009–14], or EP7) and three sessions of the House of Representatives (2005–7, 2007–9, 2009–11, known as the 109th, 110th, and 111th terms of Congress, respectively). We use these data primarily to analyze LMO membership using a variety of social network analysis tools, as our principal theoretical propositions revolve around relationships and network structures. We also analyze these membership data using more conventional quantitative methods, most importantly to identify the factors that account for the number of LMOs legislators join. For these analyses, we collect attribute data (e.g., party affiliation, committee assignments, gender, and leadership information) for lawmakers in both legislatures. Gathering LMO membership information for the EP's 7th term (2009–14) was straightforward, as membership information is readily available on the EP website.4 That information was not published for the 2004–9 term, however. Eighteen of the 24 formal groups supplied this information when we requested it; for the remaining 6 groups, we relied on the membership lists they provided to the Conference of Presidents at the beginning of the legislative term when they requested formal intergroup status.5 In comparison, collecting data on the membership of individual caucuses in the U.S. Congress is a much more tedious process because of the exceedingly large number of caucuses and because only the most formalized and well-supported caucuses maintain online membership lists and there is no comprehensive register of caucuses. A

list of registered caucuses appears on the website of the Committee on House Administration for the U.S. House of Representatives, where they are listed as CMOs; however, caucuses are required to reregister with the Committee on House Administration each Congress, and we have discovered hundred of caucuses that do not consistently register with the committee.6 Moreover, the committee does not keep membership lists—only names of caucuses with contact information for the cochairs. Therefore, to obtain the membership information we seek, we relied on the Congressional Yellow Book, a directory of information about each member of Congress. The directory is published quarterly, compiles its information largely from members’ self-reports, and lists caucus memberships. Using the winter directory for each even-numbered Page 79 → year (representing the last possible published directory for each Congress), we compiled the membership data for the 109th–11th Congresses by recording each individual representative's caucuses. We know that members continually and fluidly join caucuses throughout their terms, so by collecting the data at the end of the term we seek to maximize the probability of capturing the complete caucus memberships for each term. While collecting membership information in congressional caucuses using the Yellow Book is the most comprehensive method of which we are aware to gather these data, it is likely moderately error prone. First, we know that significant differences exist between directories published in different seasons of the same Congress. A member might report being a member of the Bike Caucus, for example, in the spring, summer, and winter editions of the Yellow Book but not in the fall. It seems probable that rather than indicating a change in membership, the omission reflects the member's failure to indicate participation in the fall survey. The Yellow Book surveys are probably completed by staffers and interns and, as a result, include some errors. Second, we hired a team of research assistants (primarily undergraduates) to record these data, one member at a time, by hand, a process that is likely to produce at least a small number of errors. To check the extent to which this was a concern, we have performed coding checks on 14 percent of randomly selected members and found an error rate of less than 1 percent. We are therefore confident that measurement error in our dataset is minimal. Moreover, we assume that surveying the caucus memberships as reported in one directory (of the eight published for a single term of Congress) provides a conservative estimate of total caucus memberships. A complete survey of all the directories from a term of Congress would likely produce measurement error in a direction where we would overestimate legislators’ memberships and thus potentially bias empirical results in favor of our theoretical propositions. Finally, we have no reason to believe that using a different edition of the directory would result in substantive findings that deviate from the patterns we identify. We took a number of other steps to reduce the amount of error introduced into the data. At the end of the coding process, we eliminated caucuses in which only a single member of Congress indicated membership—a caucus of one member is not a caucus. We eliminated interparliamentary groups, because they do not satisfy the definition of LMOs we established in our theoretical approach and in our cross-national data. We also eliminated caucuses that are state delegations, or state party delegations, since Page 80 → these subgroupings of members are already designated as connected or affiliated legislators by their state and party. We eliminated groups that are party organizations or leadership organizations. For example, a Speaker's Task Force for a Drug-Free America is not a caucus but a party leadership group. Such groups do not meet the definition of being a purely voluntary group and receive some resources from the party leadership, which caucuses do not. We went to great lengths to identify similar-sounding groups and ascertain whether they were separate entities. Since members have self-reported their caucuses, there were often differences in the ways in which caucus names were listed. For example, a member who indicated membership in the “Congressional Caucus to Combat Alzheimer's Disease” was determined to actually be a member of the “Bipartisan Congressional Task Force on Alzheimer's Disease” because the former had only one member and the latter is a long-standing caucus. Members occasionally would indicate that they were members of a subcomponent of a caucus. For example, the Northeast-Midwest Congressional Coalition has an associated Upper Mississippi River Task Force, but members of the latter are considered members of the former. Finally, with a few noted exceptions, we excluded individual members of Congress who did not serve full terms during a single Congress because of death, resignation, or special election.7 Our second major source of information consists of extensive interview data from 86 conversations with legislators, legislative staff, and outside advocates in both Brussels and Washington, D.C. These interviews were

semistructured, consisting of a series of core questions supplemented by follow-up and in-depth discussion about different subjects. We rely on both sets of data to test our expectations about the role and benefits of LMOs in legislative politics, thus increasing our confidence in our findings. We conducted 41 interviews in and around the EP between March 2009 and June 2010. Our respondents include 21 members of the European Parliament (MEPs), 8 legislative assistants (personal staff), 2 party group advisers (officials who work directly for a party group in the EP), and 10 representatives of outside interest group organizations affiliated with intergroups. In Washington, we conducted 44 interviews with congressional staffers and advocacy organizations between May 2009 and June 2010.8 Of these interviews, 37 were with members’ personal staff, while the remainder were with representatives of outside interest groups associated with particular caucuses. Finally, we conducted 1 interview with a representative from the U.S. Agency for International Development. Page 81 → The interview data were analyzed through careful coding using the qualitative data analysis software NVivo. In a first step, three researchers open-coded three randomly selected interviews. The resulting codes were consolidated into single sets of codes for both the EP and Congress interviews. Using these sets of codes, two researchers coded another 20 percent of all interviews to ensure high levels of coding consistency before one of the authors coded the remaining interviews. For the EP interviews, coding comparisons in NVivo establish that agreement between the two coders is very high—98.68 percent. The lowest level of agreement for a single code in a single interview is 81.69 percent, and for only eight of 293 codes (across all interviews) is the agreement below 90 percent. For the Congress interviews, intercoder reliability is 99.47 percent. The lowest agreement rate for a single code in a single interview is 77.51 percent, and in only 30 of 316 codes is the agreement rate below 90 percent. This chapter is largely descriptive and seeks to introduce the reader to our two legislatures and their LMOs; this is an important contribution, however, given the dearth of previous research on LMOs, especially in the case of the EP. Indeed, we present a plethora of new information about the development and structure of LMOs, including the results of negative binomial regression analyses predicting legislators’ proclivity to join LMOs.

The European Parliament: From Talking Shop to Co-Legislator The European Parliament is the first and only example of a directly elected international, or transnational, parliament. Since 1979 European Union (EU) citizens have elected the EP every five years. Together with the Council of Ministers, comprised of ministers of the national governments of the EU member states, the EP forms the legislative branch of the EU's institutions. The EP's primary seat is in Brussels, Belgium, but the EP holds its plenary sessions once a month in the French city of Strasbourg. At the beginning of its 2009–14 term, the EP had 736 members representing almost 500 million EU citizens from 27 member states. Members of the EP sit in transnational political groups rather than national delegations. These party groups are made up of national party delegations; for example, the EP's largest party group, the European People's Party (EPP), is composed of conservative and Christian Democratic parties from the EU's member states. In addition to the EPP, with its 265 members, there are six party groups Page 82 → during the 2009–14 EP term: the Progressive Alliance of Socialists and Democrats (S&D), with 184 members; the Alliance of Liberals and Democrats for Europe (ALDE), with 84 members; the Greens–European Free Alliance (Greens-EFA), with 55 members; the European Conservatives and Reformists (ECR), with 54 members; the European United Left–Nordic Green Left (GUE-NGL), with 35 members; and the Europe of Freedom and Democracy group (EFD), with 32 members. In addition, 27 members are unaffiliated. The major changes compared to the 2004–9 term are the change in name (and some membership) of the S&D, which used to be the Party of European Socialists (PES), the disappearance of the Union for Europe of the Nations (UEN) and the Independence/Democracy group (IND/DEM), and the creations of both the ECR and the EFD. Over the past 25 years, the EP has evolved from a mere consultative institution into an authoritative and assertive

actor in the EU's political and institutional framework. While it cannot directly initiate legislation, the EP has the power of amendment and veto in most policy areas, and its powers were greatly increased when the Treaty of Lisbon came into effect in 2009. While the right to initiate legislation lies with the European Commission, most proposals for EU law must receive the approval of both the EP and the Council of Ministers. In this sense, the two institutions resemble the upper and lower houses in traditional bicameral legislatures, such as the U.S. Congress. The legislative powers of the EP are restricted in some policy areas, but the “ordinary legislative procedure” (previously known as the co-decision procedure) now applies to the great majority of all EU legislation. Under this procedure, the EP can amend legislation as well as reject it outright, without the possibility of member state governments overturning this decision. This makes the EP an equal partner with the member states in most important areas of legislation, and the EP has proven very successful in shaping legislation, as 74.77 percent of its amendments ultimately became law9 during the 1999–2004 and 2004–9 EP terms. As a direct result of the increase in EP powers and the growing self-confidence of the institution, academic interest in the EP has increased substantially during the past decade. The scholarly literature provides a number of key insights into the politics of the EP. We now know that political conflict in the EP takes place primarily along traditional left-right lines, but is also based on a pro-/anti-EU division; that MEPs overwhelmingly vote in accordance with their party affiliations rather than their national affiliations; that the party system in the EP has become more consolidated and Page 83 → more competitive as the powers of the EP have increased; and that party group cohesion in the EP is lower than in national parliamentary systems but higher than in the U.S. Congress, for example (Hix, Noury, and Roland 2007). The real work in the EP is done in specialized committees, which serve as sources of expertise and suppliers of information to legislators not directly involved in particular pieces of legislation. This is one way in which the EP resembles the U.S. House of Representatives more than its national European counterparts. The EP's strong and influential committees negotiate and revise all legislative proposals before they are debated on the floor. Therefore, as in the House, the responsible committee has great control over the content of each piece of legislation. Indeed, the majority of substantive changes to legislation come out of committee. Moreover, the smaller size of the committees and their specialized membership promote compromise. Also as a consequence of the strong committee system in the EP, information, persuasion, and negotiation play a very important role in EP decision-making (Ringe 2010). While the EP sometimes appears dull because it lacks an overt government-versus-opposition dynamic, this lack of predetermined positions makes the EP a dynamic legislature. Decisions are made on the basis of ever-shifting policy coalitions rather than fixed coalitions of government and opposition parties. This makes informal politics, such as those that take place in the EP's intergroups, ever more important. Intergroups in the European Parliament The EP's rules of procedure provide that “individual Members may form Intergroups or other unofficial groupings of Members, to hold informal exchanges of views on specific issues across different political groups, drawing on members of different parliamentary committees, and to promote contact between Members and civil society.”10 There are currently 27 official intergroups and several dozen unofficial ones. The academic literature on these groups, however, is exceedingly thin. Corbett, Jacobs, and Shackleton's standard text on the EP provides the most extensive description of intergroups (2007, 184–90), which they define as groups “consisting of members from different Political Groups with a common interest in a particular political theme” (184). The authors trace the origins of intergroups back to the time when the EP was first directly elected in 1979. Page 84 → Corbett, Jacobs, and Shackleton briefly outline some of the benefits that intergroups offer (2007, 195), including a focus on issues of particular constituency or personal concerns, policy specialization, and cross-partisan contacts and cooperation. They also highlight the idea that intergroups provide MEPs with the opportunity to interact with outside interest groups. This interaction has received some limited attention in the academic literature, as a few

commentators concerned with lobbying in the EU reference and provide some basic descriptives of intergroups (Kohler-Koch 1997; Bouwen 2004b; Van Schendelen 2005; Butler 2008; Coen and Richardson 2009; Tanasescu 2009). Bouwen, for example, dedicates two paragraphs to the Intergroup on Financial Services as an instrument of lobbying the EP and concludes that “the informal character and uncertain subsistence of many intergroups make them attractive but unreliable arenas for legislative lobbying” (2004b, 12–13). Van Schendelen, in contrast, suggests that intergroups can be quite effective at influencing the EU's legislative agenda (2005, 77–78). KohlerKoch (1997, 9) briefly describes the basic concept of intergroups in the EP but does not investigate them further, noting that “it is difficult to even find out their number and even more so to assess their importance. They have been mushrooming in recent years, but because they meet irregularly, have a different format as to membership and organization it is hard to make general judgments.” More recently, Butler (2008, 575–79) discusses the role of intergroups in the context of the participation of nongovernmental organizations in EU lawmaking and suggests that “coordination between NGOs and MEPs through parliamentary Intergroups seems to represent the most efficient and consistent means of influencing MEPs” (575). The first intergroup recognized by the EP was the Intergroup of Elected Local and Regional Representatives in 1980; since then, the number of groups has increased substantially, with “some 58” registered during the EP's 4th legislative term (1994–99) and “around 80” during its 5th (1999–2004) (Corbett, Jacobs, and Shackleton 2007, 185). The reliance on estimates of the number of groups suggest one problem with studying intergroups—many groups exist only informally and, therefore, cannot easily be tracked. We thus differentiate between intergroups, which are formally recognized by the EP, and issue groups, which lack this recognition. This study focuses on the former. The significant number of issue groups that do not have intergroup status is in part the result of the EP's evolving rules governing intergroups, Page 85 → because avoiding these regulations provides incentives for some groups not to seek formal intergroup status. The first objective of the EP's rules on intergroups has been to regulate the involvement of interests outside of the EP.11 Starting in the mid-1990s, the EP sought to set in place increasingly tight rules about the financing of intergroups, and it now requires intergroups to provide annual declarations of financial interests (Article 6 of the EP's Rules Governing the Establishment of Intergroups).12 The second objective of the EP's rules has been to clarify the status of intergroups vis-à-vis the formal institutions of the EP, especially parties and committees, and to ensure that intergroup activity does not interfere with official parliamentary activities. The rules thus state that intergroups are not formal organs of the parliament and may not express opinions on its behalf (Article 1) and that they are not permitted to use the name or logo of the EP, “nor any names which might lead to confusion with the official organs of the European Parliament such as parliamentary committees, interparliamentary delegations and joint parliamentary committees” (Article 2). Intergroups also shall not undertake any activities “which might result in confusion with the official activities of Parliament or of its organs and which might adversely affect relations with the other Institutions of the Union or with non-member countries” (Article 3). The rules further stipulate that intergroups are allowed to meet only on Thursdays between 10:00 a.m. and noon or between 2:30 p.m. and 6:30 p.m. during the EP's monthly plenary sessions in Strasbourg, and not during voting time.13 Meeting rooms, simultaneous translation, and technical assistance are made available to intergroups only at those times (Article 9). However, party groups may also offer logistical support to intergroups provided that the conditions laid down in the rules are met. Finally, the EP has set a high bar for the establishment of intergroups. To receive formal recognition, groups must provide a list of members and have the support of at least three political groups in the EP (Article 4). The size of the party group determines how many intergroups it can support; during the 2004–9 legislative term, both the EPP-ED and PES could support 21 intergroups each. The ALDE could support 10 groups; the Greens/ALE and GUE, 8 groups each; the Ind/Dem, 6 groups; and the UEN, 5 groups. During the current 2009–14 EP term, the EPP and S&D could support 21 groups each; the ALDE, 10; the Greens/ALE and ECR, 8 each; the GUE/NGL, 6; and the EFD, 5. Procedures for establishing which proposed intergroups each party Page 86 → group decides to support vary. The EPP holds a formal vote in the party plenary, and the candidate groups that garner the most votes are then supported when the Conference of Presidents meets to establish all intergroups during the early stages of each

legislative term. A similar procedure is in place in the ECR.14 The ALDE also holds a vote, but it is an electronic vote on the party group's intranet,15 while the bureau of the Greens provides a list of priorities, on which the group then votes.16 The PES/S&D invites its members to present proposals regarding which intergroups it should support. These proposals are debated by the group's bureau following consultation with its members. The S&D bureau then proposes which groups to support. This proposal is presented to the group; for the 2009–14 legislative term, it was endorsed unanimously.17 The rules regulating the creation and operation of intergroups impose significant constraints on their activity, with the result that some groups choose not to seek EP recognition and thus formal intergroup status. For example, groups that want to include members who are not MEPs, such as outside organizations or their representatives, might choose not to seek formal recognition as an intergroup (Corbett, Jacobs, and Shackleton 2007, 185). Similarly, some groups are subsidized by outsiders (Tanasescu 2009, 50) and may prefer not to disclose their financial affairs. Finally, groups that anticipate that they are unlikely to garner enough support across three or more party groups may not even try to gain official recognition. Most often cited as constraining the work of intergroups are not financial regulations, however, but the rules specifying that intergroups can only meet at certain times.18 These rules create a number of problems—for example, greater competition for a limited number of meeting rooms and interpreters. Also an issue is lower participation by MEPs, speakers, and interested outsiders. MEPs are very busy during the EP's monthly plenaries in Strasbourg, and competition for legislators’ time is intense. In addition, many leave on Thursdays, the last day of the plenary session. Meanwhile, meetings in Strasbourg impose logistical and financial costs on both outsiders invited to speak at intergroup events and interested outside advocates, many of whom are based in Brussels. Small nongovernmental organizations can find the costs of regularly sending representatives to Strasbourg prohibitively expensive. For these reasons, some intergroups now hold unofficial meetings in Brussels instead of official ones in Strasbourg,19 and intergroup leaders are actively lobbying the EP to loosen the rules. For example, on February 22, 2010, the chairs of all intergroups submitted a joint Page 87 → letter to EP president Jerzy Buzek asking the parliament and the leaders of the political groups to review “these detrimental rules…to preserve the democratic and plural character of our House.”20 Issue groups are not affected by these rules. It is not clear, however, how many issue groups exist parallel to the 24 and 27 groups that were formally recognized by the EP during the 2004–9 and 2009–14 terms, respectively. These groups are listed in table 4.1. Internal documents from the two largest EP party groups provide some indication of what other groups may exist. These documents list groups that members of the EPP and S&D, respectively, put forward in an effort to gain support for official intergroup status. Table 4.2 lists the groups proposed in the two party groups that failed to become official intergroups during the 2009–14 term. Two aspects of this table are noteworthy. The first is that ten groups were proposed in both party groups, indicating a coordinated effort to find support across party lines. Second, while some of these groups likely no longer exist because they did not achieve formal intergroup status, others continue to operate as issue groups. For example, the group that used to be the Federalist Intergroup failed to gain formal status during the 2009–14 legislative term but persists as an issue group. The groups on volunteering and on sports also still operate despite failing to gain official intergroup status. The Danube group does not function across party lines, but its supporters in the EPP and S&D continue to work separately on the issue with the goal of gaining formal status in 2014.21 Precedent exists for this goal: The official Social Economy Intergroup was an issue group prior to the 2009–14 EP term. Issue groups that were not proposed in either the EPP or S&D but that nevertheless currently work actively include heart and vascular diseases, fair trade, reproductive rights, Kashmir, rugby union, caregivers, and healthy aging.22 Finally, among the issue groups that include both MEPs and external members are Eufores (which “promotes the deployment of renewable energy sources and energy efficiency within Europe”);23 the GLOBE Network (Global Legislators Organisation for a Balanced Environment);24 the long-standing Kangaroo Group (“an informal forum in which politicians, officials and the social partners can come together to exchange information and views regarding the issues of the day; and to devise further initiatives for the development of the European Union”);25 the Spinelli Group (which promotes a “federal and post-national Europe”);26 and the LEFT (Liberté, Egalité, Fraternité: Toujours) Caucus.27

Since complete membership information can be collected only for formal intergroups, we rely on these data in our quantitative and social network analyses. Our interviews in the EP were not limited to discussions of intergroups, however, but allowed respondents to also talk about any issue groups in which they are involved. These data demonstrate no systematic differences between intergroups and issue groups with regard to the relational and informational functions we ascribe to LMOs: Like intergroups, issue groups offer MEPs opportunities to establish social ties with colleagues with whom they share a common concern for a given cause, and they supply policyrelevant information to heterogeneous sets of legislators. Page 88 → Page 89 → Page 90 → Our dataset includes membership information for all 24 official EP intergroups during the 2004–9 legislative term (EP6) and all 27 intergroups during the 2009–14 term (EP7). The average number of members per intergroup is 53 in EP6 and 67 in EP7. The current intergroups with the greatest number of members are Climate Change, Biodiversity, and Sustainable Development (138 members); Small and Medium-Sized Enterprise (122); Lesbian, Gay, Bisexual, and Transgender Rights (115); Sky and Space (115); and Water (112). The total number of intergroups joined by MEPs ranges from 0 to 9 during both terms, with an average of 1.68 during EP6 and 2.51 during EP7. Figure 4.1 shows the distribution of the number of memberships. In 2004–9, 31.2 percent of MEPs joined no intergroups, and 24.2 percent joined a single group; 44.6 percent of members joined more than one intergroup. In 2009–14, only 18.4 percent of MEPs joined no intergroup at all, 19.2 percent joined a single group, and 62.5 percent joined multiple intergroups. In other words, MEPs have developed a greater proclivity to join intergroups, and they join slightly more groups overall. Tables 4.3 and 4.4 list the legislators with the largest number of intergroup memberships in EP6 and EP7, respectively. Page 91 → Page 92 →

The U.S. Congress: The Lynchpin of Legislative Studies The U.S. Congress is the bicameral legislature that anchors the legislative branch of the government. Its powers are enumerated in the U.S. Constitution and are cross-checked by the executive and judicial branches. The president has no proposal power in the Congress but has veto power over legislation. The lower chamber, the House of Representatives, has 435 members elected in single-member districts of approximately 700,000 constituents each. Page 93 → The upper chamber, the Senate, has 100 members—2 senators for each of the 50 states. Members of Congress are elected in majority rule elections with strong two-party competition. The two major parties, Republican and Democratic, hold various types of primary contests in the states and districts to determine their general election candidates. Once elected, members of Congress have close constituency ties, and party loyalty is weaker in Congress than in many other legislative chambers around the world. Party discipline is quite a bit stronger in the House of Representatives than in the Senate because House members serve two-year terms, while senators serve six-year terms. Senate terms are staggered so that one-third of the Senate is up for reelection every two years. Therefore, two years make up one Congress. The first Congress sat in 1789–90, and 2012 was the second session (year) of the 112th Congress. The short and known terms of House members and the highly professional nature of the legislature have resulted in a very pronounced incumbency advantage. It is not unusual for 90 percent or more of House members to be reelected, and the longest-serving members have been in office for more than 40 years. The U.S. Congress is the political and physical center of Washington, D.C., sitting atop Capitol Hill. Congress has a long history with deep traditions and norms that govern its members, especially in the Senate; however, the institution's real stability comes from its formal organizational features, parties and committees. While party loyalty may not be as strong in Congress as in many parliaments, its leadership structure is vital to its productivity. Leaders have a variety of tools, such as agenda setting and committee appointment power, that allow them to maintain legislative control. These tools are more powerful in the majoritarian, rule-laded House than in the discourse-driven Senate, but rank-and-file members abdicate powers to committee and party leaders to achieve legislative (and therefore electoral) goals.

Caucuses in the House of Representatives Members of Congress have always had mechanisms for fostering informal and social interactions with their colleagues. In the antebellum era, before Washington, D.C., was a developed city and before members of Congress were expected to spend much time there, most members stayed in boardinghouses or “messes” (Young 1966). These shared living quarters provided significant social interaction among legislators that spanned party lines, geographical boundaries, and ideological interests. While not much is Page 94 → known about members’ informal organizations between the Civil War and the 1950s, by the mid-twentieth century, a few legislative “clubs” had been formed, among them the Chowder and Marching Society; they were dominated by Republicans (Caldwell 1989, 629–30). In 1958, a group of newly elected liberal Democrats formed the Democratic Study Group because they were dissatisfied with the more moderate ideological approach of the party's leaders (Stevens, Mulhollan, and Rundquist 1981). In 1963, House Republicans formed a counterpart organization, the House Wednesday Group (Caldwell 1989). In addition to countering dissatisfaction with party leaders, these organizations provided mechanisms for research, facilitated the spread of information, and served as policy forums. By 1969, a bicameral organization, Members of Congress for Peace through Law, had been formed. These three groups are generally considered to be Congress's first three caucuses. Hammond (1991) notes that caucuses mushroomed in the post-congressional-reform era, with the number of caucuses and their memberships rising significantly after 1974. She attributes this growth to increased frustration with the party leadership (280). These congressional reforms decentralized power and broadened the authority of subcommittee chairs, leading observers to refer to the post-reform period as the era of subcommittee government (Rohde 1991). Hammond (1991) reports that members complained about disproportional ideological and geographical representation in committees and leadership, neglected policy issues, and being forced to take difficult votes on issues that would never become law. Caucuses emerged as a means to bridge the schisms these frustrations caused, providing opportunities for information exchange and political coordination across social groups in a more fragmented Congress. Party leaders initially were put off by the formation of these groups and felt threatened by the perceived attempts to skirt party leadership; however, caucuses were relatively successful at organizing members, spreading information, and asserting agenda influence through their energy and momentum. Between 1970 and 1974, the caucus population is estimated to have increased to 13; over the next six years, however, it jumped to an estimated 70 groups (R. Singh 1996). Party leaders gradually began to take advantage of the expertise and organized collective action of the caucuses, in turn encouraging the formation of more caucuses. For example, Speaker Tip O'Neill (D-MA) ensured that representatives of the Congressional Black Caucus, Congressional Hispanic Caucus, and the Congressional Caucus on Page 95 → Women's Issues sat on the Democrats’ various party committees (Hammond 1991, 283). Starting in 1979, the House Administration Committee established a procedure for caucuses to receive formal recognition. Caucuses could apply for status as an official legislative service organization (LSO) and receive office space, franking privileges, financial accounts, and staff budgets. By 1981, the committee established strict rules on expenditures and income, required quarterly reporting, and prohibited LSOs from accepting funds from outside organizations. Many caucuses applied for LSO status, and for about two years the House Administration Committee liberally granted that status to about 30 caucuses. Groups that did not apply or did not receive approval as LSOs remained unregistered, unregulated caucuses—typically referred to as congressional member organizations (CMOs) (R. Singh 1996). The caucus population continued to expand throughout the 1980s. Although the number of LSOs was limited by the House Administration Committee through a de facto moratorium on LSO formation because of resource limits, the CMO population continued to expand. By 1987, roughly 118 caucuses existed, and all members belonged to at least one LSO or CMO (Hammond 1991). By 1993, Congress had an estimated 127 caucuses (Richardson 1993).28 The rise in the popularity of caucuses has been attributed to a number of factors. Without disputing Hammond's (1991) assertion regarding the role of congressional reforms during the 1970s, others have pointed to additional

contributing factors. Robert Singh suggests that the concerns of members seeking reelection in the era of declining party power increased the need for representatives to communicate their individual, constituent-sensitive concerns to voters back home (1996, 81). As elections became more candidate-centered, media-focused, and professional, Singh argues, incumbents had increasing incentives to create, join, and participate in caucuses for electoral gain, since caucuses allow members to signal policy preferences and priorities to constituents. Davidson and Oleszek (1981) argue that in addition to “structural weaknesses” in Congress and electoral incentives, the rise in the interest group population and regional rivalries between members also contributed to the expansion of the caucus system. The 1980s also saw the establishment of caucus institutes and foundations. These organizations were typically associated with LSOs and were established in response to the regulations put in place by the House Administration Page 96 → Committee between 1979 and 1981. While technically outside organizations with no official or formal links to caucuses, these institutes and foundations were closely associated with their parent caucuses and typically had very similar names (e.g., the Congressional Black Caucus Foundation, the Congressional Fire Services Institute), and some evidence indicates that a few of these organizations may have had rent-free House office space (U.S. House, Committee on House Administration, Ad Hoc Subcommittee on Legislative Service Organizations 1982). The foundations and institutes kept records, engaged in research and fund-raising, recruited members into the caucus, and provided logistical support to the caucus. Ethical concerns soon were raised about the financial relationship between LSOs and their foundations, however, and the House Administration Committee established a series of regulations that included banning private funding of LSOs (Caldwell 1989). The controversy over the relationship between caucuses and outside organizations persisted into the 1990s. Although all members of Congress were members of caucuses or LSOs, many expressed concern regarding the financial relationships between caucuses and outside groups and the means by which LSO staff were funded. Many began to see the caucus system as a way to get around House ethics rules. For example, a June 1993 article in the Washington Times noted that in the post-Watergate era, members of Congress were prohibited from hiring spouses or children, engaging in moneymaking on Capitol Hill, and wining and dining colleagues and staff—except for LSOs (Rodriguez 1993). The writer provides examples such as the Congressional Black Caucus spending nearly $900 on a dinner for its volunteers and staff and the Environmental and Energy Study Conference (an LSO) producing a newsletter that its affiliated institute sold for a profit (Rodriguez 1993). In May 1992, Representative Jim Walsh (R-NY) called for the elimination of all caucuses. Representative Bill Thomas (R-CA), the ranking member of the House Administration Committee, complained that caucuses are “an umbrella for the mixing and comingling of private and public money” (quoted in Burger 1992). Representative Jim Bates (D-CA) called the caucus system “out of control,” and Pat Roberts (R-KS), one of the biggest opponents of caucuses in the House, called LSOs “an accident waiting to happen” (Burger 1992). In 1994, when the Republican Party won a majority of seats in the U.S. House for the first time in more than 40 years, one of their first official actions was to disband the LSOs. The rules change, sponsored by Roberts, Page 97 → took effect on January 30, 1995 (Love 1995). CMOs could no longer have an allowance, hire staff, or have office space. Nearly all of the former LSOs opted to become CMOs. Ironically, the CMOs then became more reliant on outside groups to engage in lobbying, publicity, research, and other support services (Love 1994). Some observers have speculated these changes resulted primarily from Republicans’ desire to dismantle a few specific politically formidable Democratic groups such as the Congressional Black Caucus and the Democratic Study Group.29 However, evidence from the House Administration Committee and its reports suggests widespread concern about ethical violations in the caucus and LSO system (U.S. House, Committee on House Administration, Ad Hoc Subcommittee on Legislative Service Organizations 1982, 87-905O). While some suspected that the GOP sought to dismantle the caucus system entirely, others suggested that the Republican leadership simply sought to dismantle LSOs and that CMOs were the GOP's “preferred route” (Love 1994).30 Today's rules governing caucuses (CMOs) are simple and straightforward. According to the rules published on the website for the Committee on House Administration, caucuses may not have a separate corporate or legal identity, employ any staff, have office space, support an outside entity, accept goods or services from an outside entity to

support the caucus, use personal funds to support the caucus, use a frank to support the caucus, have an independent caucus web page, or have official stationery.31 The House gift rules also apply to caucuses. If the changes enacted in 1995 were intended to dismantle the caucus system or its influence, they failed. Figure 4.2 shows that caucus membership expanded from the 103rd (1993) through 111th (2010) Congresses. Figure 4.2 also shows the steady growth in the average size of caucuses. While the 1995 reforms caused the average caucus size to drop from 18 to 9, membership rebounded after the Democrats regained a majority in the 110th Congress. In the 110th and 111th Congresses, the average caucus sizes are 22 and 26, respectively. The steady growth in the number and size of caucuses suggests their utility in spanning structural holes during this period of increased political partisanship and polarization, a point to which we return in our concluding chapter. Figure 4.3 shows that in the 109th Congress, more members joined relatively few caucuses, while few members joined many caucuses. More recently, however, the number of members who join few caucuses has declined, while the number of members who join many caucuses has increased. In addition, comparing figure 4.3 to figure 4.1 demonstrates how many more caucuses members of Congress join in comparison to the number of intergroups joined by MEPs. Page 98 → In addition, groups have found creative ways around the restrictions on caucuses. Several caucuses have hired dedicated staffers, seemingly in direct violation of these rules, by having their salaries paid by a different caucus cochair each pay period or by having partial salaries paid by each cochair during a pay period. Essentially, these specialized caucus staffers are regularly hired and fired, sometimes on a monthly basis. Not only were staffers open in telling us about this creative accounting but we were able to verify the information on a website that tracks legislative staff salaries.32 Only a very small handful of caucuses have reached the level of formality where they are skirting the rules on hiring staff. They tend to be single-party caucuses, but we found them on both sides of the aisle. Today, a number of caucuses still use foundations and institutes as a source of support. Caucuses with affiliated foundations, such as the Congressional Black Caucus Foundation, the Congressional Sportsmen's Caucus Foundation, and the Congressional Fire Services Institute, provide significant resources for the caucuses with which they work. These foundations are established for the explicit purpose of assisting their affiliated caucuses and perform no other functions. Foundation personnel were again not shy about describing in detail the activities they help perform for their affiliated caucuses. They maintain membership lists, recruit new members, help to plan events and activities, host briefings, advertise events, and perform general logistical and administrative duties for the caucus. On its face, the foundations system seems to violate the House rules against caucuses accepting goods or services from outside organizations, but the foundations and their legal counsels are diligent in assuring that they do not violate any laws, House rules, or House ethics standards. Page 99 → We discovered only one group that has entirely overcome the restrictions on caucuses. The Tom Lantos Human Rights Commission is a formal institution of Congress as of 2008. This group started as a caucus and was “promoted” to an official commission of Congress by authorizing legislation passed during the second session of the 110th Congress. The Human Rights Caucus is one of the oldest caucuses in the House and its longtime chair, Tom Lantos (D-CA), was nearly synonymous with its cause. Speaker Nancy Pelosi championed the effort to convert the caucus into a congressional commission shortly after Lantos died in February 2008; the conversion allows the Tom Lantos Human Rights Commission to have office space, budget, staff, and other House resources, such as access to court reporters who transcribe hearings.33 The commission still operates very much like a caucus, and unlike other official congressional commissions and advisory boards, members can choose to join the Human Rights Commission rather than being appointed.34 Single-party caucuses, such as the Blue Dogs and the Republican Study Page 100 → Committee, do not qualify as

LMOs as we define them in this book. These organizations operate as voting coalitions that focus on a broad range of issues rather than as voluntary groups of legislators that form around a particular issue or cause. Single-party caucuses also follow different rules of operation than most LMOs. For example, the Blue Dog Coalition meets every Tuesday when Congress is in session at 5:00 at a restaurant near the House office buildings. The meetings are for members only; staffers do not attend. Coalition members cannot miss more than three or four meetings over a six-month period. The penalty for violating the attendance rule is a loss of voting privileges in the Coalition. The group's leaders and its subcomponent task forces rotate on a regular basis and are elected from within the Coalition. Moreover, members of Congress cannot choose to join the Blue Dog Coalition: They must be nominated and accepted as members. The Blue Dog Coalition also caps its membership at a percentage of the entire Democratic Caucus to maintain cohesion and allow the Blue Dogs to operate as a homogeneous voting bloc on issues that the Coalition prioritizes. The Blue Dogs are, in other words, an exclusive club; for LMOs, in contrast, membership is voluntary and open. The Blue Dog Coalition is the strictest single-party group we found—the New Democrats and the Republican Study Committee, for example, do not have attendance requirements. They do, however, host events that are targeted primarily at members rather than staff and take steps to coordinate votes and leverage the caucus's voting power for or against legislation.35 The single-party caucus organizations ultimately are important features of the legislative landscape in Congress, but they operate more like party factions than LMOs.36 It is remarkably easy for a member of Congress to create a caucus. Doing so merely requires writing a letter to the Committee on House Administration that states the name of the group, its purpose, its officers, and some contact information. Once the committee receives the letter and engages in a pro forma approval of the caucus, the group officially exists.37 However, our method of collecting data on the population and membership of caucuses has shown hundreds of caucuses that exist without formally registering. For example, the 111th Congress has 342 official caucuses, but we found 419 caucuses. When we asked respondents about such oversights, some seemed unaware that they were not registered, some indicated they were experiencing a shift in cochairs and were waiting for the leadership to Page 101 → be sorted out before re-registering, and some seemed genuinely unconcerned about their lack of registration. This indicates that there are few benefits to registering or few costs to not registering.

Who Joins, and Why? Respondents report joining LMOs for a variety of reasons. In the EP, the most frequently cited reason is personal interest in the topic or cause promoted by the intergroup.38 Other reasons mentioned are that the group's issue relates to relevant committee work,39 that no single committee covers the intergroup's cause,40 and that joining allows MEPs to build their professional profiles.41 Finally, several respondents suggest that their membership in intergroups reflects constituent interests and concerns.42 There is, in other words, a constituency dimension to MEPs’ involvement in intergroups, as just over one-third of our respondents indicate.43 We include as references to this constituency dimension any mention of constituents, voters, citizens, elections, region or district, particular industries with seats in members’ districts, and the media “back home.” Some MEPs acknowledge, however, that they do not advertise their involvement in intergroups in their districts or that their constituents likely do not know about this part of the MEPs’ work.44 To determine what factors explain legislators’ proclivity to join LMOs, we conduct negative binomial regression analyses with the number of LMOs individual legislators join as the dependent variable. We first predict the proclivity of MEPs to join intergroups during EP6 and EP7.45 Table 4.5 shows the results, while figure 4.4 presents the predicted probabilities generated from this estimation.46 MEPs from the four party groups with the greatest number of seats after the 2004 election, which are also the most mainstream party groups in the EP ideologically, join more intergroups than MEPs from smaller fringe parties on the left and right and nonattached members (EPP: 2.49, PES: 2.18, ALDE: 2.08, Greens: 1.95). Differences in the predicted probabilities are not statistically significant, however, with the exception of the difference between the Greens and the EPP.47 Yet members of the mainstream parties join statistically significantly more LMOs than two

of the smaller and more ideologically extreme parties, GUE/NGL (1.31) and ECR (1.39), again with Page 102 → two exceptions (the Greens and the ECR). MEPs of the far-right, EU-skeptic parties (IND/DEM, UEN, and EFD) and nonattached members join the fewest intergroups (0.51). Members from Northwestern Europe join more groups (2.01) than their colleagues from East-Central Europe (1.6), but neither the difference from members from Southern Europe (1.91) nor the difference between MEPs from Southern and East-Central Europe is statistically significant.48 Female MEPs (2.2) join more intergroups than do male legislators (1.75), while leaders join fewer intergroups than do nonleaders, as the statistically significant negative coefficient for the leadership variable indicates. Differences in the predicted probabilities between leaders (1.63) and nonleaders (1.93) are not statistically significant, however.49 Also not significant is the difference between MEPs who have LMOs in the national legislature of their member states (1.98) and those who do not (1.71). Finally, a statistically significant difference exists between the predicted number of intergroups joined in EP7 (2.28) and EP6 (1.58).50 Not statistically significant in explaining the number of intergroups joined are tenure in office51 and whether MEPs are elected on the basis of preferential voting.52 Page 103 → Congressional respondents give a number of reasons for joining caucuses. Most emphasize that they join because of their personal interest in or involvement with a caucus's issue or cause.53 In at least some cases, the caucus seems to be a substitute for serving on the relevant committee.54 Others explain that they join caucuses that represent constituency concerns55 or note that they joined a specific caucus because of a direct request from either a constituent56 or an outside advocate.57 In other instances, the request comes from other members of Congress.58 Using the complete membership data from the 109th–11th Congresses, we examine the factors that predict the number of caucuses members of the House of Representatives join. Some independent variables are specific to either the EP or Congress, however, because their inclusion is not meaningful Page 104 → in the other legislature. In the EP, for example, we control for preferential voting and account for the existence of LMOs in MEPs’ member states of origins. In the case of Congress, we include predictors for legislators from racial minority groups and control for the voting percentage by which legislators won their previous single-member district elections. Consistent with Victor and Ringe (2009), we expect that Democrats join more caucuses than Republicans; that more senior members (in terms of how many Congresses they have served in) join more caucuses than less senior members; that members with larger electoral margins of victory join more caucuses; that party and committee leaders join fewer caucuses; and that women and men and ethnic or racial minorities and Caucasians show no difference in membership behavior. Table 4.6 displays the results of this negative binomial estimation; figure 4.5 shows predicted probabilities.59 The results generally confirm our expectations.60 The negative and statistically significant coefficient for party in the regression model indicates that as in the EP, party affiliation is an important predictor of LMO membership, since Democrats join more caucuses than Republicans. The predicted probabilities show that Democrats joined roughly 18 caucuses each during this period, while Republicans joined about 13 caucuses each, all else being equal. Page 105 → We also observe that rank-and-file members of Congress join more caucuses (16) than party and committee leaders (12). Again, this finding is in line with the results from the EP, although the finding is more tenuous in the latter case.61 In contrast, no statistically significant difference exists in the caucus-joining behavior of male and female members of Congress, whereas female MEPs join significantly more intergroups than do their male counterparts. We also find differences between the EP and Congress with regard to tenure in office: Senior members of Congress join more caucuses than do their junior colleagues, while in the EP, seniority does not achieve statistical significance. Caucasian members join an average of 16 caucuses, whereas members of racial or ethnic minorities join only 13 caucuses on average, a statistically significant difference.62 This finding is somewhat surprising because we have

not previously observed such differences, and they contradict our expectations. Moreover, as in the EP, more recent Congresses have had members Page 106 → join more caucuses. In the 110th Congress, the estimate is 25.8, while in the 111th it is 29.8. The LMO system appears to be becoming more popular in both institutions. Finally, legislators who won their most recent election by a narrow margin are less likely to join caucuses than members who won their election handily.63 As figure 4.5 illustrates, the greater a member's margin of victory, the more caucuses she is likely to join, and the more Congresses in which a member has served, the more caucuses he joins. Tables 4.7–4.9 list the 10 members who joined the most caucuses in the 109th–111th Congresses. A significant jump in the number of caucuses joined by these representatives occurred from the 109th to the 110th Congress. Control of the House also switched parties during this time, perhaps suggesting that Democrats join more caucuses when they are in the majority than they do when they are in the minority. Page 107 → Page 108 → Page 109 → While most of the analysis in this chapter concentrates on the details of caucus membership in the 109th–11th Congresses, we have obtained additional data covering the 103rd–8th Congresses (1993–2010). During this period, we identify 647 unique caucuses. Thirty-two of the caucuses persisted through the entire period (listed in table 4.10); the average caucus existed in five of the nine Congresses. Tables 4.11–4.13 show the 10 caucuses with the most members in the 109th–11th Congresses.

Conclusion The purpose of this chapter has been to introduce the reader to our two legislatures and their LMOs, and to set the stage for evaluating our propositions about the relational and informational benefits of LMOs. In this effort, we offered a great deal of descriptive information about intergroups in the EP and caucuses in Congress—an important contribution given how little we know about LMOs, especially in the EP. We also described the nature, collection, and processing of our social network and interview data. Finally, we examined who joins LMOs, and why, in both EP and Congress. In both legislatures, members join LMOs because they have a personal interest in their topic or cause, because the LMO relates to or is a substitute for committee work, or because of constituents’ concerns. Party affiliation helps explain legislators’ proclivity to join LMOs (members of mainstream parties in the EP and Democrats in Congress join more LMOs), as does leadership (rank-and-file members join more), and legislative term (members join more LMOs during more recent legislative terms in both EP and Congress). Gender has no effect in Congress, while female members of the EP join more intergroups; in contrast, tenure does not have an effect in the EP, while senior members of Congress join more caucuses than do their junior colleagues. Finally, MEPs’ geographic regions of origin affect their proclivity to join intergroups, with those from North-Western Europe joining the most, while Caucasian members of Congress and those that are electorally safe join more caucuses. In the next chapter, we examine the network functions of LMOs by exploring how they help legislators build social relationships across institutional boundaries, especially those imposed by party affiliation and committee jurisdictions. We provide qualitative evidence emphasizing the relational role of LMOs and examine LMO network structures using a variety Page 110 → of social network analyses tools. Chapter 6 demonstrates the informational benefits of LMOs, building on our extensive interview data. We present evidence for information flow inside the legislature as well as information exchange between legislators and outside advocates that provide information and legislative subsidies to LMOs. We also consider the extent to which LMOs fulfill their signaling and coordination roles in legislative politics, finding these functions to be of secondary importance for participants in the LMO system.

Page 111 →

CHAPTER 5 Networking the Legislators One thing [I get out of the intergroup meetings] is the ability to know and discuss with colleagues that are out of our delegations and our committees. So it provides you with an opportunity to communicate with colleagues. This is a very big parliament…so it is very difficult to meet people. Even a few minutes of discussion can make a connection…. It is very important to be able to function as a parliamentarian to have this, to create these personal links. People know you, and you know people, know their agendas, know their character. So, it is very important also in terms of the actual human functioning of the Parliament. (Respondent 60, European Parliament) Legislative member organizations (LMOs) give legislators opportunities to establish relationships with colleagues with whom they would not otherwise share social ties, thereby providing an institutional solution to a collective problem. Such relationships give lawmakers access to information that would be at best costly to acquire or at worst unattainable. We investigate the relational benefits of LMOs in this chapter by providing evidence from our interviews that highlight the groups’ role as arenas for social interactions or indeed as social networks, as many of our respondents in the two case studies call them.1 Hence, this chapter not only confirms our expectations about one of the key functions of LMOs but also demonstrates that legislators both recognize and value this function. We also evaluate one of our key propositions about social relationships established in LMOs—that is, that the majority of them tend to be weak ties. We make this case by examining the extent to which LMOs give legislators opportunities to establish strong ties with each other. Specifically, we consider the frequency of LMO meetings, the proclivity of legislators or representatives of legislative offices to attend these meetings, and the extent to which there is fluidity in the set of participants between meetings. From this investigation, we conclude that the participation of most legislators in Page 112 → LMO activities is too limited to allow for the establishment of strong ties: LMOs meet too infrequently and most legislators do not attend often enough for the majority of LMO ties to be anything but weak. This weakness is not a disadvantage, however, as our theory emphasizes. First, the creation and maintenance of weak ties is inexpensive, which means that legislators can reap the benefits of LMOs without having to invest substantial resources. Second, weak ties tend to be cross-cutting and therefore give lawmakers access to new information. Our qualitative data thus suggest that LMOs play an important relational role in legislative politics, that lawmakers themselves view this role as a key benefit of LMOs, and that most LMO ties are weak ties. These observations do not, however, establish that the structures of LMO networks are such that the informational benefits our theory outlines are realized. To make this case, we must take two further steps. First, we must test empirically whether LMO networks create social networks that are structured in such a way that they facilitate the efficient diffusion of new information and ideas. We have to show, in other words, that LMOs help establish bridging ties that cut across institutional and social boundaries imposed by party affiliations and committee assignments. We show that this is indeed the case. Second, we must move beyond presenting the potential for the efficient diffusion of information in LMO networks and demonstrate actual informational benefits. We leave this task for chapter 6, in which we provide evidence both for information flow through LMO networks inside the legislature and for the exchange of information with outside actors associated with specific LMOs. In particular, we emphasize the inflow of what we call high-utility information and the efficient diffusion of this information throughout LMO networks. We begin this chapter by examining our qualitative data to demonstrate the relational benefits of LMOs. After establishing that LMOs provide opportunities for lawmakers to build social relationships with each other that often cut across party and committee lines, we move on to investigate LMO network structures using social network analysis tools. We provide a series of social network descriptive statistics for our two legislatures and then empirically test our propositions about the structures of LMO networks. We first illustrate the bridging nature

of LMO ties by examining to what extent LMO ties replicate existing social ties. For example, if Legislators A and B are connected to each to other because they are members of the same LMO, Page 113 → but they also serve on the same committee, their LMO tie is less valuable because they are already socially connected—the LMO tie replicates the committee tie. We expect to find that many LMO ties connect pairs of legislators who would not otherwise be directly connected to each other, and we illustrate this reality by examining the ego networks of a small set of focal legislators. We then consider whether members of LMOs tend to connect colleagues who are otherwise connected to each other, thereby making these LMO connections redundant. If Legislator A has LMO ties to both Legislators B and C, these ties will be more valuable for the transmission of information if B is not also connected to C, because A's LMO ties are needed for information flow between B and C. In this context, we rely on Burt's constraint (Burt 1992), which will be lower for Legislator A if B and C are not connected to each other. More generally, constraint decreases as fewer of ego's alters are connected to one another. We expect to find that LMO members have lower constraint scores than non-LMO members, because they connect a greater number of colleagues who are not otherwise connected to each other.

The Relational Value of LMOs Our interview respondents emphasize the importance of both intergroups and caucuses for the establishment of relationships inside the legislature. In the European Parliament (EP), the function of intergroups as facilitators of personal contacts among interested actors is explicitly highlighted by 17 of 21 members of the European Parliament (MEPs) we interviewed as well as 5 of 8 staffers, 1 of 2 party advisers, and 5 of 10 outside advocates.2 These respondents discuss the value of intergroups for building relationships, making contacts, meeting people, getting to know them, and even building friendships. In a statement that meshes with our theoretical propositions about the relational value of intergroups, one MEP said that intergroups are “a meeting place.”3 Almost one-third of our EP respondents even use the word network in reference to the functions of intergroups.4 One MEP describes his intergroup as “a forum for discussion [that] brings people together and creates a network in order to promote one idea and support this idea.”5 A second respondent maintains that an intergroup is “a networking body…a place Page 114 → to listen to new ideas and to network.”6 Notably, intergroups create “networks [that] take you beyond your usual group of suspects,”7 in that they allow MEPs to establish relationships with colleagues whom they do not already know from their other legislative activities.8 They build networks across EU member states, political groups, committees, and delegations: The intergroups are a good chance to create friendships between political groups, between different countries.9 [What I get out of the meetings] is the ability to know…colleagues that are out of our delegations and our committees.10 MEPs, in other words, both meet each other in intergroups (“there are certainly contacts that I have only because of the intergroup”),11 and they get to know each other (“especially at the beginning of the term it happens quite regularly that [MEPs] don't know each other so well but…then they see each other on a regular basis in intergroup meetings and they get to know each other”).12 Some MEPs provide specific examples of having met close collaborators from other party groups through intergroup activities. One Liberal MEP from a Benelux country explains that he met a British Conservative colleague “in the intergroup. I didn't know him before. So when I need something now, [he] is someone I can go to.”13 Another Green MEP makes a similar point about a colleague from the center-right EPP group who is not from her member state: If I were looking for somebody from the EPP to do something on [this issue], probably one of the key people I would think of would be…the co-president of the intergroup. [Question: But wouldn't you know that anyway because you know [him]?] But I don't know [him]. He's not on my committee. You know, so this is one of the things that [intergroups] give you: an opportunity to meet other members

who are not on your committee.14

Along similar lines, one of the outside advocates we interviewed describes members of the intergroup with which she works as having “contacts in other meetings…. In some committee meetings, they sit together and they talk to each other—like informal discussions…even if they are not from the same party.”15 Such discussions also take place immediately after meetings Page 115 → of the intergroups,16 both between MEPs and outside advocates who attend the meeting and directly between legislators. According to one MEP, after meetings end, “I can consider if there is somebody I would like to get in touch with and ask, ‘What is your experience? You said A, B, C. Can we spend another half an hour talking about this?’”17 Another respondent suggests that “it is not only the meeting itself, but also the opportunity to speak to people informally without making difficult appointments, and sometimes speak to somebody who you normally wouldn't speak to.”18 In fact, several respondents emphasize that their collaboration with colleagues whom they met in the intergroup is not limited to the intergroup's issue area.19 One Portuguese member, for example, says that she and a Dutch colleague from another party group “have been working a lot together on drafting joint amendments on issues that have nothing to do with [the intergroup]. [But] we met through the intergroup.”20 Caucuses in the U.S. Congress are also seen as an important mechanism that allows legislators to build relationships with one another, especially for those who may not be connected through formal legislative institutions. One staffer describes relationship building among legislative offices as “obviously the main thing, the number 1…. I can't emphasize that more…. Everything that they do is about building a relationship.”21 Respondents indicate that they use caucuses for various types of relationship building and keeping contacts. Thirty-three of our 44 congressional respondents mention relationship building, emphasizing in particular how the caucus organization helps them to stay in touch with other offices.22 One staffer puts it bluntly: People come to you and all you are doing is introducing them, so the caucus can be a great mechanism for people to meet—networking basically. So they come to these briefings [or] receptions mainly in that regard, and they meet other people…and share cards, and then keep us in the loop, basically…. It gives us some mechanism for people to meet.23 Another staffer concurs: Caucuses “give you a reason to be talking with another office.”24 A third says succinctly that “there is no reason” she would know her counterpart in the other cochair's office “if it weren't for the caucus.”25 A number of staffers talk about how the caucus serves as a way for their Page 116 → “boss” (the member for whom they work, as aides commonly refer to legislators on Capitol Hill) to develop and maintain relationships with other members. Said one, “I think they are a great avenue for members to be together, to be able to have an idea of what is going on…. That is what we need.”26 Another respondent explains that “if it weren't for the caucus” his boss and their cross-party cochair counterpart “probably wouldn't end up working together that much.”27 Several respondents again use the word network or networking specifically when referring to the benefits of caucus membership and leadership.28 In addition, many use the term relationship when they describe the benefits of caucuses.29 One maintains that the caucus “is like a web,”30 and another reports that “it is a great format for building relationships.”31 One respondent specifically refers to the decline of comity in Congress and the fact that most legislators travel back to their districts on the weekends rather than spending time with their colleagues in Washington (see, e.g., Uslaner 1993). His boss feels that the caucus helps to make up for the lack of social opportunities to develop personal relationships with other members,32 and that these relationships help build trust between colleagues.33 The informality of caucuses also seems to help legislative offices, as opposed to members themselves, to develop and maintain relationships. The staffers we interviewed emphasize that they have a fair degree of control over caucus events and agendas, especially compared to their level of input with committees, for example.34 This

additional control and input provides opportunities for offices to connect with one another. In addition, respondents surmise that members may feel more at ease and open in a caucus setting, as opposed to a committee, because caucuses are informal and there are no transcripts or official records of meetings.35 This informality encourages people to speak more frequently and openly, which in turn creates opportunities for legislators and their staff to communicate and maintain or develop their relationships. Hence, the relative informality of LMOs is seen as an asset in the U.S. Congress, as it is in the EP. One MEP who emphasizes the social importance of the intergroup's less formalized nature explains, I believe that intergroups should be more relaxed and informal [and] a lot more social as well, because it is kind of that informal grouping of people where you can have very frank discussions. It does not always have to be, “My party's line is this.”…We're very busy people, and sometimes it's nice to do something in a relaxed and informal environment.36 Page 117 →

Ties That Bind across Parties and Committees Social interactions in EP intergroups almost always take place between members from various party groups and committees. This fact is critical with regard to political coordination in intergroups, which allow for the creation of policy coalitions across both party boundaries and committee lines. Several responses highlight the importance of intergroups for building compromise, consensus, common positions, majorities, or policy coalitions.37 As one MEP explains, The intergroup really is broader than any committee and broader than any party group. It combines people from different committees and people from different party groups, which is very effective because then if you want to influence a piece of legislation, you have people in all the parties with their mandates for negotiations…and you have people in all the committees…. Often a piece of legislation will be going through different committees so it is important that you have people in all of the umbrella committees taking the [intergroup's] issue into consideration.38 While some respondents highlight that intergroups consist of “like-minded individuals,”39 many more emphasize that members of a group share not a set of policy preferences but rather policy priorities and outcome preferences.40 Consequently, intergroups are important in terms of building policy coalitions because they do not consist of groups of members who necessarily agree with each other even on the intergroup's cause. As one MEP explains, “The people you get are the ones who are interested in the issue. Where you get more of a divergence would probably be in the policy solutions.”41 The cross-partisan membership and interactions in intergroups is emphasized as a significant feature by 18 of 21 MEPs, all 8 staffers, both party advisers, and 9 of 10 outside advocates, for a total of 37 of 41 respondents.42 This cross-partisanship lies at the core of the intergroup system,43 and “the main thing” these groups have to offer, according to one respondent, is “a cross-party small meeting.”44 This is, as another one puts it, a “huge value, yes. Huge value…. It's the biggest value of the intergroup.”45 In many intergroups, cross-partisanship is institutionalized in the leadership, which usually comprises representatives from various party groups.46 Different members of the bureau may, for example, take turns chairing the intergroup's meetings, as is the case in the SME Intergroup.47 The Trade Page 118 → Union Coordination Intergroup rotates the chair position between the two largest EP party groups, with an S&D member chairing the group for half a legislative term and an EPP member taking the other half.48 Moreover, the coordinators (that is, spokespeople) for the five largest party groups in the Employment and Social Affairs Committee serve simultaneously as the Trade Union Coordination Intergroup's bureau.49 In this context, with the LMO leadership network we observe clearly the value of strong, embedded ties between frequent intergroup participants who interact with each other on a regular basis. Crucially, cross-party interactions in the intergroup are described as less competitive and confrontational than in

committee or plenary. One MEP notes, “I don't have to present myself with elbows to reach my goal.”50 Others explain that while committee work is “often quite confrontational,” intergroups allow MEPs to have discussions “about what we can do, how we can influence [policy] in the right direction, which is a fun way to work with political opponents, actually.”51 Along similar lines, an outside advocate describes his interaction with a freshman MEP, who told him that “she is very fascinated to be a member of that intergroup because she realized she could talk freely with other MEPs…. She said, ‘That's what I like.’”52 In this process, MEPs can identify colleagues in other parties who may be receptive to negotiation and compromise: “It is quite important to know, for example, for the EPP, where there is room to convince, and how to convince.”53 Such exercises in cross-partisanship are not purely symbolic, as one respondent emphasizes: Having all parties represented in the bureau of the intergroup, which is a smaller grouping of ten MEPs who sort of lead the work on the intergroup, gives you a very good overview of who you can count on, with whom you need to do more work, who is going to champion the issue. That is very, very helpful.54 The intergroup membership data also confirms that intergroups are truly cross-partisan organizations. The average number of party groups represented in the 24 intergroups is 5.13 during EP6 and 5.74 during EP7 (out of 7 for each term). Only two intergroups during EP6 (Globalisation, Trade Union Coordination) and one intergroup during EP7 (Ways of Saint James) report having members from fewer than four party groups. Page 119 → Respondents emphasize the importance of intergroups as organizations whose topics and membership cut across not only party lines but also committee jurisdictions.55 Intergroup meetings bring together MEPs from a range of committees, and since their topics often do not fall neatly into the jurisdiction of one single committee, intergroups can “think across committees”56 or “address the issue from all the committees’ perspectives.”57 The goal in such cases is to find a compromise that spans various committees. In other instances, intergroups may be important not simply because of their cross-cutting memberships but also because their principal topics are not a top priority in any single committee, yet taken as a whole, the subject is recognized as significant.58 Ultimately, however, deliberation and bargaining in intergroups do not at all obligate MEPs after they return to their party groups and committees, which makes some MEPs skeptical of the utility of intergroups. One respondent maintains, for example, that “the cross-partisanship is relevant in the sense that you actually have representativeness or whatever you want to call it, but it does not bind you in any way.”59 Others, however, emphasize the value of having intergroup members return to their party groups and committees with new insights and information that they can use to influence party positions. For example, one MEP maintains that if he wants to make a difference with regard to a specific topic, “I will circulate it to the intergroup members and say, ‘Look, this is coming up. Can you please speak on this matter in your group meeting?’”60 Others suggest that if they want colleagues to “do the right thing in committee”61 or to “influence other committees,” then “intergroups are crucial.”62 Moreover, cooperation beyond party lines in committee is not limited to deliberation but extends to highly consequential legislative activities such as the drafting of amendments.63 In other words, EP intergroups do not merely provide mediums for information exchange but also serve as instruments of political coordination. This coordination is “really specific: You meet someone, you have an amendment, you talk to him. He has an amendment, he talks to you,” as one MEP put it.64 A staffer echoes this assessment: In the intergroup in which her MEP is involved, “We are all in different committees, and when someone makes an amendment in the committee, then he calls you over to also sign the amendments and make the thing stronger.”65 An MEP from one of the most active intergroups explains in more detail, Page 120 → I use this network when I have amendments that I want support for…. Usually you get support from

your own group, but you don't know about the other groups. But here I know that even though this guy is conservative, he is in the intergroup, so obviously he supports [its cause]. So I might get him to cosign an amendment that I put forward…and then usually it will go through. So that's probably the most important reason for me to be active in the group. It gives us power that way. [I do this] very often…. And then really whenever a colleague that's also a member of the intergroup proposes something, I will almost always support it.66

One of his colleagues concurs but again emphasizes that such collaboration is not limited to the realm of a particular intergroup but also extends to other issue areas: “You start to know the other people and also the way they think concerning certain issues…. Then they can join you and cosign amendments with you even if that has nothing to do with the subject of the intergroup.”67 Respondents in Congress also emphasize that relationships established in caucuses cross party lines and allow for greater bipartisanship. Since many of the formal institutions of Congress (e.g., committees and parties) are so strongly partisan, caucuses offer a means for members to communicate and be connected across party lines. Said one respondent, Things can get pretty tribal pretty quick and pretty lively around here. Caucuses, it's not the only way, but it's one of the ways, it's one of the contexts in which…that's not supposed to matter. And so it creates an opening for members to build relationships,…not just any relationships, but relationships across the aisle.68 Many others similarly emphasize the bipartisan nature of caucuses,69 explaining that bipartisanship is “an utmost goal. That's a top priority” for many caucuses.70 Numerous respondents also highlight how their caucus helps them maintain cross-partisan relationships.71 One staffer explains, You have got areas of common interest…whether you're liberal, conservative, whether you're from the South or the Pacific Northwest or New England. We have three Democratic cochairs and three Republican cochairs, and we work really closely together, and the staff work really closely together. And, you know, and we all get along…. I just see it as a blessing.72 Page 121 → The bipartisan benefits of caucuses are most easily reaped, of course, when an issue crosses the aisle relatively easily;73 in other instances, however, the Republican and Democratic cochairs may deliberately avoid bringing an issue on which they disagree forward to the rest of the caucus membership for fear that doing so would cause friction in the caucus.74 Such behavior emphasizes the priority that members place on the bipartisan nature of caucuses. If members did not value the bipartisan aspect of their caucus so highly, they would make agendasetting choices that are more likely to disrupt the group. For many groups, maintaining a bipartisan caucus is akin to solving their collective action problem with respect to caucus membership—without a commitment to bipartisanship, the group unravels. A number of respondents discuss how being bipartisan helps them to recruit members to the caucus, present caucus ideas to party leaders or the president, and promote legislation in ways that do not alienate other members.75 Indeed, bipartisanship is such a key feature of some caucuses that they will sometimes be stymied by the departure of a cochair, as the caucus attempts to recruit a member from a specific party to take the chair's place. In such instances, the caucus may prefer to lay dormant for a while rather than appear to have leadership from only one party.76 In other instances, caucuses will not meet if not enough representatives from both sides of the aisle are present: “If we can get at least one Republican, one Democrat, we will hold the meeting. If we only have two Republicans, we'll try to hold off…. I can't speak for all of them, but in the [caucuses] that [my boss] is a member of, that's very common.”77 Some respondents, however, also recognize the imbalance in the general membership of some caucuses even

though they value bipartisanship and work to ensure that they have bipartisan cochairs. One recognizes that bipartisanship is important, “although it is not equal by any means. I think given the 96 members of the Caucus, there [are] probably 6 that are Republican? Eight that are Republican?”78 Yet the same respondent then emphasizes the importance of those few, because “if we need Republicans for something, we're always going to ask them because they're more reliable to support it…. Democrats, we can get any one of them, but when you need a Republican, there's only a few that are going to say yes.”79 In the end, impressions of partisanship or bias are seen as potentially problematic for many caucuses, as respondents fear that the relational and informational benefits of caucuses might be diminished. Hence, those two benefits are not fully distinguishable or separable from one another. Page 122 →

The Prevalence of Weak Ties Respondents in both legislatures emphasize the importance of LMOs for the establishment of relationships and contacts, and many of the quotations above suggest fairly close working relationships between specific pairs of legislators. This phenomenon is indeed what we observe in both intergroups and caucuses—a relatively small group of lawmakers establishes close working relationships, or strong ties, with each other. Such close collaboration is particularly evident between LMO leaders—for example, the cochairs of congressional caucuses or the members of EP intergroups’ bureaus, which usually consist of between five and eight MEPs and often hold additional meetings that are separate from the public intergroup events.80 However, legislators do not know all colleagues from their LMOs quite as well, and some describe their weak ties to other members of the group. According to one interviewee, “It is impossible to know all the members and to have the same type of connection and relation with all the members. I know much better the people who are part of the bureau than the other members.”81 In chapter 2, we argue that the great majority of ties between members of the same LMO are weak ties, because lawmakers generally do not participate in LMOs with enough frequency to create strong connections with colleagues. Indeed, this characteristic helps keep participation in LMOs cheap from the point of view of individual legislators, because creating and maintaining weak ties requires little time and few resources. In this section, we consider the extent to which caucuses give legislators opportunities to establish strong ties with one another by examining the frequency of LMO meetings and patterns of participation. Examining patterns of participation in LMOs to estimate tie strength is far from ideal. Indeed, the amount of time two persons spend together is only one of four elements Granovetter (1973, 1361) emphasizes as contributing to tie strength, along with emotional intensity, intimacy, and the degree of reciprocation characterizing a relationship. However, it is not possible to capture the level of intensity, intimacy, and reciprocation involved in the relationships between legislators with the data available to us, requiring us to focus on opportunities for legislators to establish strong LMO ties with each other. Intensity, intimacy, and reciprocation are at least partially a function of opportunities for social interaction, however; LMO members who rarely interact are therefore unlikely to establish strong ties. Page 123 → Intergroups in the EP tend to have a small number of members who are strongly tied to each other—in particular, those members who assume formal or informal leadership positions in a given group—but most MEPs have only limited opportunities for social interaction in intergroups, for several reasons. First, intergroups meet at most once a month and on average every other month—in fact, a few groups meet only three or four times a year. Most intergroups, in other words, meet far less often than legislative committees, for example, which meet for between four and eight half days per month (Corbett, Jacobs, and Shackleton 2007, 134).82 Second, many MEPs tend to invest a fair amount of time in a single intergroup but are far less active in others.83 They frequently attend events of their primary intergroup84 but only occasionally attend meetings of the other intergroups of which they are

members.85 Most report attending only when the particular topic on the agenda holds special interest for them,86 which further decreases the number of opportunities for interaction. Third, the number of MEPs who attend intergroup meetings is fairly small. Based on our respondents’ estimates, the mean number of participants is 12; the median number is 10.87 Many more MEP offices tend to be represented at meetings, since a majority of respondents suggest that MEPs send assistants to intergroup meetings if the topic is of interest and the legislators are unable to attend;88 nevertheless, these numbers indicate that most MEPs do not have a great number of opportunities to meet and interact. Two additional factors contribute to the scarcity of interaction. First, attendees usually comprise a small, core group of members89—generally, bureau members who are likely to share strong ties. Second, a fair amount of fluidity exists in the attendance by intergroup members from outside this core, since MEPs will attend only if a particular topic interests them. These realities indicate that activity in intergroups alone is unlikely to establish strong ties between pairs of MEPs other than those who serve on a given intergroup's board and perhaps an additional select few. For the great majority of legislators, in contrast, intergroup activity is too limited to allow the establishment of a great number of strong ties. We come to the same conclusion in the U.S. Congress, where we first consider the frequency with which caucuses hold events that bring together members. Great variation exists in this regard, in that the most active caucuses hold events on a regular basis, while some others seem to be entirely dormant for long periods of time. Many respondents tell us that caucus events are held with varying frequency,90 while some caucuses have regular Page 124 → meetings—weekly,91 monthly,92 or quarterly.93 Others meet only when doing so seems necessary, maybe once or twice a year.94 In general, outside of the largest and most organized caucuses, group events tend to occur irregularly and not more than a few times per year. This phenomenon contrasts starkly with the frequency of committee hearings. In the 109th–11th Congresses, committees held a median 100 hearings per Congress,95 a number that dwarfs the number of events held by caucuses. In other words, caucus participants have relatively few opportunities to interact directly on a regular basis, thereby impeding the development of strong ties. About one-third of U.S. respondents provide estimates of head counts at caucus events, with a median attendance of 45,96 although this number may overestimate the actual number of participants, since several of our respondents represent large and active caucuses. Caucus events are generally targeted at staff, and most participants are legislative assistants (or sometimes interns), who gather handouts, take notes, and meet other attendees.97 One respondent bluntly identifies the biggest lure for staffers: “Most of the staff that come to the briefing are only coming because of the free food. It does not mean it is bad. It means that you have got an audience and…they get a nugget of information or they get contacts.”98 Events tend to be organized so that the caucus cochairs can stop by and make brief remarks and shake a few hands before leaving the event to staffers. In the words of one staffer, “The cochairs oftentimes will come and sort of kick off—bring, you know, opening remarks”;99 according to another, “If you get any more than three or four members of Congress, it is a huge success. But even at more formal events, if you get more than two members of Congress at the same place at the same time, you can consider that a victory.”100 As in the case of EP intergroups, variation exist in how often representatives from different congressional offices attend caucus events. Most respondents maintain that their attendance varies depending on the topic as well as the demands on their time. Staffers also report that the quality of events varies from caucus to caucus, and sometimes they choose events based on their anticipated usefulness. One respondent explains, I always attend the Progressive Caucus. The Progressive Caucus is a very active caucus. They put out vote recommendations, and I mean, they really do stuff a lot. They get invitations from the president. The India caucus probably wouldn't get an invitation from the president. The Black Caucus is like that, I think the Hispanic Caucus is like that.101 Page 125 →

Staff generally attend events put on by their members’ core caucus(es), especially groups of which their bosses serve as chairs or cochairs. Attendance at other caucuses will be much less frequent. A core of 15 or so people attends regularly; other attendance varies.102 Caucuses thus generally do not provide consistent enough communication and events for most of their members to use this venue to form strong relationships with one another. Rather, caucuses are forums in which legislators or their offices can establish weak ties. Weak ties can be quite valuable, however, in that caucus members can readily tap into these networks if the need arises; in this sense, many of the relationships legislators develop in LMOs are utilitarian but have a sort of ad hoc nature. As one EP respondent explains, I know the people that are there often. Some of them I know also from committee work. I know maybe 10 of them, I would guess. [The others] I just know by name, and I know that they are a member…. If I need some support or something, I would then ask the secretariat from the group to just give me a list of the members, of the people that attend the meetings, and then I would…just write an email to this person or get a meeting with this person.103

LMOs as Social Networks: Descriptive Statistics Having established that legislators view LMOs as important arenas for social interaction and the establishment of social ties and that most of these ties are weak ties, we now test our propositions about the bridging nature of LMO ties using social network analysis tools. We begin by providing a number of descriptive social network statistics to give a general overview of LMO network structures and the positions of key actors within those networks. We use LMO membership information to generate square relational matrices for each legislature for each legislative session. Ties between persons are determined on the basis of membership in LMOs (Borgatti, Everett, and Freeman 2002). The resulting networks use common membership in one or more LMOs as a measure of tie strength. In other words, for each network, we are looking at an n × n adjacency matrix A representing all the LMO-based ties such that aij represents the total number of joint LMO Page 126 → memberships. Aij = 0 if the ith legislator does not share membership in any LMOs with the jth legislator and 1 ≤ aij ≤ X if he or she does, where X is the maximum number of joint LMO memberships of any two legislators. In the EP, we analyze two LMO networks, one each for the 2004–9 (a matrix composed of 751 × 751 MEPs) and the 2009–14 EP term (730 × 730).104 For the U.S. Congress, the matrix sizes are 433 × 433 for the 109th Congress (2005–7), 437 × 437 for the 110th Congress (2007–9), and 433 × 433 for the 111th Congress (2009–11).105 Our data are undirected, or symmetric: If actor A and actor B are in at least one LMO together, then they are connected, and we make no assumptions about the direction of their connection. A first notable difference between intergroup networks in the EP and caucus networks in Congress is that members of Congress join a much greater number of LMOs than do their counterparts in the EP. While the average number of caucuses over the three Congresses increases from 8.6 (109th) to 19 (110th) to 25 (111th), MEPs join only 1.68 intergroups on average during EP6 and 2.51 during EP7. It is thus not surprising that caucus networks are much denser than intergroup networks, where density is simply the proportion of existing ties between legislators as a function of all possible ties. As table 5.1 shows, density for the 2004–9 EP term is 0.14, a number that increases to 0.28 for the 2009–14 term.106 In the U.S. Congress, density increases from 0.36 in the 109th Congress to 0.77 in the 110th and 0.75 in the 111th. These basic differences between EP and congressional networks bring us back to an important question: Why do members of Congress join so many more LMOs than their EP counterparts? Our answer involves the incentives provided by the first-past-the-post, single-member district system in the United States, which encourages legislators to signal support for a particular cause or policy issue by joining a large number of caucuses, incentives that are absent from the proportional representation system by which MEPs are elected. Since joining LMOs is voluntary and inexpensive, members of Congress have little reason not to sign up for whatever caucus their constituents might care about. This dynamic implies that it is very difficult to differentiate between members who

join caucuses with the intention of actively participating (indeed, our interviews support the notion that many, if not most, legislators in both Congress and the EP have a small number of core LMOs on which they spend time and resources)107 and members who join LMOs to lend symbolic support to a cause and to signal policy priorities to their constituents. These incentives to join a great number of congressional caucuses introduce a great deal of noise into our membership data because membership information alone does not tell us much about individual caucus members’ activity levels, and these more specific data are unavailable aside from the limited qualitative evidence we provide in other parts of the book. Page 127 → These realities, in turn, raise another question: How do the greater proclivity of members of Congress to join caucuses and the resulting greater density of the caucus network relate to our empirical findings about LMO network structures? The answer is fairly simple: The data from the U.S. case are denser, noisier, and entail less variance than the EU data. Consequently, we are less likely to find meaningful substantive effects in the U.S. data, all else being equal. The implications of the greater densities of the congressional networks are observable in the remainder of this chapter. It is, for example, not surprising that the geodesic distances between legislators (the number of edges in the shortest path between any two nodes) are shorter in Congress than in the EP (table 5.2). While the average distances in the caucus networks are 1.57 (109th Congress), 1.18 (110th) and 1.23 (111th), they are somewhat longer in the case of EP6 (1.75) and EP7 (1.69). In other words, legislators in the Congress can reach each other more easily than can their counterparts in the EP. Page 128 → Nonetheless, overall distances are short even in the case of the EP. Indeed, 25.4 percent of MEPs during EP6 and 33.2 percent of MEPs in EP7 are directly connected to each other via joint intergroup membership. In EP6, 74.2 percent of MEPs can reach each other in just two steps; for EP7, that number is 66.8 percent. These results mean that fewer than 1 percent of legislators need three steps to reach one another via the LMO network. In the 109th Congress, 44.4 percent of members can reach each other in one step, and that number rises to 82.5 percent in the 110th and remains high, at 77.1 percent, in the 111th. In all three Congresses, virtually all legislators can reach each other within two steps. These overall short distances are significant because they imply that information exchanged in LMO networks does not have to travel far between members. Short distances facilitate the efficient diffusion of information and increase the probability of knowledge transfer (J. Singh 2005). As a point of contrast, we compare LMO networks and networks made up of common committee assignments in both legislatures (table 5.3). This comparison is interesting from an informational point of view, because we argue in chapter 2 that committee networks are inferior to LMO networks with regard to information exchange. Our network data show that legislators are indeed more connected through LMO networks than committee networks. First, distances in the committee network are greater than in LMO networks. In EP7, for example, the average geodesic difference in a network composed of committee ties is 2.69, as opposed to 1.69 in the intergroup network.108 In the case of the 111th Congress, the geodesic distance in the committee network is 1.88, much greater than the comparable distance in the caucus network (1.23). Of course, legislators do not have complete freedom over their committee assignments. While members of Congress can request committee assignments, there is no guarantee that their requests will be granted, and members can sit only on a small handful of committees (typically 1–4) but can join caucuses at will. The situation is similar in the EP, where members sit on 1–3 committees. In light of these differences, we expect densities to be greater and geodesic distances to be shorter in the LMO networks of both legislatures than in their committee networks. Page 129 → In sum, legislators in both chambers are more tightly connected through LMO networks than committee networks, as evidenced by the shorter paths and the higher percentages of nodes that can be reached in just one or two steps.

Such networks are ideal for efficient information flow because they enable nearly all legislators to follow a number of paths to reach other legislators; if one path breaks down, alternatives are available. To investigate further the connectedness of legislators in LMO networks, we identify the pairs of lawmakers with the most LMO ties. In EP6, five pairs of legislators have the maximum number of joint intergroup memberships (five). Natalie Griesbeck (ALDE, France) has five direct ties with Thomas Mann (EPP, Germany), Philippe Morillon (ALDE, France), and Edit Bauer (EPP, Slovakia), as do Thomas Ulmer (EPP, Germany) and Jacques Toubon (EPP, France). During EP7, two MEPs from France, Michele Striffler and Jean-Pierre Audy (both EPP), have the maximum number of direct ties (seven). In the 109th Congress, the pair of legislators who share the most caucus memberships is James McDermot (D-WA, 7) and Michael McNulty (D-NY, 21), who have 24 joint caucus memberships. In the 110th Congress, this pair drops to the number 2 slot on most common caucus memberships with 31, and the lead pair is Richard Larsen (D-WA, 2) and Dennis Moore (D-KS, 3), who are in 35 caucuses together. In the 111th Congress, Henry Waxman (D-CA, 30) has 39 caucuses in common with three other Democrats: Donald Payne (D-NJ, 10), James McDermot (D-WA, 7), and Christopher Van Hollen (D-MD, 8). For a more formal measurement of the level of connectedness between any two actors, we rely on the concept of maximum flow, which considers how many actors that are directly adjacent to Actor A lead to pathways to Actor B. If this number is large, A and B are more connected, since there are numerous ways for them to reach each other. The logic of this measure suggests that the availability of pathways between actors makes a linkage strong, as opposed to distance or some other measure of connectedness. For example, if member A needs to send a message to member Z, and she can send it only via member C, the connection between A and Z is weak. Page 130 → Conversely, if A can send a message to Z via C, D, E, F, or G, then the connection between A and Z is stronger (Hanneman and Riddle 2005). The maximum flow algorithm thus takes into account all connections between all actors, not just the most direct paths between actors (Wasserman and Faust 1994). Maximum flow ranges from 0 to 524 in EP6, with an average of 47 (standard deviation = 66); in EP7, the range is 0 to 820, with a mean of 109 (standard deviation = 116). Table 5.4 lists the pairs of MEPs who have the highest maximum flow scores–values of 389 or higher in EP6 and 689 or higher in EP7. In both EP6 and EP7, the list of most connected pairs of legislators is composed exclusively of the 10 most central actors. This group of most connected MEPs cuts across both party and national lines: In EP6, 25 of the 45 pairs are not from the same party group or the same EU member state. Sixteen pairs are from the same party group but different member states; two are from the same country but different party groups; and only two share both party group and national affiliation. EP7 shows less heterogeneity given the greater dominance of a single party, the EPP. Nine of 45 pairs are from neither the same party group nor the same member state. Thirty-four pairs are from the same party group (EPP) but not from the same country, and only two pairs share a common party affiliation and member state of origin. In the U.S. Congress, the maximum flow scores range from 0 to 1219 (109th), 0 to 3708 (110th), and 0 to 4936 (111th). The average maximum flow in the three congresses is 156 (standard deviation = 157), 902 (standard deviation = 652), and 1129 (standard deviation = 927), respectively. Tables 5.5–5.7 show the pairs of legislators with the highest maximum flow scores in each Congress. The same names tend to pop up on these tables, both within a Congress (because these members join many caucuses) and across Congresses (since a joiner in one Congress remains a joiner in other Congresses). The dyads with the greatest flow tend to be Democrats, but a few Republicans—Phil English (R-PA, 3) in the 110th Congress and Frank LoBiondo (R-NJ, 2) in the 109th—appear on these lists. These pairs represent some geographical diversity. In other words, as in the EP, the congressional dyads with the greatest flow between them cut across party and geographical lines, at least to some extent. Finally, we consider the centrality of individual legislators in the LMO network, relying on betweenness centrality, which describes the number of times a node (or legislator) lies on a path between two other nodes. This centrality measure thus captures the idea that legislators’ importance in the LMO network is a function of the extent to which they connect other network members. Table 5.8 presents normalized betweenness centrality scores, where the betweenness score is divided by the maximum value and expressed as a percentage (see

Freeman 1979), for the most central actors in the EP. In EP6, centrality ranges from 0 to 0.74, with an average of 0.05 (standard deviation = 0.09). In EP7, it ranges from 0 to 0.61, with a mean of 0.06 (standard deviation = 0.09). Page 131 → Page 132 → Page 133 → Page 134 → In Congress, normalized betweenness centrality in the 109th measures between 0 and 0.62, with an average of 0.11 (standard deviation = 0.11). In the 110th, the range is 0 to 0.1, with an average of 0.04 (standard deviation = 0.02), and in the 111th, the range is 0 to 0.2, with an average of 0.05 (standard deviation = 0.04). Table 5.9 shows the members of Congress with the highest betweenness centrality scores. Betweenness centrality captures the structural positions of legislators in the LMO network only (that is, the social network that describes ties between legislators based solely on their membership in LMOs), not the legislative network as a whole (that is, the social network made up of all connections between legislators).109 In other words, betweenness centrality is a measure of the internal brokerage of LMO networks and by extension of the bridging power of individual legislators within the LMO network. This differs from our primary concept of interest, however, which is the bridging nature of LMO ties in the legislative network as a whole, or the extent to which LMO ties connect legislators who are not otherwise connected through any other legislative institution. We now turn to our investigation of the extent to which LMO ties are bridging ties in the legislative network. Page 135 →

Are LMO Ties Bridging? The bridging nature of LMO ties is apparent in some basic descriptive statistics for the EP and Congress. To begin, we consider the extent to which Page 136 → intergroup ties in the European Parliament replicate other kinds of social ties between legislators and therefore do not provide new connections to colleagues. Our focus is on EP6 (2004–9), the last EP term for which we have complete data. We seek to model every institutional tie between MEPs during this term—that is, to capture every connection that exists between a given pair of MEPs by virtue of serving in the same institutional body. We count as institutional ties joint participation in every body that finds mention in the EP Rules of Procedure—party groups, national party delegations, parliamentary committees, delegations,110 the Bureau of the EP,111 the College of Quaestors,112 the Conference of Presidents,113 the Conference of Committee Chairs, and the Conference of Delegation Chairs. To this list we add the MEPs’ members states of origin to account for ties through shared nationality.114 A first indication of the bridging nature of intergroup ties is that 47.74 percent (17,467 of 33,058) of ties between MEPs who share at least one common intergroup membership do not replicate existing institutional ties. In other words, almost half of all intergroup ties connect lawmakers who would not otherwise be connected. These ties are genuinely cross-cutting: 66.29 percent of intergroup ties connect legislators who are not in the same party group, 85.53 percent connect legislators who do not serve on the same committee, and 90.73 percent connect legislators who are not from the same EU member state. To illustrate the bridging nature of intergroup ties in the EP, figures 5.1a and 5.1b show a network composed of both intergroup and other institutional ties. For presentational reasons, the network shown in the figures is much smaller than the actual intergroup membership network, with only 10 legislators and 45 ties; however, the number, distribution, and strength of the institutional ties are proportional to those in the actual network. These networks are comprised only of MEPs who share at least one intergroup tie; since only one-quarter of MEPs actually share direct intergroup ties, this presentation avoids showing a great number of MEPs with no intergroup ties. The 24 solid lines connecting the actors are institutional ties other than intergroup ties. The thickness of the line indicates the number of common institutional ties between two actors. The 21 dashed lines in figure 5.1b indicate intergroup ties that do not replicate the institutional ties in figure 5.1a. Comparing the top and bottom panels of figure 5.1 thus illustrates the great number of connections that exist between MEPs by virtue of their intergroup membership. Page 137 →

We further demonstrate the bridging nature of intergroup ties by investigating in more detail the ego networks of four focal legislators (that is, the social networks composed of all MEPs with whom each focal legislator is connected and their connections with each other), who differ from each other in intergroup membership. The first focal lawmaker, Mathieu Grosch (EPP, Belgium), was randomly drawn from the 35.2 percent of MEPs who are members of a single intergroup; Robert Sturdy (EPP, UK) was drawn from the 42.7 percent who have joined two or three intergroups; and Françoise Grossetête (EPP, France) was chosen from among the 22.1 percent who are affiliated with between four and eight intergroups. We also examine Page 138 → the ego network of Jacques Toubon (EPP, France), one of two MEPs who have joined the maximum number of nine intergroups during EP6, again following a random draw.115 We consider how intergroup ties affect the number of first- and second-degree ties in each focal legislator's ego network. We determine first how many colleagues an MEP establishes a direct connection to via intergroup membership with whom he has no other institutional ties. Grosch has 69 intergroup ties, of which 51 (73.91 percent) are with colleagues with whom he shares at least one other institutional membership and 18 (26.09 percent) are to MEPs with whom he does not share any other institutional ties. Sturdy is affiliated with three intergroups, memberships that tie him to 172 colleagues. Of those ties, 129 (75 percent) replicate institutional ties, while 43 (25 percent) do not. Grossetête joined six intergroups and shares 251 intergroup ties, of which 160 (63.75 percent) replicate institutional ties and 91 (36.25 percent) do not. Finally, Toubon has the largest number of nonreplicative ties of the four lawmakers, as 186 (47.82 percent) of his 389 intergroup ties do not replicate his institutional ties. Next, we examine how many second-degree ties (connections with friends of friends) become first-degree (or direct) ties when we consider intergroup ties in addition to all other institutional ties between MEPs. For Grosch, intergroup membership turns 4.3 percent (18 of 417) of his second-degree ties into first-degree ties. For Sturdy, this number is 13.35 percent (43 of 322); for Grossetête, it is 28.09 percent (91 of 324); and for Toubon, it is 50.41 percent (186 of 369). Hence, by joining intergroups, MEPs can turn a substantial number of friends of friends into just friends. Table 5.10 summarizes these results. In the 109th Congress, 31 percent of ties (10,378 of 33,489) between members who share at least one common caucus membership do not replicate other institutional ties. For the 110th Congress, this number is 34.6 percent (25,381 of 73,295), and it is 33.4 percent (23,548 of 70,445) for the 111th. Caucus ties are also clearly crosscutting in that 41.6 percent connect members who are not in the same party in the 109th Congress; for the 110th Congress, that number is 46.9 percent, and for the 111th Congress, it is 44.5 percent. In the 109th Congress, 79.1 percent of caucus ties connect legislators who are not on the same committee, numbers that rise to 78.0 percent for the 110th and 79.5 in the 111th. And lawmakers who are not from the same state are linked by caucus ties at rates of 94.8 percent Page 139 → in the 109th Congress, 95.8 for the 110th, and 95.5 percent for the 111th. We also examine the ego networks of four legislators from the each of the 109th, 110th, and 111th Congresses, again selecting samples that vary by the total number of caucuses joined. We randomly draw one legislator from each quartile in the distribution of total caucus memberships in each Congress.116 As in the case of the EP, many LMO ties among members of Congress are bridging ties that do not replicate other institutional connections (established in this case by virtue of joint legislative leadership, party affiliation, committee membership, and home state) (table 5.10). One legislator has no caucus ties that link him to colleagues with whom he is not already connected; this particular finding is instructive in that it highlights the possibility for caucuses in the Congress to be made up entirely of members of a single party. Representative Jeff Flake (R-AZ, 6) was a member of only two caucuses in the 110th Congress, and both are made up of only Republicans. One is the Republican Study Committee, a single-party caucus of House conservatives that operates more like a party faction or voting coalition than an LMO.117 The second is the Congressional Western Caucus, which at the time was composed exclusively of Republicans, though it has included Democrats during other sessions. The caucus ties of the other focal legislators in the Congress are bridging, however. For example, keeping the

focus on the 110th Congress, Kevin Brady (R, TX-8) joined 13 caucuses and has 342 caucus ties, of which 61 percent (209 of 342) replicate other institutional ties. Moreover, caucus ties convert 68.6 percent of his seconddegree ties into first-degree ties. Ron Lewis (R, KY-2) joined 19 caucuses and has 385 total caucus ties, with 55.6 percent of them replicating other institutional ties; caucus ties convert 81.82 percent of his second-degree ties into first-degree ties. Finally, Edward Brown joined 36 caucuses and had 374 total caucus ties.118 His caucus ties replicate other institutional ties 63.6 percent of the time and convert 82.9 percent of his second-degree ties into first-degree ties. The percentage of nonreplicative ties in the U.S. case varies, but in 10 of the 12 focal cases in table 5.10, it exceeds 25 percent, and it ranges as high as 48.83 percent. In other words, for most of our focal legislators in Congress, between a quarter and half of their total caucus ties provide connections to colleagues with whom they share no other institutional ties. This finding is notable in a legislature with only two parties, where every lawmaker is connected to a very large number of colleagues through party membership alone. Looking at seconddegree ties also illustrates the bridging nature of caucus ties, as caucus membership converts large numbers of second-degree institutional ties into first-degree ties. The average rate at which these 12 focal legislators convert friends of friends into friends through shared caucus membership is 54 percent, and that number has a large range—from 0 to 91.21 percent—with greater conversion rates in the most recent Congress studied. Page 140 → Page 141 → Measuring Legislators’ Constraints Having illustrated that many LMO ties do not replicate other institutional ties, we now continue to test our theoretical propositions about the bridging nature of LMO ties by examining whether LMO members connect pairs of colleagues who are not otherwise tied to one another. We rely on Burt's measure of constraint, which describes the extent to which a person's network is concentrated in redundant contacts (Burt 1992, chapter 2). The objective is to conduct paired t-tests to compare the constraint of LMO members and non-LMO members. Since a lower constraint score indicates that an actor connects more people who are not otherwise tied to each other, our first expectation is that the constraint scores of LMO members are statistically significantly lower than those of legislators who are not members of any LMOs. In other words, members of LMOs connect more fellow legislators who are not already connected to one another than do lawmakers who are not members of any LMOs. Next, we compare the constraint scores of LMO members to their constraint scores in the absence of their LMO membership. We expect that LMO members connect more fellow legislators who are not otherwise connected to each other than they would if they were not members of any LMOs—that is, LMO membership (statistically significantly) decreases individual legislators’ constraint scores. At first glance, these hypotheses may read like truisms. After all, one would expect that any ego with a greater number of LMO or other social ties is more likely to link otherwise disconnected alters and therefore have a relatively lower constraint score. Finding confirmation for our hypotheses would thus not be surprising. With regard to our second hypothesis in particular, this raises the question of whether comparing networks including LMO ties to networks without LMO ties considers the relevant counterfactual. Page 142 → Perhaps a more appropriate point of comparison would be a counterfactual network with a randomly distributed set of simulated LMO ties. We maintain that our expectations and comparisons directly follow from our theoretical story. To start, our hypotheses reflect our principal argument about the value of weak, bridging LMO ties: they facilitate the flow of information in the legislative arena by connecting (subsets of) legislators who would not be connected to each other in their absence. A counterfactual in which LMO ties are absent is thus the appropriate point of comparison against which to evaluate our propositions. In contrast, a comparison to a counterfactual network including a set of randomly distributed simulated LMO ties would only follow logically from our theoretical propositions if we argued that legislators joined LMOs in a conscious effort to affect the number of bridging ties in their ego networks. If we maintained, for example, that lawmakers joined LMOs in an effort to increase the number of

bridging ties in their ego networks, we should indeed expect to find greater bridgingness in the LMO network than in a random counterfactual. This is not what we argue, however. Our theory highlights weak and bridging LMO ties as an informational benefit, not as their raison d'être. Indeed, this proposition is borne out in our qualitative data: While legislators consistently emphasize the value of establishing relationships with colleagues from other parties and committees, they do not claim to join LMOs in an effort to reap those benefits. Building relationships that cut across party and committee lines is a valuable outcome associated with joining LMOs, but legislators join for all kinds of different reasons (discussed in chapter 4). Accordingly, the appropriate comparison of legislative networks with LMOs is to a counterfactual without LMOs. In our examination of constraint, we no longer seek simply to model every institutional tie between MEPs during EP6. After all, two legislators might have shared no institutional ties in 2004–9 but have established a relationship during a prior legislative term. To account for this possibility, we collect institutional membership information not only for EP6 but also as far back as the first direct EP election in 1979. To capture variation in membership during a given term, we code this information for each year since 1979 and count as a tie any joint service in any one of the institutional bodies listed earlier. We also consider two MEPs to have a tie if they represent the same member state. Not every such tie is an actual social tie, of course. For example, two Page 143 → MEPs may have served in the same committee in 1979 (or in 2009, for that matter) but never exchanged a single word. Consequently we likely overestimate the number of institutional ties between most pairs of MEPs. Conversely, we likely underestimate the strength of the ties between some specific pairs of MEPs, since we cannot, for example, account for social connections such as personal friendships. However, it is unlikely that many pairs of legislators fall into this category, and the effect of not counting personal friendship ties will be outweighed by our overestimation of the strength of ties for most other pairs. In other words, this approach is unlikely to bias the results in favor of our propositions; on the contrary, our estimation is the most conservative possible. We calculate constraint scores on the basis of the complete 1979–2009 institutional ties data plus our intergroup ties data for EP6. The resulting constraint scores range from 0.007 to 0.033, with a mean of 0.013 (standard deviation = 0.004). We begin by comparing the constraint scores of MEPs with no intergroup memberships to those with one or more memberships. This comparison demonstrates that, as hypothesized, the networks around MEPs who are members of intergroups are (statistically significantly) less constrained than the networks around MEPs who are not intergroup members. Furthermore, constraint decreases successively (and statistically significantly) with a greater number of intergroup memberships. Specifically, MEPs who are not members of any intergroups have a constraint score of 0.016; those who are members of one or more intergroups have a score of 0.012; and those who are members of five or more intergroups have a score of 0.01 (for all differences, pr(t) = 0.001). Intergroup ties are thus genuinely cross-cutting. Next, we compare the constraint scores for MEPs who are members of intergroups in this network to a network that ignores intergroup ties and thus only includes the 1979–2009 institutional ties data. In other words, we compare the actual constraint scores for intergroup members to what their constraint scores would be if they were not members of any intergroups. This comparison shows that the average constraint for the network that includes intergroup ties is lower for members of intergroups (at 0.012) than their constraint in the network that ignores intergroup ties (at 0.014), a difference that is statistically significant (pr(t) = 0.001).119 Hence, both of our expectations are confirmed in the case of the EP: Intergroup ties bridge structural holes. Page 144 → We also calculate Burt's constraint using institutional ties data for Congress, this time counting ties going back to the 89th Congress (1965–66), when the two longest-serving legislators in our sample, John Dingell and John Conyers (both D-MI), served their first terms together. We count such ties for all members of the 111th Congress, taking into account membership in the same party, service on the same committee, legislative leadership (serving

as party or committee chair), and being from the same state. Calculating and comparing constraint scores for members of Congress is more problematic than in the case of the EP, however, because of the greater proclivity of members of Congress to join caucuses. Indeed, conducting a paired comparison between caucus members and members who choose to not join any caucuses is all but meaningless because so few members fall into the latter category. In the 111th Congress, for example, only five members chose to stay outside the caucus system. It is thus not surprising that we do not find a statistically significant difference between the constraint scores of those five members (0.0114) and everyone else (0.0112, pr(t) = 0.46). We do, however, find confirmation for our expectation that caucus members have lower constraint scores (0.011185) than they would have if they were not members of any caucuses (0.011203, pr(t) = 0.01). The substantive difference is very small, but finding any statistically significant difference that confirms our hypothesis is striking given the great density of this network and the resulting low variance in the constraint measure.

Conclusion In this chapter, we provided a series of systematic tests of our expectations about LMOs’ role in helping legislators develop relationships. Specifically, we expected that LMOs solve an informational collective problem for legislators by allowing them to develop relationships that cut across such social and institutional cleavages in legislative politics as party and committee lines. LMOs do so because they are voluntary organizations that are inexpensive to join and maintain. Without LMOs, legislators would have fewer relationships that cut across formal institutional divisions and therefore more limited access to nonredundant information. We have demonstrated this relational function of LMOs through our interviews with respondents in the EP and Congress, who consistently and Page 145 → emphatically note the importance of LMO relationships to their work. Indeed, several respondents highlight the idea that they maintain important relationships with colleagues from other parties and committees that are grounded in common LMO membership and that these relationships would not exist in the absence of LMOs. Interviewees also emphasize the cross-partisan nature of LMOs and the relationships they foster and how the relative informality of LMOs facilitates the establishment of ties that cross the aisle. Most LMO ties are not, however, strong social ties. Indeed, one important characteristic of LMO ties is that they are weak because LMOs do not provide frequent enough contact between their members to foster strong ties. This weakness is significant for two reasons: Weak ties are cheap to create and maintain, thereby contributing to the low-cost nature of LMOs, and they tend to be bridging, thereby facilitating the flow of policy-relevant information. We illustrated our argument about the bridging nature of ties by examining the ego networks of several focal legislators; we found that a great many LMO ties in both the EP and Congress provide unique, direct connections between lawmakers in that they do not connect legislators who are already linked to one another through other institutional ties, such as joint membership in parties or committees. We then proceeded to a more rigorous test of our proposition using Burt's constraint. The resulting empirical test is quite conservative because we examine the extent to which LMOs provide connections between legislators above and beyond all their other institutional connections, including those developed earlier in their legislative careers. Our expectations—that LMO members connect more colleagues who would not otherwise be connected than do non-LMO members—are again confirmed. In sum, our qualitative and quantitative evidence is consistent with our theoretical story about the role of LMOs in developing and maintaining interlegislator relationships that cut across existing formal institutional ties. While each of our empirical tests may not be conclusive by itself, together they provide a preponderance of evidence in support of our key proposition about the bridging nature of LMO ties. LMO networks thus provide opportunity structures for the efficient diffusion of information throughout the legislature. In our next chapter, we move beyond demonstrating the potential for information exchange and offer evidence for the informational benefits of LMOs.

Page 146 →

CHAPTER 6 There's No Such Thing as Too Much Information (in Legislative Politics) Caucuses are generally headed up and they're actually driven by whoever are the chairs and the cochairs, and they're the people that you can count on to go to for information, for contacts, to pull briefings together on these kinds of issues, to know where some of the best resources on these pieces of information are. And that's I think why [a member of Congress] enjoys being a member of a lot of caucuses. Because it just kind of helps us maximize our resources when you have such a limited number of staff. I mean, House members have excruciatingly small staffs. (Respondent 31) Our analyses of the legislative member organization (LMO) networks in the European Parliament (EP) and the U.S. Congress confirm that LMO ties are bridging ties that connect legislators who, in the absence of LMOs, would be less connected to each other. Such networks facilitate the flow of policy-relevant information in legislative politics. Showing that LMO networks are structured in such a way that they should facilitate the flow of information does not, however, demonstrate that they indeed fulfill this function in legislative politics. This chapter, therefore, investigates the informational role of LMOs. We begin by examining information flow inside the legislature—that is, between legislators themselves and between their offices. Our interview data offer evidence for the exchange of both policy and political information. This exchange of information relates to policy proposals on which legislators are actively working as well as attempts either to help set the legislative agenda or to keep an issue or cause on lawmakers’ radar. LMOs thus serve as educators on both current issues and potential future concerns. Our respondents again see the relative informality of LMOs—compared to legislative committees, for example—as an asset, highlighting this characteristic when they discuss LMOs’ relational benefits (see chapter 5). LMOs, Page 147 → they maintain, provide arenas where honest discussions between members of different parties and committees may take place and where participants can exchange ideas in the absence of party competition or predetermined ideological battle lines. Some differences exist in terms of internal information exchange, however—for example, there is greater reliance on LMO meetings for the distribution of information in the EP, whereas Congress prefers written materials disseminated via email. We also find that the exchange of political information, as opposed to policy information, is more prevalent in the EP than in Congress, a finding that makes sense in a political setting where cross-party coalitions are needed to pass legislation. Information exchange in LMOs extends beyond the legislative arena itself because many LMOs maintain privileged relationships with specific outside organizations that are institutionalized in joint LMO activities and regular interaction between key staff both inside and outside the chamber. Outside advocates play two critical roles in the LMO systems in both Brussels and Washington: They provide legislative subsidies to LMOs and their members and supply them with political and policy information. We consider the relationships between LMOs and outside advocates in the second part of this chapter, focusing on both the legislative subsidies outsiders provide and the exchange of policy-relevant information. We focus in particular on incentives for outside advocates to provide high-utility information. These incentives revolve around legislators’ ability to triangulate and thus double-check information they receive from the outside; the high levels of internal social embeddedness in single-LMO networks; and outsiders’ fear that legislators might take away their privileged access should they provide redundant or even misleading information. In sum, LMOs increase the likelihood that high-utility information enters the legislative realm and is then efficiently diffused through internal information networks. We begin this chapter by examining the exchange of information through LMO networks inside the EP and Congress. Next, we discuss how the personal commitment of LMO leaders and the involvement of outside actors in LMOs help keep the costs of LMOs low from the point of view of individual participants. The inclusion of

outside advocates and the privileged, institutionalized access LMOs provide are also important from an informational point of view, so we examine patterns of information flow between insiders and outsiders. We close this chapter by investigating in more detail our propositions about the flow of high-utility information from the outside in. Page 148 →

Information Exchange inside the Legislature Our qualitative data collected in the EP highlight the informational function of intergroups, a role all 41 of our EP respondents discusses. Some quite explicitly emphasize the importance of exchanging policy and political information, while others discuss in more general terms how intergroups allow members of the European Parliament (MEPs) to learn, exchange ideas, be briefed, or receive input. One of the principal functions of intergroups is the exchange of high-quality policy information,1 which allows MEPs to make better policy choices, as one respondent indicates: I think it is very important to learn what is going on in different areas and to hear what people have to say, and normally it is a confrontation of different ideas. I think all these groups are quite important [as] part of the opinion building.2 Most of this information is available in the intergroup meetings, which are attended by MEPs who are interested in finding out more, and therefore they can come learn more information from people that come to present at the groups or come to the meetings. That is what happens at the meetings—we have a presentation from somebody, we discuss issues that are coming up on that specific topic, and then we have a social side to the intergroups as well…. The intergroups allow you to select the areas that you are interested in and to give you more in-depth knowledge of certain, specific areas…. The intergroup is another way for MEPs to receive information and discuss matters more in depth. It is important that we do that.3 Other respondents concur, emphasizing that intergroups are a “way of gathering information and of learning a lot”4 and that involvement in intergroups means “being a recipient of information. That is of great assistance for me and my office and my work.”5 The information to which MEPs have access in intergroups relates directly to MEPs’ legislative work and prepares them for the more formal parts of their work.6 It is often “technical information [that] you do not necessarily have access to. You could go and search for it, or eventually it Page 149 → could come to you, but…there is an intergroup that allows you to do it. I would say it is very useful from that point of view,” as one respondent explains.7 Another agrees that “intergroups are there to prepare the debates in the committee and party groups. And if one has been in an intergroup discussion, when one comes to the group discussion or the committee discussion…the quality of one's engagement will be higher.”8 In sum, intergroups allow legislators “to get reliable information and use it efficiently”9 in their legislative work. Intergroups also allow legislators to identify those colleagues to whom they can turn when they require policy information.10 While not every member of an intergroup is necessarily an expert in the relevant policy areas, those who participate actively or even fill leadership positions are likely sources of information. One MEP suggests that “the support network is really so they know who to go to to tackle any given issue,”11 while another describes how intergroups allow MEPs “to make the contacts and to address the right people…. It is small synergies that develop.”12 According to one intergroup leader, When I was elected as intergroup president, that's when I found a lot of MEPs started to make contact with me, even those who hadn't beforehand…. Because as intergroup president I guess they know that they will have probably more productive work outputs if they know they can contact me and I can

provide them information.13

Similarly, another MEP explains that when she and her staff are “writing about disability law,” they go to a colleague who holds a leadership position in the Disability Intergroup “because he knows more about it. Same when he comes to me about climate change or something on climate policy or energy policy.”14 However, MEPs use intergroups not only for the exchange of policy information but also to exchange political information with colleagues from other parties and committees. One MEP explains that interaction with colleagues from other party groups in the LMO's bureau allows legislators to talk openly and honestly about the likely arithmetic in the different political groups…. What will be the breakdown inside the EPP group, the Liberal group, the Greens, the GUE, and the Socialist group? Do we get the qualifying majority we need at the crucial second-reading stage? [Participants] are Page 150 → very, very honest. They do not come along to bullshit us. They come along to say, “Well, in the Liberal group, we are going to have a problem here.”…And because of that honesty, we get a really good indication of where things will go.15 Another member confirms this idea: “I know what my political party's views are on certain topics, but I do not know what other parties’ are, so this is certainly a way that the intergroup does prove its value.”16 Intergroups provide a “test field where one can see what the reactions are, what the opinions are.”17 These interactions are not necessarily limited to the bureau meeting, as another respondent emphasizes, but also take place during and after formal group meetings. Indeed, you could use such a discussion to listen to the different views and…after the meeting speak to people and say, “Well, do you think you could go there?” So it is not only the meeting itself but also the opportunity to speak to people informally without making difficult appointments and sometimes speak to somebody who you normally would not speak to.18 A second type of political information exchanged in the context of intergroups relates to information about what issues or specific policy proposals are likely to make it onto the EP's agenda. In other words, intergroups allow legislators to more efficiently track legislation, both before it enters the agenda and once the legislative process is under way. Intergroups are generally concerned with issues that do not fall neatly into the jurisdiction of any one committee, which means that lawmakers must keep track of legislative proposals in a variety of committees that may affect their issue of choice. This is one of the key functions fulfilled by intergroups19 and in particular by their leaders, who either track legislation directly or rely on intergroup members from various committees to provide information about policy proposals that relate to the intergroups’ causes. The relative informality of intergroups, whose meetings are generally open to anyone who wishes to attend,20 also provides other opportunities for individual MEPs to gather political information. For example, a number of respondents affiliated with the Trade Union Coordination Intergroup describe how a British Conservative MEP, Philip Bushill-Matthews, who served as the coordinator of the EPP in the Employment Committee during Page 151 → the 2004–9 term and who had a reputation of not being a friend of trade unions, would regularly attend their meetings “for strategic reasons” and “to see what we are up to.” Bushill-Matthews “always wants to know what the enemy thinks, so he is always present.”21 Our respondents see the openness that allows Bushill-Matthews to attend the Trade Union Coordination Intergroup as a major benefit. They describe intergroups as forums for discussion, deliberation, and even negotiation that are more informal and less constraining than parties and committees. This informality is highlighted by a majority of respondents,22 who discuss the benefits of having more informal meetings, honest and straightforward discussions, and speaking, discussing, and asking questions freely. One MEP, for example, told us,

It is important that there is an informal forum for people…to meet, to exchange opinions, to brush up, to be briefed and then to go back into their groups, their party groups, to take part in the group discussion on these matters at a more specialist, expert degree of information and confidence…. There is a certain importance of being able to speak to each other in more private circumstances.23 Several respondents compare interactions in intergroups to those in another small-group setting, especially committee meetings, and emphasize the advantages of the relative informality of intergroups. One views committees as “more formal, more technical, more narrow-minded in a way. [An intergroup] is more openminded…. Because they are as informal it is easier to produce ideas…. Everything is open, we can say whatever we want…. Even if it is stupid, we can do it.”24 Another respondent notes how “everyone is at ease…. Everyone speaks more freely than they would perhaps in a committee meeting.”25 A quote from a respondent who frequently and closely works with the Trade Union Coordination Intergroup nicely sums up the benefits of these informal exchanges and even ties them to a specific legislative proposal: I perceived the intergroup bureau as a valuable place where there is time for MEPs to exchange their ideas without being interrupted and without time, pressure, or the public. [They have an] atmosphere of exchanging among equals without the feeling one has to defend one's own political line, and with being allowed to ask advice. That is where I see its specific role. And Page 152 → not being held liable for what they say…. The Trade Union Intergroup is an open body, open discussions…. On the European Works Council Directive, the Trade Union Intergroup bureau was an important place to resolve the conflict. [It] was, to be frank, the one place where [people] could shout at each other.26 In the U.S. Congress, 33 of 44 respondents also emphasize that information exchange is a benefit to caucus participation. In some cases, the caucus leadership even sees its role as an “educator,”27 and as one respondent suggests, “Nine times out of ten, [our activity consists of] sending out information.”28 Caucuses can deal with technical, complex, or unfamiliar issues. In such cases, leaders use the caucus as a way of educating and disseminating substantive information to other congressional offices because the primary constituency they wish to serve is members.29 Caucuses are helpful in this regard because they allow for the quick and easy identification of other offices that are likely to be interested in this information.30 Without the caucus and its list of members, the information provider would either have to spend more time figuring out which offices to target or have to use a blanket form of communication that goes to all offices and is more likely to be ignored. Caucuses have no legislative authority; they cannot use House resources or hold hearings or obtain transcript services. In this sense, one respondent is quite right when he notes that “caucuses can't do anything.”31 For precisely this reason, the same respondent highlights the value of caucuses not in reference to their formal power but as conduits of information exchange: “It's more like a candid forum to do more informal discussions on the issues. [It's] a mechanism for them to get their message out or their information out.”32 Some caucuses do not even take formal positions on issues or endorse legislation. In an effort to maintain their bipartisan appeal and their roles as facilitators of relationships and information, they deliberately do not get involved in advocating, promoting, or opposing legislation. For these groups, the informational role is the caucus's sole function.33 The primary vehicles for dispersing information through caucuses are briefings and emails, some of which include newsletters, opinion pieces or editorials, or other attachments with information. Thirty of our 44 U.S. respondents emphasize the importance of their email lists for communicating with caucus members and distributing information.34 According to one respondent, the mailing list is “one of the primary functions of a caucus.”35 Thirteen respondents discuss the immediate benefits provided by email listservs,36 Page 153 → although others also specifically mention the costs associated with such an approach. After all, staffers are inundated with so much information and such a volume of email that it is easy for the informational value of caucus emails to become diluted.37 It may also be difficult to keep lists up-to-date enough for them to be useful.38 Finally, 17 respondents discuss using briefings to spread information.39 As discussed previously, however, many respondents also indicate that caucus events occur irregularly, with their frequency fluctuating with the House's active legislative agenda. With regard to their function as an internal advocacy group, caucuses fall into two categories: those that are

committed to not taking any explicit policy positions and those that use the caucus organization to help advocate on behalf of legislative objectives. Caucuses of the first type refrain from advocacy in an effort to maintain their membership: They are umbrella organizations that try to keep diverse sets of actors engaged. These groups frequently face collective action dilemmas as they try to attract and maintain members, and they are conservative in their choices about legislative activity. Many caucuses are therefore not legislative coalitions or advocacy groups. Indeed, more than one respondent talks about not taking a position on bills for fear of offending portions of a group's membership.40 These groups are careful to distinguish “lobbying” or “advocating” from “educating.” Examples of such caucuses include the Congressional Fire Services Caucus, Congressional Heart and Stroke Coalition, High Performance Building Caucus, the Congressional Hazards Caucus, and the Congressional International Conservation Caucus. For these groups, remaining bipartisan and having a broad appeal is more important than championing a particular legislative objective. Also, the legislative measures related to these caucuses tend to be large bills that cover broad policy topics. For example, it is unlikely that Congress would ever consider a heart-health bill and more likely that Congress would consider health care legislation, part of which would be related to heart health. But such a large bill would be too controversial and ideological to attract the backing of a large bipartisan group. The fact that some caucuses choose to prioritize their bipartisan nature and large coalition size demonstrates the value they place on the caucus as an institution that provides worthwhile informational benefits to members. A number of representatives from different caucuses relate their experiences with the second type of group, which use the caucus infrastructure to help with internal legislative advocacy.41 Said one respondent, Page 154 → If the caucus were to put its weight behind [a bill,] we can communicate [our support] through the caucus—through our communication channels. More people can communicate through those channels to get more people to cosponsor it, and then we can say “as 30 members of the Congressional [X] Caucus, we represent those views.” And so it just adds a certain weight to the message.42 Another interviewee concurs, explaining that “you have an ability to band together and say, ‘Hello, leadership. We have 40. We have 30. We have 60. We have, you know, 118 people here who can push this through, or block such and such.’”43 These caucuses pride themselves on their large numbers and try to use their membership to their advantage in their advocacy efforts. However, we generally found that caucuses of the first type have an easier time attracting members and growing than those of the second type. In the opinion of at least one respondent, the informational nature of caucuses makes them much more likely to be the first type than the second: “More often than not [my boss] will sign up for that caucus because usually caucuses are nonpartisan and more information-driven.”44 In addition to emailing information and holding briefings, caucus chairs use their position to organize other legislative activity for a cause, such as engaging in letter writing of all sorts. Whether it's writing to an administration official, writing to another head of state, writing to a governor, there's a number of different people he might write to. So the congressman would cosign a letter on behalf of a certain issue that's being spearheaded by a caucus.45 Many caucuses in the House seem to make regular use of “Dear Colleague” letters, circulated to every member of Congress.46 Caucuses also engage in Capitol Hill receptions, press conferences, and roundtable discussions,47 and some caucuses hold special events to draw attention to their cause. The Congressional Bike Caucus, for example, sponsors Bike to Work Week, and the Traumatic Brain Injury Caucus sponsors an annual Traumatic Brain Injury Fair. Social events are important because they can be an easy way to draw attention to a group or cause, and staff or members may find it easier to attend an after-hours reception than a briefing in the middle of the day. Page 155 → Weeklong special or interactive events can be ways to attract attention in the service of spreading information

about a cause. A few of our respondents discuss using the caucus infrastructure to exchange political information about legislation. One respondent talks about using the caucus as a way to contact a committee chair to try to move a piece of legislation. Using the caucus's name gave the staffer more credence when he contacted the chair, making the staffer stand out from the other 434 congressional offices calling to make “an ask” (a request of a member, typically to support or oppose a proposal).48 By and large, however, most respondents who mention the importance of information exchange concentrate on policy information rather than political information. In this way, congressional caucuses seem to differ from intergroups in the EP, where the exchange of political information appears to be more commonplace. Some staffers and their bosses seem to think of the informational role of caucuses as going hand-in-hand with their role as representatives of constituencies. One respondent emphasizes that the caucus system allows participants to be “ahead of the curve” on issues because a caucus has prompted them to sign on to a bill or have awareness of an issue, in many instances before constituency groups come calling and asking questions.49 Respondents cite the relative informality of caucuses as another benefit: “Caucuses are not mandatory, nothing in it is mandatory. It's just—my personal understanding of it is that it's nice to have more information. More information is better. More capability is better. And a lot of the bonus of being part of a caucus is that it's…a setup that you can choose to use or not.”50 Caucuses’ informality means that they are flexible institutions and that members can use them however is most utilitarian; it gives members more control, especially when compared to committees: “We take marching orders from the committee in a lot of ways, but with the caucus event we control it.”51 Furthermore, some respondents indicate that members are more at ease and open in the caucus setting because of the informality: “People are going to be more open when they know they're not under the watchful eye of a recording.”52 In addition to providing coordination between offices, caucuses facilitate the flow of information across parties and between leaders and rank-and-file members. First, caucuses provide a mechanism for members to have equivalent information across parties. Some caucus chairs—or more accurately, their staffers—regularly share information between offices. If Page 156 → one cochair has a meeting with a group in which she receives a policy brief or position paper, she will likely obtain a second copy and send one to her counterpart's office.53 Caucuses can serve as a means of breaking through strict partisanship and the information inequality that can come with it. Second, caucuses can help members coordinate and communicate with leaders. Participants see value in sending a message to colleagues or leadership that a group of members (as opposed to a single member) supports something. As one respondent puts it, “It increases [our] voice, and it coordinates those activities with a greater number of voices.”54 Another calls this phenomenon “strength in numbers.”55 Caucuses may have added informational benefits over committees because they can react more quickly to some issues and are more willing to take on controversial topics: “We take up issues that are sometimes a little too hot for the committees to handle…. Things where the United States hasn't signed onto treaties, which some of the committees will see as maybe just a little bit too much of a hot potato.”56 In a related vein, because a caucus's focus is so much narrower than a committee's, the caucus can help to coordinate activity across committees when an issue touches the jurisdiction of many panels.57 Moreover, caucuses can emphasize issues differently,58 provide information to which non-committee members do not have access,59 and be used to complement committee experiences.60 These differences emphasize that LMOs fulfill roles relating to the spread of information in Congress that committees do not.

The Low-Cost Nature of LMOs LMOs are cheap, a key characteristic that allows them to serve as solutions to an institutional collective action dilemma concerning the great demand and insufficient supply of information in legislative politics. The costs associated with participating in LMOs are quite low for most members and, given the groups’ voluntary nature, members only ever need to invest as much as they see fit. The low-cost nature is important because the

informational deficit in legislative politics has much to do with lawmakers’ lack of time and resources. Given the great demands on their time, legislators would be unwilling to invest substantial effort in LMOs if participation made them worse off. Hence, LMOs offer net benefits only if their relational and informational advantages outweigh the costs of taking part. Our respondents highlight that the costs associated with participating Page 157 → in LMOs are indeed small enough to not serve as a disincentive.61 Attending LMO meetings does “not [take up] a huge period of time,”62 especially when “there's a lunch. I have to eat anyway.”63 In other words, legislators can reap the relational and informational benefits associated with LMOs without putting in a great deal of effort. These groups offer an open invitation to free-ride because “no member is obliged to join an intergroup.”64 However, someone must assume the costs associated with creating and maintaining LMOs, and there are both insiders and outsiders who bear those costs. Inside the legislature, individual members with enough of a personal stake in an LMO's issue or cause are willing to serve as LMO leaders, a role that requires them to collect and disperse information, organize LMO meetings and other activities, and coordinate efforts when LMOs seek to influence the legislative agenda or affect a particular piece of legislation. Legislators may choose to engage in these efforts because an LMO's cause is especially close to their hearts, because they represent constituents who have a keen interest in the LMO's issue, or because they seek to build reputations as experts and go-to people for the issue at stake. For example, says one leader of an intergroup, “I have a genuine interest in these issues, and I think they are fighting a just cause.”65 Another member of an intergroup's bureau similarly maintains, “I joined because [it is] a topic that's of special interest to me.”66 A third explains that she is chairing her group because “nobody else would do it.”67 In other instances, legislators join or lead particular LMOs because they have personal stakes in the groups’ issues or causes. Examples would be a farmer chairing an LMO that deals with agricultural issues, a multiple sclerosis patient leading the multiple sclerosis group, or the relative of a firefighter running a fire services LMO.68 This dedication on the part of individual leaders and their successful cooperation are essential for LMOs’ effectiveness, as several of our respondents emphasize: Groups succeed only when the cochairs are committed and active.69 In return for their engagement, LMO leaders receive not only the knowledge that their efforts help advance a cause that they value but also recognition of their leadership both inside and outside the legislature. As one MEP puts it, “That's certainly one of the base political motives, that it can help [your] profile.”70 In Congress, the payoff in terms of professional profile or reputation is a recurring theme, and a number of respondents maintain that their bosses benefit from serving as caucus cochairs because their involvement causes others to perceive them as leaders on a given issue, with capacity, expertise, and legitimacy.71 Using a hypothetical example, one respondent says, Page 158 → If my boss wants to go to Liechtenstein or introduce a bit detailing tax treatment of Liechtensteinian wheels or something, he can say, “Well, as a member of the Liechtenstein caucus for the past 15 years, I've been following the situation blah, blah, blah.” It's got legitimacy. And he also has the standing, then, to go to other members of the Liechtenstein caucus and say, “Hey, will you sign on to my bill about tax treatment of Liechtensteinian wheels?” So it's really, it's establishing a framework that is there if you need it. It's just kind of capacity building, I think.72 Another respondent notes that caucus involvement makes members more influential and that “that's one of the reasons why you would join a caucus anyway—so that you can become more knowledgeable and more influential.”73 He goes on to claim that even the White House contacts this representative's office more frequently because of its well-established caucus leadership.74 There are, in other words, potential payoffs in serving as LMO leaders and these payoffs provide an important incentive to cover some of the costs associated with running the groups, and having committed insiders is critical to the LMOs’ success. However, LMOs are created not only by particular legislative offices but also by outside

advocacy groups that maintain privileged relationships with LMOs and support the groups by providing legislative subsidies to its members. This involvement of outsiders is an important reason why the costs associated with LMO activity inside the legislature are quite low. Providing legislative subsidies to LMOs and their members is one of two key functions outside advocates fulfill; the other is serving as a supplier of both political and policy information. We discuss these two functions in the remainder of this chapter, highlighting the nature of the relationship between LMOs and outside advocates, the ways in which outside advocates supply legislative subsidies to LMOs, and patterns of information flow, both from the inside out and especially from the outside in.

Privileged Partnerships with Outside Advocates Intergroups in the EP and caucuses in the U.S. Congress establish ties between insiders and outside advocates who share a concern for a common cause—interest groups, trade associations, lobbyists, or foundations. In Page 159 → many instances, these ties constitute privileged relationships that are institutionalized in LMOs in the sense that legislators and outside advocates regularly interact, share resources, and exchange information. Strong ties, based on regular and iterative interactions, provide the basis for mutual trust as well as an incentive structure for insiders and outsiders to maintain the privileged partnership, from which both sides benefit. Outside advocates have exceptional access to a set of legislators who, by joining a given LMO, have signaled their commitment or support for a cause. This access gives outsiders opportunities to supply information to these lawmakers, to help set the legislative agenda, or to ensure that their issues remain on legislators’ radars. Legislators, in turn, enjoy the legislative subsidies provided by the outside advocates and gain access to reliable information sources. Indeed, the institutionalization of privileged relationships in LMOs creates an incentive structure for outside advocates to supply high-utility information that is reliable, research-based, and easily digestible for busy policymakers and their staff. We address the question of information exchange between insiders and outsiders after discussing the nature of their relationship in more detail. Outside Advocates as Providers of Legislative Subsidies In Brussels, the most institutionalized relationships between intergroups and outside advocacy organizations are those where the outside organization provides “human resource” assistance, which generally means serving as the secretariat of the intergroup. In addition, a number of outside advocacy organizations have privileged relationships with specific intergroups, even if their cooperation does not amount to the outside group formally acting as the intergroup's secretariat (although indications are that in some cases the groups may informally do so). A number of these relationships go a long way back in time, such as the links between the Trade Union Coordination Intergroup and the European Trade Union Confederation, the Animal Welfare Intergroup and the Eurogroup for Animals, the Hunting Intergroup and the Federation of Associations for Hunting and Conservation in the EU, as well as the Intergroup on Ageing and the AGE Platform. Table 6.1 provides a list of intergroups with links to specific outside organizations.75 While some intergroups are established because one or more legislators decide to pursue a particular issue or cause—for example, the Tibet Intergroup, which was established in 1989 by Michel Hervé (Socialist Group, Page 160 → France) (Mann and Reuter 2010, 405), or the Animal Welfare Intergroup76—the impetus for others comes from outside advocates. ATD Fourth World, an interest group fighting extreme poverty,77 was the driving force behind the establishment of the Fourth World European Committee Intergroup, which later became the Intergroup on Extreme Poverty and Human Rights.78 Similarly, L'Union Sociale pour l'Habitat, an interest group that represents social and low-income housing organizations in France,79 actively promoted the establishment of the Public Services Intergroup, first by seeking the support of French MEPs during the run-up to the 2009 EP election and later by also targeting MEPs from other member states.80 The activities of outside advocates trying to support the establishment of a new intergroup or ensuring the continuation of an existing one can be quite extensive and can include directly contacting large numbers of MEPs, seeking support from key MEPs across different political groups who may be able to bring others on board, and having constituents contact their representatives to urge them to support a particular group.81 One MEP describes such efforts explicitly as “campaigns…funded by

outside organizations.”82 Page 161 → Outside advocates are also often closely involved in setting the agenda of the intergroup once it is established,83 and the impetus for action can generally come from both sides.84 As one outside advocate explains, her intergroup is “coorganized” by insiders and outsiders.85 The association between specific intergroups and outside organizations may be so close, in fact, that the groups do not engage in any activities that exclude their primary outside partner.86 In some instances, the two may even at times be seen as a single entity. One MEP observes that “in the Parliament, MEPs get sometimes confused about what is done by the [Animal Welfare] Intergroup and what is done by the Eurogroup for Animals, and sometimes MEPs call the intergroup ‘Eurogroup.’”87 MEPs are, of course, aware that sensitive ethical questions can arise when interest groups finance an intergroup's secretariat, invite intergroup members on trips, provide sponsorship for receptions and meals, or pay for interpretation at intergroup events.88 Intergroups are subject to the EP's rules concerning transparency and the declaration of financial interest.89 The financial interest declarations for the EP's 2004–9 term show that 10 intergroups received human resources support from outside organizations, as did 8 intergroups in the early stages of the 2009–14 term (table 6.1). No intergroups listed other types of support in 2010 or 2011, and only two groups did so in their 2007 financial declarations: the European Disability Page 162 → Forum covered travel and accommodations expenses for speakers at six intergroup meetings in Strasbourg, while the Lesbian and Gay Intergroup received unspecified financial support from the International Lesbian and Gay Association. Only the Peace Initiatives intergroup declared that it received financial support directly from the political groups and MEPs, a practice that is also common in some issue groups.90 Despite concerns about the intergroup system and its links to outside interests, MEPs are quick to emphasize that the contacts they establish with outside advocates through intergroups are critical to their work91 as long as a balance is struck and different points of view are taken into consideration.92 Intergroups can even be helpful in this regard because they promote access to different points of view and thus the triangulation of incoming information.93 Furthermore, many intergroups are associated with public interest groups rather than particular industries (table 6.1), a reality that some MEPs maintain mitigates the ethical concerns raised about the intergroup system.94 Finally, MEPs stress that they ultimately make all decisions and that they bear responsibility for evaluating the input they receive from outside advocates.95 To avoid appearing overly dependent on outside advocacy organizations, some intergroups have made attempts to professionalize their secretariat and to keep it in house. The Intergroup on Lesbian, Gay, Bisexual, and Transgender (LGBT) Rights, for example, has a secretary who works inside the EP who is formally a member of the intergroup leader's staff, but the secretary's responsibilities relate exclusively to the intergroup.96 Others, however, see it as an advantage if “the secretariat is done by an organization that is not really a part of the European Parliament [because] the intergroup is organized in a more coherent and consistent way.”97 It may also increase the group's outside visibility, help prevent the intergroup from being associated with a single MEP (as might be the case if the intergroup is run out of a legislator's office), and ensure that the group continues to operate smoothly in the case of personnel turnover. Of course, outside advocates are aware of the potential shortfalls of straddling the line between providing administrative support as a “legislative subsidy” (Hall and Deardorff 2006) and their core functions as interest groups. In our conversations, therefore, outside advocates emphasize that they try to separate their activities as lobbyists and their roles as secretariats of specific intergroups. One claims that “we keep rather neutral,” explaining, “If I am one-sided as the secretariat, the MEPs would not appreciate my Page 163 → work any more…. I mean, this should be providing a service, facilitate some work for them, but not to be too one-sided.”98 Another outsider agrees: We are not there to defend our ideas, our interests. We are there to put everybody on the table, so we're not there to lobby the Parliament. So in practical words, it means that we do everything except

the political stuff, which is done by the MEPs. So we prepare everything; we're in charge of the logistics. We're in charge of drafting press releases, reports. We are meeting with MEPs and maintaining close contacts with them.99

A third outsider similarly maintains, My role is only to give the input and to help them to do what they want to do. But I never interfere in the decision. I am not a member of the board. I never set the agenda of the meeting…. So it's totally managed by the board…and I am only there…. I'm totally shadow acting in this one. And I think it is very important because I have no legitimacy…. I try to create consensus…. I try to make a bridge.100 If successful in such efforts, outside organizations can establish a long-term, trusting relationship with intergroups and their leaders. One outside advocate describes such a relationship based on strong social ties: The previous president [and I] knew each other very, very well, and he knew that he could trust us. We knew the areas that he was interested in, we knew the kind of workload that he had, we knew…what we could ask from his office, what we could not ask for, what was our job, what was his job.101 We also find evidence for privileged relationships between insiders and outsiders in the U.S. Congress, where many respondents highlight that the caucus organization allows them to stay in touch with outside organizations.102 Indeed, a large number of our interviewees discuss their caucuses’ relationships with outside groups, or the “usual suspects”103 with whom they have routine contacts.104 The number of these “usual suspects” seems to range from one or two (more common) to between seven and ten (less common), but no matter how many, outside groups are catalysts for much of the activity of many caucuses. According to one staffer, Page 164 → Advocacy groups are a large driving force behind these caucuses. They sort of give us our direction. We of course have to look over it and approve it, and we don't follow them blindly by any means, but we look to them as sort of experts in the field, and they bring a representation of America.105 Table 6.2 lists select caucuses and their primary outside affiliates. Outside advocates are quite conscious of the legislative subsidies they provide. One representative of an outside organization puts it bluntly: “We do essentially all the work, and frankly, if it wasn't for [our organization], I don't think the caucus would have survived.”106 An insider agrees that the outside organization “tends to sort of be the leader and do the legwork and be the organizer.”107 This work consists of organizing events and bringing in speakers,108 sending out invitations for meetings,109 drafting letters,110 generating reports,111 direct advocacy,112 recruiting new caucus members,113 research,114 communicating with offices,115 and hosting events.116 Moreover, outside advocates can be the main impetus behind the creation of caucuses.117 As a representative of one outside group maintains, “It's part of Lobbying 101 to set up a caucus. If you want to support an industry, the first thing you do as a matter of course, is to find friendly [members of Congress] who will support the establishment of a caucus.”118 Another concurs, “Most [caucuses] started because a group came to the member.”119 Some outside groups also support a caucus event by paying for refreshments: At our briefings we always provide some type of food. It gets the 21-year-old staffers to come out. [It] is usually nothing more than sandwich platters, cookies, a fruit bowl, and iced tea. And we spread the costs among the cosponsors of the event, so it usually comes to $100 or something. Not, you know, not a huge amount. But it is a big hit in terms of visibility.120 Like all legislative subsidies, this type of support by outside advocates to caucuses is subject to congressional gift

and ethics rules. Various respondents seem to understand or interpret the limits of what outsiders can or cannot do somewhat differently, however. While some were open about telling us about maintaining membership lists,121 writing letters,122 and recruiting members,123 others were more cautious: Page 165 → Page 166 → [Outsiders] have to be a little careful about the services we provide. We really can't do the work of the caucus. Otherwise, the Ethics Committee would be coming down hard on everybody. We do briefings, we'll put together white papers. We facilitate meetings. We can't write letters. We can't do their caucus recruitment. The caucus has to do the caucus recruitment.124 In some instances, legislative subsidies go as far as helping to draft legislation. According to one representative of an outside organization, “Sometimes we get phone calls from members of Congress saying, ‘We're about to introduce a piece of legislation. Here's the draft. Can you take a look at it for us?’…We'll take a look at that bill and tell them whether or not…this particular piece of legislation will actually have a positive impact.”125 A respondent inside Congress tells a similar story: “We maintain relationships with [outside] organizations and bounce things off them occasionally. If we see something that is coming down the pipe, [we] ask them what they think.”126 The involvement of outsiders in the conceptualization of legislation may extend further than just double-checking legislative proposals, as one outside advocate explains: We just recently had a bill introduced in the House and the Senate that we developed in conjunction with the [outside organization]. So I did the draft, and then we worked it out between the advocacy organizations and then brought it to the Hill staffers, who then brought it to their members.127 In whatever form, legislative subsidies are generally welcome from the point of view of legislative offices involved in the caucus system, since they help to “take the burden off”128 and alleviate the “onerous workload.”129 Given the often tremendous investments outside groups make in “their” caucuses, the insiders might seem on the surface to be the main beneficiaries of the privileged relationships between legislators and outside groups. In reality, however, this relationship is quite symbiotic,130 with organizations also reaping important benefits from the privileged access they gain via the caucus system. One rather simple advantage that is nonetheless of tremendous value to outside groups concerns the fact that an outside group cannot set up an event, book a room, organize a reception, or host virtually any public event on Capitol Hill without the help of a member. Consequently, Page 167 → if…an outside organization wants to come up and do a briefing, they are required to go through a member office. A member has to secure a room [and] has to set that up for the logistics. But as far as paying for communications, things like that, if the outside entity is doing it, they have to bear those costs.131 Therefore, groups that want a regular or physical presence in the Hill's competitive lobbying environment benefit greatly from having well-established ties to members. The primary benefits outside advocates enjoy are not simply logistical, however, but also relational and informational. One outsider, for example, points out “that we have access to lawmakers, that we can educate them and their staff, [and] that we handle issues important to our community.”132 Particularly compelling from his point of view is the fact that the caucus gives him access to lawmakers who have already expressed an interest in the topic. Another outside advocate concurs: “It's a way for us to identify legislators that will help us on an issue.”133 Caucuses, in other words, not only connect legislators to other legislators with whom they would not otherwise be connected but also help outside groups develop relationships that they would otherwise lack. Emphasizes one outside advocate, the caucus “allows us to build a relationship with someone that we wouldn't otherwise

know.”134 Since representatives typically want to talk only with groups affiliated with their constituencies, some groups with very broad-based interests can have difficulty gaining access to offices. The caucus solves this problem because members have already indicated their sympathy toward an issue, thereby providing an opening for the outside advocate. One staffer describes the relationship-building aspect of the caucus network as a sort of critical link in a chain of communication: Outside groups “come talk to the cochairs, and then we can put them in touch with other people or help them with an event. It really helps to link up everyone who's interested.”135 Caucuses can serve as a clearinghouse for issues that do not necessarily map easily into congressional jurisdictions. For example, the same respondent notes the widely dispersed industries and advocacy groups that have some interest in their topic, among them a manufacturing industry, health groups, and municipal planning organizations. Without the caucus, these outside groups would have no way to coordinate their activities or to learn about what is going on with their issue because they could rely only on their relationships with disparate legislative committees. Page 168 → A caucus, in contrast, is not subject to restrictive jurisdictions and is therefore more flexible. Information Exchange between Insiders and Outsiders In line with our theoretical propositions, the exchange of information lies at the core of the relationship between LMOs and outside advocates. Some of this information may be political—for example, concerning the preferences of key constituencies and stakeholders in a given policy area. In the European Parliament, one respondent maintains that the positions of the European Disability Forum are “representative of the interest of people with disabilities,”136 which is “very valuable” from the perspective of legislators intent on making good and efficient policy.137 In addition, outside advocates may even provide information about the preferences, positions, and strategies of other EU institutions, gathered independently through their own “Brussels networks.” One respondent reports that some of her colleagues in the interest group have “nothing to do with the intergroup” but work primarily with the European Commission, enabling her to learn about developments there and subsequently to pass along that information to her contacts in the EP intergroup. She notes, “If you are well informed, you are appreciated as an information deliverer.”138 Insider information about events in other institutions also allows outside organizations affiliated with particular intergroups to help track legislation and let their contacts know if something of interest is in the policy pipeline.139 These contacts in other EU institutions and the great time constraints intergroup leaders face140 often allow outside advocates to have information about proposed legislation before MEPs become aware of it.141 Most information exchanged between insiders and outsiders, however, is substantive information about the content and likely consequences of policy proposals. Intergroups are, after all, designed to serve as forums that encourage “experts and expertise coming into the house…. They are a place to listen to new ideas.”142 One source of these ideas is the participation of outside advocates as members of the audience at intergroup meetings, where they can ask questions, make statements, and seek conversations with legislators after the meetings. Outside advocates also serve as formal participants in panel discussions and help identify and invite experts on a given issue. The presence of these outsiders turns intergroup meetings into “almost a kind of academic environment,” according to one MEP, where “you have longer discussion and the discussions are a little bit more theoretical Page 169 → in nature…. So intellectually it is more interesting.”143 This aspect of the intergroup system even appeals to one MEP who generally discounts the importance of intergroups but acknowledges that if there is an interesting speech, if they invite someone interesting, I go and listen. So for me it is a possibility to learn. And there was this speaker some weeks ago…on conditionality of development linked to the welfare of the farm animals. Okay, so this was something I had never heard about, so it was interesting for me to go and listen. Very useful…. I think that the Animal Welfare Intergroup is useful as…a learning tool…and I think they are very much having more of a research agenda. They really—they provide high-quality presentations…. I would say it is more an academic seminar. You can learn a lot.144

The high substantive quality of the information provided to MEPs in intergroup meetings is one of the key features that qualifies it as high-utility information. A second is that it is presented to MEPs in an easily accessible and digestible format.145 Intergroup meetings offer the opportunity to meet and learn from a variety of stakeholders at the same time, to be briefed efficiently on an issue of interest, and to begin connecting substantive information with the political side of the deliberation and negotiation process. Intergroups also provide access to information in the form of position papers or policy briefs prepared by outside organizations and distributed to members either via email or at the intergroup meetings.146 In some cases, intergroup meetings may in fact be the only place where legislators have access to position papers by the outside organization.147 Relevant materials are often distributed directly to attendees as they enter the room, a practice that some MEPs appreciate because it allows them to “pick up materials neutrally, when otherwise I would have to ask an interest group directly, which would then think, ‘Great, we got another one.’”148 The informational materials distributed through intergroups are generally easily digestible in that they are short and to the point, as several representatives of outside organizations emphasize.149 One, for example, highlights that “we try to produce a very short paper” that allows MEPs “to understand what they do and what they have to vote for.”150 Another outside advocate highlights that “MEPs like to have” a short briefing that is “formulated in key sentences” and includes “information they can use in their speeches.”151 Several MEPs emphasize that intergroups provide an opportunity to Page 170 → stay in touch and communicate with outside advocates “without having to have a lot of appointments.”152 Indeed, one respondent suggests that “one of the MEPs from another intergroup told me that he is creating an intergroup in order to bring together all the lobbyists and stakeholders, because otherwise he is never finishing and listening to people again and again and again.”153 When asked what makes intergroups useful as a source of information, a third legislator similarly explains that “this place…is full of information…like a waterfall of information. What intergroups can do is they can filter out information related to the topic in a connected manner,…and that's very helpful.”154 The benefits in this regard are mutual, since outside advocates can use intergroup meetings to identify potential supporters and to reach many MEPs at once.155 Furthermore, intergroup meetings also give outside advocates opportunities to network with each other, as one MEP told us: “They come there and they meet each other, and they can make plans as well, while they are at the meeting.”156 The policy information MEPs receive in intergroups often directly relates to a specific dossier that is currently on the EP's agenda. Boessen and Maarse (2009, 225), for example, describe how the Health Forum Intergroup organized a meeting on tobacco advertising at a time when a ban on tobacco advertising was actively being debated in the EP. Similarly, the Trade Union Coordination Intergroup held meetings on the Working Time Directive,157 the SME Intergroup (Small and Medium-Sized Enterprise) held meetings on the directive on combating late payment in commercial transactions,158 the Public Services Intergroup held meetings on the socalled Monti-Kroes package159 on state aid,160 and the Animal Welfare Intergroup held meetings on the regulation of animal transport.161 In these instances, an intergroup may invite a representative of the European Commission, the sole institution with the power of legislative initiative;162 a representative of the Council, the EP's co-legislator;163 and someone from the relevant EP committee that handles the dossier. The invited speaker from inside the EP often will be the rapporteur, who is charged with ushering the proposal through the legislative process on behalf of the Parliament.164 In addition, “two or three experts” may be in attendance to “do their presentations”;165 these experts may be interest group representatives,166 scientists, or scholars.167 Such intergroup meetings seek to provide information about the content and expected consequences of the proposal, bringing together interested actors from various committees and parties, outlining the positions of Page 171 → stakeholders on the outside, and allowing all interested parties to discuss the matter in an informal setting. One MEP explains that “if there was one MEP who is making a report on disability and human rights…to support their work, we will discuss this at the intergroup, and then everyone else can also make proposals and interventions and to see that work through.”168 According to another MEP, It is…almost education…. Every so often one of the intergroups will…have speakers on a particular piece of legislation or a particular issue that is really interesting. And because it is not a huge period

of time…you can go there. I think it is a quite productive way of being briefed almost…. Sometimes what you want to do is actually pursue something in more detail so that you can actually understand what it is that you are being asked to do…you know, what are the implications of this piece of legislation?169

Not only is policy information exchanged at intergroup meetings, but MEPs may also request it from the outside advocates their intergroup is affiliated with.170 Moreover, if those outsiders do not know the answers to the legislators’ questions, they know where to find those answers. One outside advocate maintains that “when they ask me something, in like 90 percent of cases, I'm able to answer their question. When I cannot answer the question, I will then forward it to a national organization…. I'll find a contact for them.”171 In this sense, MEPs can act as “experts in using experts” (Heclo 1978, 103). The involvement of some outside advocates affiliated with intergroups goes as far as drafting specific amendments concerning a particular piece of legislation.172 Such is, for example, the case for the Trade Union Coordination Intergroup and its outside ally, the European Trade Union Confederation; the Animal Welfare Intergroup and the Eurogroup for Animals; the Disability Intergroup and the European Disability Forum; the Fourth World European Committee and the ADT Fourth World; and the Intergroup on Ageing and the AGE Platform Europe. One outside advocate173 acknowledges that drafting amendments “is one of our key activities,” and a member of the bureau of the affiliated intergroup highlights how he appreciates receiving “the expert knowledge that you need when you make the legislation…. With my limited staff, it would be very, very difficult for me to make such detailed amendments.”174 In fact, MEPs at times explicitly request amendments: In the words of one outside advocate, “Sometimes Page 172 → [MEPs] ask to send some proposals of amendments on specific reports. [They] send us a draft report and they ask us, ‘If you have some amendment, please send it to us,’ and that's it.”175 A third outsider confirms this process: We draft possible amendments for the rapporteurs, give suggestions. And then if the report is ready, we will draft amendments for others…. It's highly appreciated by members of Parliament if they don't have to do the work themselves, especially if you offer value. If we are capable of making good amendments, we are highly appreciated…and we can do the work for them.176 Of course, the drafting of amendments relates to specific pieces of legislation that are actively debated in the EP. Yet not all information exchanged in intergroups relates to such specific legislative proposals. In fact, intergroups’ informational mandate seems to be quite a bit broader and include a general educational function for MEPs, as several respondents discussed how they learned something important at an intergroup meeting that might later come in handy. One MEP maintains that There are so many things that go on and that's one of the reasons why I went to the diabetes intergroup and then I took interest in that because they brought up things I did not know, on the development, on the costs, and so on…. They bring up many things you do not know, what the size of the problem is. It is always interesting to be informed about that…. I learned a lot at these meetings. And when it comes up again, when one of these things comes up in reality you can say, “Okay, yes, we will have this and that.”177 In a similar vein, another MEP indicates that You never get out of one of these events empty or with less than when you went there, you have access to the information even if you were not there. Somehow, there is always something that then creates resonance when you are dealing with other issues, that you make the links and connections, and you always learn something…from this.178 In Congress, 27 respondents mention the importance of information exchange when discussing the relationship between caucuses and outside Page 173 → groups, including all of the representatives of outside organizations with whom we spoke.179 One staffer explains, “You're bringing someone from the outside in that is either a

specialist in, say, voting [or] elections…or you're dealing with the nuclear issue, so you're bringing in a very specific scientific minded person, [or] the business community if you're investing in that.”180 In other instances, caucuses may bring in visitors who are not experts on technical issues but practitioners or laypeople who can relay firsthand, personal experiences on an issue.181 Finally, some outside advocacy groups organize field trips so that members and staffers can learn about an issue. According to one respondent, “There was an actual day trip. I don't know much about the [issue], but I learned so much that day.”182 Similarly, the Fire Services Caucus and its institute organizes an annual trip to a national firefighting training facility near Washington, D.C., where members and staff have the opportunity to operate equipment: “We put them in a burning building,” as one respondent puts it.183 Another maintains that caucuses “give us some mechanism…to call upon if we need data from outside sources, ”184 which often have resources and expertise that staffers and members lack. In fact, caucuses and groups that deal with more technical issues seem to rely more on the information provided by outside groups than groups that focus on less technical issues.185 Said one outside organization representative, “I have a list of all the staffers for all the members [of the caucus], and I'm constantly sending them stuff.”186 One respondent on the inside confirms that the groups “provide a lot of technical assistance and recommendations, because we're not the experts.”187 According to another insider, staffers don't want to have to learn how to speak the language of gastroenterology. Matt does not have the time to be his own translator on that. So I will spend two hours on the phone,…and then I'll distill that into five bullet points and email that to Matt.188 In the words of still another insider, The caucus can only do so much. It goes back to personal relationships that stakeholders have with staff and members of Congress. As staff on the Hill, we deal with a lot of stuff from the legislative constituents. It makes our lives a lot easier if people from the outside…point us in the right direction to the right information so that we can quickly read…. We don't have the time Page 174 → to do the research ourselves because there's a wealth of information out there, and it's hard to—and given the fast-paced nature up here, you have to get—have access to that information quickly.189 And yet another says, I have a place to turn to that can get me information fast, right? That's extraordinarily helpful when I'm already putting in 16-hour days every day. And the congressman is putting in even more. So they're extraordinarily helpful resources.190 The outside advocates themselves are not necessarily able to provide the desired information, but they can identify who may be able to help in a given circumstance. One staffer explains, When we need information, we're able to say, “How I can find this information? Can you help me find this information? Put me in touch…with the right person.”…They're very knowledgeable about their sort of area of expertise, and we're able to tap into that, which is really beneficial.191 Caucuses also rely on groups to provide on-the-ground information, often about policies that are needed or that need to be changed. Interest groups typically have close contacts with practitioners or are made up of practitioners and consequently have high-quality information about how laws work in practice or about the potential effect of a new law. One outside advocate explains that there may be “sections in the piece of legislation, provisions in a bill that don't have the impact that the members intended, and so that's where our area of expertise can really come in handy.”192 Another respondent cites a case where constituents provided information to an outside group about the desirability of a particular policy change; the group brought that information to the caucus, which then pursued the necessary policy change. The respondent speculates that the relevant people in Congress might not have otherwise known about the importance of this particular issue.193

Of course, outside groups associated with particular caucuses are not the only sources of information available to congressional offices, but a great advantage is that outsiders are quite eager to help out their contacts inside the chamber. For this reason, they work quickly to provide legislators and Page 175 → their offices with information. While Congress has access to various forms of in-house expertise (e.g., committee staff, the Congressional Research Service, the Government Accountability Office), outside groups may have an advantage over these institutions because of the speed with which they can respond to requests for information. The in-house institutions designed to address members’ needs for information serve all members of Congress, and while they produce information and reports that are of very high quality, they typically have to wait for a request to put a report together before one is established. Staffers must wait for research to be assembled. In contrast, if staffers contact reliable interest groups with whom they “already have a working relationship,” the outsiders “will quickly provide information.”194 In addition, we find some evidence of groups providing political information. A number of caucuses, for example, hold meetings with outside organizations to define common objectives. One respondent describes such an occasion: “The whole point of the meeting is to get the [outside organizations] in one place,…and we talk about priorities—mainly priorities for the upcoming year.”195 Other respondents acknowledge that having contacts with outside organizations gives caucus members “ideas of what is going to be well received.”196 Political information also flows from inside the legislature to the outside, however, since outside advocates “want to know what Congress is saying and what Congress is hearing.”197 One staffer told us that “they'll certainly call us and ask, ‘The…bill, is that going to move through committee next week? What have you heard?’ And we'll provide whatever intel we have…. They rely on congressional staff for updates.”198 However, we found fewer examples of political information exchange in Congress than in the EP. This result is in line with our earlier finding about information exchange inside Congress and the EP, where political information seems less common in the case of the former. The Supply of High-Utility Information The reliance of legislative offices in both the EP and Congress on outside advocates for the provision of political and policy information raises questions about legislators’ undue dependence on lobbyists and interest groups and its consequences. Of particular concern is the possibility that outsiders with institutionalized access to LMOs would take advantage of their privileged positions by feeding legislators information that is biased in favor of their own policy objectives. Respondents are quick to emphasize, however, Page 176 → that outsiders’ input is not accepted wholesale.199 In fact, one respondent in the EP claims that the intergroup allows her to be more independent of outside interests, because rather than approach someone on the outside and risk “a kind of dependency,” she can go to an intergroup meeting and listen “neutrally” to what is being presented.200 Outside advocates also emphasize the limits of their influence. One sees his role as “explain[ing] what are the facts, what are the consequences,” not “direct[ing] someone.”201 Another similarly suggests that “it is more a question of training and education of MEPs than a way to influence their way of thinking.”202 As a result, some MEPs do not even perceive intergroup meetings as lobbying events: “I would say it is more providing information than lobbying. Lobbying is done with people who are coming in and visiting us, giving us their view…. I do not see it as lobbying. I do not have the feeling of lobbying when we participate.”203 We observe similar dynamics in the Congress, where some outside advocates do not interpret what they do as lobbying in the conventional sense. One argues that our job is to put good information in front of them. Education obviates lobbying. We don't walk into the gray area. We don't ask for votes, just ears…. They don't see us as advocates but as good sources of information…. We are seen as a place where a lot more information is presented in a very balanced fashion. The Hill does not see us as advocates but as educators.204 Another similarly claims that their “mission is societal. We're unbiased. We're just a bunch of engineers with…a certain area of expertise. And we're just providing the science. ‘Here's the information. You guys do with it as you will.’”205

Such statements are, of course, cause for skepticism, not least because outside advocates acknowledge that they hope that legislators affiliated with their LMO ultimately will “influence and…speak to member colleagues who can have the power to make a change.”206 Accordingly, interest groups should not be expected to perform a purely informational role given their high stakes in a given policy area. And indeed, respondents do not maintain that the information that is exchanged in LMOs is entirely neutral. Our respondents do, however, describe how LMOs help prevent outside advocates from supplying information that is redundant, inconsistent, unsubstantiated, or even misleading and how they offer incentives for the provision of Page 177 → high-utility information. Our interviews identify three principal incentives, all of them in line with our expectations: the high levels of embeddedness in a network of knowledgeable actors, the possibilities for triangulating information received from various sources, and the incentives associated with the privileged relationships enjoyed by outside advocates with close ties to a given LMO. The institutionalization of the relationships between legislators and outside advocates through LMOs simultaneously provides all three incentives. The high level of social embeddedness of active participants in the LMOs from both inside and outside the legislature means that, as one MEP put it, “Everybody knows everybody else, so it is quite transparent.”207 Moreover, many (though not all) legislators affiliated with intergroups are policy experts themselves,208 and this statement almost universally applies to the intergroup leaders.209 Therefore, outside organizations would have difficulty advocating positions that lack proper foundation; in fact, MEPs often “do not agree with the speakers…and they have a debate.”210 The social embeddedness of outsiders in the LMO system means that they may even be consciously included in the search for a policy solution. One congressional staffer told us, “My boss just introduced a bill yesterday…and we were working with all of those organizations to get their input on the legislation before we introduced it so we could get their support for it.”211 Similarly, an EP respondent emphasizes that institutionalized exchanges between MEPs in Trade Union Coordination Intergroup and the European Trade Union Confederation “include the labor unions in a search for consensus inside the Parliament.”212 LMOs also allow legislators to triangulate and thus verify information received from various sources.213 Because LMOs are one of several sources of information for legislative offices, they provide opportunities to check information received elsewhere. One respondent in Congress notes, “Everybody provides us with information. Everybody lobbies Congress. And every staff develops their network of people that they rely on”;214 along the same lines, an MEP emphasizes, “I have said to lobbyists…if you are telling me nonsense, I will find out, because I will always check the information you give me.”215 Moreover, some LMOs are set up in such a way that they provide access to varying points of view—for example, by inviting speakers who represent different stakeholders’ views on a given issue.216 Three lengthy quotations describe this kind of setup. Page 178 → When I have a question about the castration of pigs and there is an [Animal Welfare] intergroup meeting on the topic, they may invite various groups. I may have Pfizer with a vaccine; people who are advocating castration using better anesthetics; those who maintain male pigs ought to be slaughtered at an earlier age; and university representatives presenting their research. I would end up with five different positions on how to avoid the painful, traditional castration procedures. I could talk to each of the five individually, or I can ask the intergroup to invite all five, where they present their positions and then discuss amongst themselves. I can listen and get a lot out of it…. And everybody can come and listen and we can think about this in political terms as well. And I can have all that in one afternoon.217 When we organize a meeting [of the Intergroup on Climate Change, Biodiversity and Sustainable Development], we try to have a balanced discussion. It means that we invite the three institutions, the presidency—it depends. But, always the European Commission plus NGOs, private sector, and scientists. So in one hour and a half or in three hours, [MEPs] have access to a lot of information, and they can use more easily this information than the information that they can have through bilateral meetings. [At our events,] there will be Bird Life [a global alliance of conservation organizations] and Syngenta [a global agribusiness company]. Well, they do not share the same views…. After all the

presentations there is always a debate. So, yes, they can confront [each other's] ideas.218 It is very common and regular that we contact those actors we recognize as spokespersons and…try to get from them all their positions and their statements on specific subjects related with the subject we are dealing with…. The main intention since the beginning was to hear as much as possible and as different positions as possible in order to have a broader vision and scope of the issues…. Intergroups are open windows to outside institutions, NGOs, organizations, consumer associations, and other actors and stakeholders…. Here you have a room and an open door and an open window to interact with other stakeholders and groups of interest, which have something to say on these subjects.219 We also find instances of this kind of triangulation in Congress, where one staffer describes a caucus event: The group “kind of had a roundtable discussion on it. We invited speakers—say, five, four or five speakers—and a Page 179 → moderator, and we tried to get all different points of view on topic…. And so we had a bunch of people, we had five or six speak.”220 Along similar lines, one staffer describes how “Exxon and Audubon can sit at the same table” during a caucus meeting.221 However, not all intergroups or caucuses make the same efforts to accommodate every perspective on a given issue, and not all provide balanced discussion panels at their meetings.222 One respondent in the EP describes how she left a meeting of an intergroup early because she “found the speakers biased…. They were trying not to educate me but to indoctrinate me.”223 However, such a position seems to be in the minority among our respondents. In addition, several maintain that triangulation of information takes place at LMO meetings even if the invited speakers fail to represent all possible positions on a given issue, simply because of the openness of the LMO system. One outside advocate who acknowledges the “exclusive relation” her organization has with one of the EP intergroups, explains that other interests are not left out: “MEPs all have a lot of different influences coming from the outside, and not only from [us]. All the NGOs are invited, all the actors are invited. It is not closed.”224 As a result, MEPs can not only interact with outside advocates but also see the stakeholders interact and debate each other.225 The situation is similar in Congress, where a variety of outsiders show up at caucus meetings. Indeed, participation is such that “there's always a group that I've never heard of before that comes.”226 As a result, it is unlikely that all information provided would be entirely one-sided. Finally, our respondents confirm the expectation that LMOs and the privileged access and relationships they grant specific outside actors increases the quality of the information that flows from the outside into the legislature. Outsiders are well aware that caucuses give them “a platform and access to Congress that they might not otherwise have”227 and that “it's in their interest to have good working relationships with member offices too…. At the end of the day, it's going to take pushing on the inside to get some of these things done.”228 Losing this privileged access would be costly, but is unlikely to happen as long as outside advocates maintain a reputation as a credible and reliable source of information—the bread-and-butter of any advocate in either Brussels and Washington. One outside advocate in Brussels whose organization has ties with an intergroup dating back almost 30 years, succinctly said that “good information makes the credibility of an organization: If you make mistakes, then you lose your reputation, and that is the worst thing which can happen to Page 180 → an interest group.”229 One MEP confirms that this same outside advocate “knows that if he lies to me once, the thing is dead. So he won't lie to me.”230 Other respondents highlight the idea that outside advocates “want to be on the good side of the argument”;231 when they are providing information to legislators, advocates seek “an opportunity to show off the goods.”232 These sentiments are also echoed in Washington: “Our reputation is important, and we have to uphold it.”233 The importance of providing high-utility information was a recurring theme in interviews with outside advocates, especially in Brussels. One respondent describes how her organization influenced the final content of one of the most high-profile pieces of legislation to make it through the EP during the past decade, the Services Directive. Had she and her colleagues “not been able to come up with technical, legal, good, high-quality proposals, [the

problem] would not have been solved.”234 An outside advocate in Washington similarly explains, “I know from that experience that [a member of Congress] is going to make you a priority not just because you gave him a check, but because you understand the universe of issues that they're facing.”235 And according to an interest group representative in Brussels, For us it is a very important principle: To really be a reliable partner…. That's one of our core principles: Always be honest…. If you have information,…it must be reliable, and not that someone can prove that it is just a lie to direct someone, because then you will completely lose the trust…. If you are not a serious player, you will soon lose your credibility…and that is your main asset here…. We want to show the world how good we are.236

Conclusion This chapter investigated the informational role of LMOs in more detail, building on the findings of chapter 5. Here, we demonstrated that LMOs are indeed valuable arenas for the exchange of both policy and political information, both between legislative offices and between insiders and outside advocates. The two chapters thus confirm our key proposition about the role and value of LMOs in legislative politics: They provide opportunities for legislators to establish social networks based on common policy interests, and those networks provide opportunities for the exchange of policy-relevant information. Page 181 → We find support for these core functions in both EP and Congress, yet our data also provide evidence that LMOs may serve as a means for lawmakers to signal policy priorities and preferences to constituents and that these groups are useful as arenas of political coordination. This finding is not surprising, given that the three functions of LMOs are not mutually exclusive. Signaling and coordination are secondary for most participants in the LMO system, however, and they are not of equal value for members of Congress and MEPs: Signaling is more important in Congress than it is in the EP, while the coordination function of LMOs is more valuable in the EP than it is in Congress. The informational role of LMOs, in contrast, has universal value for the majority of participants in both LMO systems. As we hypothesized earlier in the book, the plurality, single-member district system in the United States encourages legislators to signal support for a particular cause or policy issue by joining caucuses that reflect constituents’ preferences and priorities.237 Hence, electoral institutions not only help determine whether LMOs exist in a given legislature, as we established in our cross-national comparison in chapter 3, but also influence the role of LMO systems in different legislatures. Since joining caucuses is voluntary and participation is cheap, members of Congress have an incentive to join even if they have little intention to expend any of their limited resources on participation. Indeed, a number of congressional respondents emphasize that they join caucuses at the direct request of constituents.238 As one respondent put it, “A constituent may call and say, ‘Hey, I'm from Kazakhstan, and I'd love for you to be on a Kazakh caucus.’ The constituent might not even know if there is one, but they just want you to be on it.”239 Signaling support for a given issue by joining a caucus is quite cheap: I am not even sure what benefit it provides to us back home, but presumably for some [members] who want to emphasize that they care about an issue but do not really actually do very much, it is a convenient way of doing it…. You can vote against every appropriation that supports diabetes and [when your] constituents…tell you, “You know, we want you to support diabetes-related issues,” [you can still say,] “I am a member of the diabetes caucus. I care passionately about it.”240 In other words, as some respondents candidly explain, joining a particular (set of) caucus(es) has few costs,241 and the benefits may have a positive constituency effect.242 This does not mean, however, that caucuses should Page 182 → be dismissed as inconsequential signaling devices. Indeed, signaling is a form of political information exchange and thus a meaningful political action; Miler (2011, 889) quite appropriately speaks of “symbolic representation” in this regard. More important, LMOs’ various functions are not mutually exclusive, which means

that even legislators who join caucuses for signaling purposes have access to the information exchanged in the caucus network. Indeed, one of the key benefits of LMOs is that members are encouraged to free-ride on the relational and informational benefits. Given the incentives to use congressional caucuses for signaling purposes, it is not surprising that members of Congress join many more caucuses (on average 12 in the 109th Congress, 22 in the 110th, and 26 in the 111th) than MEPs join intergroups (on average, 1.68 during EP6 and 2.51 during EP7), as we saw in chapter 4. After all, the link between constituents and representatives is not nearly as tight in the EP as it is in the Congress, because MEPs are elected using proportional representation systems. Despite the use of preferential voting in some EU member states, MEPs are generally not directly accountable to a territorially defined subgroup of voters. Indeed, MEPs elected using preferential voting do not join more intergroups than their counterparts elected from closed party lists (see chapter 4), indicating the limited utility of intergroups as signaling devices. That public awareness of the EP and the salience EU citizens attach to it tend to be quite low further undermines incentives for MEPs actively to signal support for a policy issue to their constituents because signaling is less relevant to electoral success. A second meaningful point of difference between EP intergroups and congressional caucuses that emerges in this chapter relates to their cross-partisan nature. In both chambers, respondents emphasize the importance of crosspartisanship, yet it appears to take different forms in intergroups and caucuses. In the EP, the distribution of seats is such that only the two largest party groups, the center-right EPP and the center-left S&D, can together muster a majority large enough to pass legislation without the support of third parties.243 If the EPP and S&D do not agree, a coalition of at least three party groups is required for legislation to pass. It appears that this need for crosspartisan cooperation creates a setting in which intergroups are more likely to serve as arenas for political coordination, such as the search for compromise and a common position, attempts at coalition building across party lines, and the incidence of political Page 183 → activities (e.g., the introduction of amendments by MEPs from different party groups). Congress, in contrast, has only a limited need for cross-partisanship (or rather, bipartisanship)—for example, if the majority party is split on an issue or because of attempts at logrolling. Bipartisanship is less important because the majority party dominates the legislative process and usually does not need support from across the aisle, at least within one of the two chambers of Congress.244 Nevertheless, respondents emphasize members’ desire to engage in bipartisan action. While bipartisan cooperation is not explicitly needed (for the majority party), members recognize its benefits, perhaps because every majority anticipates being in the minority at some point in the future, because they believe that constituents value cooperative behavior, or because members believe that they get relational and informational benefits from being in bipartisan organizations. While we cannot be certain of the impetus and it may vary from member to member, the result is that some caucuses are nearly apolitical. As respondents tell us, some caucuses refuse to take explicit positions with regard to legislation to avoid introducing conflict that may threaten their bipartisan nature. Accordingly, we find fewer references emphasizing the value of caucuses for activities such as coalition building or the introduction of amendments. Attempts at political coordination have the potential to lead to contestation and conflict, which many caucuses find undesirable. These dynamics also explain why we find more examples of political information exchange in the EP than in Congress. In a political setting where cross-partisan compromise is necessary to pass legislation, intergroups can be valuable forums where representatives of different parties can find common ground on issues that they consider to be political priorities. In this sense, the exchange of political information is a key part of any attempt at coordination. Yet both the exchange of political information and coordination efforts are limited to a small subset of participants in the two LMO systems—that is, LMO leaders and a small number of particularly involved members (e.g., members of an EP intergroup's bureau), who are strongly tied to one another. In this context, we observe the “strength of strong ties” (Krackhardt 1992) rather than the “strength of weak ties” (Granovetter 1973). Despite the incidence of both signaling and coordination, the principal benefit for the majority of rank-and-file LMO members is relational and Page 184 → focused on the exchange of policy information. LMOs assume the role of educators on both current legislative proposals and issues that active participants in the LMO system might

like to see on the legislative agenda. Yet policy ideas are often infused not by inside advocates but by the interest groups with whom many LMOs maintain regular, institutionalized relationships. Indeed, outside advocates contribute in important ways to LMOs’ informational function, frequently serving as sources of expertise on which LMOs rely. An outside advocate may draft policy briefs, engage in research and analysis, participate in roundtable discussions, or know whom to contact should a legislator have a question that the outside advocate cannot answer. The information outsiders provide tends to be of high utility to LMO members because it is generally research based, reliable, and easily digestible. Outside advocates are well aware of the privileged relationships they maintain with LMOs and their leaders, and losing this privilege would cost them dearly. LMOs provide what all interest groups strive for: long-standing, trusted relationships with legislators who share the groups’ policy concerns. The institutionalized access LMOs provide, therefore, gives outside advocates strong incentives not to supply information that might be seen as redundant, unreliable, or misleading. Legislators, for their part, benefit from the privileged partnerships between LMOs and particular interest groups in that these groups allow them to tap into external networks of policy experts who are eager to provide policyrelevant information. However, those eager outsiders provide not only information but also other types of legislative subsidies in that they help set up LMOs, provide administrative and logistical assistance, recruit members, organize meetings, and sponsor events. This kind of subsidy is crucial to the creation and maintenance of LMOs, because it helps keep LMOs cheap. Members are free to enjoy the informational benefits of LMOs without having to invest their limited time and resources. In sum, LMOs provide easy access to high-utility information regarding legislators’ policy priorities. They encourage the inflow of high-utility information and facilitate the efficient diffusion of this information through LMO networks. Yet the costs associated with these benefits are minimal from the legislators’ point of view because no one is forced to participate or to expend resources on LMOs. In sum, LMOs—voluntary, low-cost organizations—help mitigate the inherent informational collective action dilemma in legislative politics.

Page 185 →

CHAPTER 7 Conclusion: How LMOs Do and Do Not Matter How LMOs Matter, Sometimes: LMO Impact on Legislative Outcomes One big question yet to be considered is to what extent the activities of legislative member organizations (LMOs) are consequential in that they affect the content of legislation. This is a difficult issue to address, because we expect the influence of LMOs to be diffuse and indirect. In fact, our theoretical argument about the strength of weak, cross-cutting ties, the exchange of policy-relevant information, and the collective gains reaped from both describes a kind of influence that should be difficult to measure. LMOs are most likely to affect the legislative process at the agenda-setting stage, or in the early stages of the bill formation process, when legislative proposals are first considered, deliberated, and negotiated. LMOs and Legislative Outcomes Identifying the direct impact of any particular player, group of actors, or other contextual influence on a final policy decision is exceedingly difficult because we cannot conclusively reject the counterfactual. For example, making the case that a particular amendment was introduced as a result of discussions in an intergroup or tracing a policy initiative back to ideas first raised in a congressional caucus will almost always be difficult and will often be impossible. The reality is that LMOs do not play a formal role in the legislative process, which means that any influence they wield will be indirect, through the activities of LMO members inside parties or committees. For example, amendments introduced by a European Parliament (EP) intergroup do not show up as such in the legislative records; as one staffer explains, “It is always an internal process. Of course you will never see it.”1 Page 186 → In this light, it is not surprising that a number of respondents in both the EP and Congress share our sense that LMOs do not frequently directly affect legislative outcomes. Indeed, there is a fair degree of skepticism among some respondents in both chambers when asked if and how LMOs influence policy proposals.2 One, for example, maintains, “I don't think you can really say one caucus really did or didn't do something that made a difference at the end of the day.”3 Given their lack of formal legislative power, most LMO activities focus on agenda setting or agenda maintenance. Intergroups in the EP, for example, seek to affect policy outcomes by influencing the agenda of either the EP4 or the European Commission,5 by trying to ensure that particular causes do not get lost in the myriad of competing policy concerns, or by helping members of the European Parliament (MEPs) make up their minds on specific policy issues.6 Intergroups and their members use a number of tools to achieve these objectives—issuing parliamentary questions,7 writing letters to the president of the Commission or the Council signed by leaders and /or members of the intergroup,8 introducing initiative reports,9 or promoting written declarations.10 Congressional caucuses also help set the legislative agenda by keeping topics at the forefront of legislators’ attention11 and by engaging in internal advocacy efforts to build support and keep legislation moving. In the opinion of one respondent who discussed the reauthorization of the National Earthquake Hazards Reduction Program in 2010, “the [hazards] caucus was the only way to keep it out there, to keep it alive.”12 Similarly, in reference to the reauthorization of a special diabetes program, the Congressional Diabetes Caucus “has been effective in getting a lot of good, strong bipartisan support for those bills.”13 One respondent credits the Korean caucus with helping the Korea Free Trade Agreement to pass since “most members of the caucus…tried to push hard” on this legislation, holding briefings “to try to gain interest in it, to try to inform congressional staffers and members of the benefits the free trade agreement might bring.”14 Our respondents maintain such efforts have contributed to the realization of numerous different policies; table 7.1 lists the items our respondents believe have

been included on the legislative agenda in part as a result of the activities and advocacy efforts of various caucuses. Page 187 → In the EP, we found a particularly instructive example of an intergroup setting the agenda in the case of the European Union Strategy for the Baltic Sea Region. This initiative can be traced back to the EP's resolution on the Baltic Sea Strategy for the Northern Dimension (INI/2006/2171), which in turn is based to a significant extent on a working paper issued by the Baltic Europe Intergroup, Europe's Strategy for the Baltic Sea Region (Beazley et al. 2005). The timeline for the course of events is as follows: November 2005: The Baltic Europe Intergroup publishes Europe's Strategy for the Baltic Sea Region. November 2006: The EP's resolution passes, urging “the Commission to come up with a proposal for an EU Baltic Sea Strategy” (European Parliament 2006; A6-0367/2006). December 2007: The European Council invites the Commission to present an EU strategy for the Baltic Sea region by no later than June 2009 (European Council Conclusions, December 14, 2007, point 59). June 2009: The Commission issues a communication to the European Parliament, the Council, the European Economic and Social Committee, and the Committee of the Regions concerning the European Union Strategy for the Baltic Sea Region (Commission of the European Communities 2009). Page 188 → October 2009: The European Council adopts the EU Strategy for the Baltic Sea Region and endorses the Council conclusions on the subject (European Council Conclusions October 29–30, 2009, points 35–36)

The idea for the creation of a Baltic strategy, which is now in the process of being implemented, thus emerged within the Baltic Europe Intergroup (Joenniemi 2009).15 Particularly notable are the similarities between the intergroup's working paper and the EP resolution urging the EU to adopt the strategy. Table 7.2 compares some of the wording. Moving beyond the agenda-setting stage and trying to observe the direct impact of LMOs on the content of legislation becomes even more challenging, yet a number of our respondents provide specific examples for the direct influence of LMOs and their members on particular legislative outcomes. In the Congress, for example, the Pro-Life Caucus insisted on language in the 2010 Affordable Care Act (i.e., “Obamacare”) that ensured that public funds could not be used for abortions (the so-called “Stupak Amendment”).16 Similarly, the efforts of the Congressional Traumatic Brain Injury Caucus affected the Traumatic Brain Injury Act, which was reauthorized in 2009. Even though the caucus chairs were not members of the committee of jurisdiction, they “were literally in committee staff communications and helped pass the bill.”17 Another example of a congressional caucus contributing to a legislative accomplishment comes from the 2004 passage of the Congo Basin Forest Partnership Act. The measure gained strong support in both House and Senate and was signed into law by President George W. Bush following active advocacy efforts by the International Conservation Caucus (Royce 2004). Finally, the Congressional Steel Caucus both advocated and provided language to support the domestic steel industry in the fiscal stimulus bill of 2009 (American Recovery and Reinvestment Act), promoting in particular the “Buy American” amendment. The caucus sent a letter to party leaders that was signed by 50 members of the caucus and held a well-timed, well-attended event in which steel industry CEOs highlighted the issue. The House passed a version of the bill with the language that the steel caucus had advocated.18 Similarly, in the EP, a number of respondents give specific examples for the impact of intergroups by describing instances where amendments on specific policy proposals that were drafted within the context of the intergroup

were incorporated into the final policy product.19 Several respondents Page 189 → affiliated with the Trade Union Coordination Intergroup cite the group's role in the negotiation of the Services Directive, one of the most highprofile and most contested pieces of EU legislation in recent years. As one MEP suggests, “With the trade unions, we totally converted [the Services Directive] from a very hostile piece of legislation from the perspective of the trade unionists [into] something that we could not only live with but to a great extent embrace.”20 According to another respondent, the “final amendments regarding the country-of-origin principle” were drafted by representatives of the Trade Union Intergroup and its outside affiliate, the European Trade Union Confederation.21 Even a respondent who was not happy with the outcome of the Services Directive acknowledges that the Trade Union Cooperation Intergroup had “partial influence in the discussion” of the legislation; in his mind, the discussion in the intergroup “certainly played a role.”22 However, given that intergroups do not play a formal part in legislative negotiations, “only few people know how important we have been for the success of that.”23 Page 190 → Page 191 → Respondents also describe a series of legislative proposals where the Animal Welfare Intergroup* achieved success. One MEP is 100 percent sure that, for instance, with the subsidies for long animal transports, that was the intergroup. I think that also cat and dog fur, sealskin, when we made REACH, which is the big chemical legislation, there are a lot of different amendments on animal experiments that went through. It is difficult to say whether or not it is only because of the group, but we did work within the group, and we did get a lot of them through. And I have been negotiator on [a specific legislative dossier].24 And you know, you can always argue that, well, even if you had not been in the group you would have done these things anyways. That might be true, but the fact of the matter is that I got a lot of my input from the group. I even got amendments that were [drafted] by the secretariat, and I got that into legislation.25 Page 192 → Another example of the Animal Welfare Intergroup's impact was provided not by a respondent affiliated with that group but rather by one from the Hunting Intergroup, which did not support the outcome: The negative [example] was last year. There was the proposal to ban the trade in seal products…. Our Nordic members have a small number of seals that they hunt. They do not club them, but they hunt, where they use every part…. So it is a different hunt, but to explain this to 750 MEPs is difficult. So for them it is vital, and so we organized this meeting, and we thought that is a very good important statement…. And then a month later, the Animal Welfare Intergroup, they are in some cases more efficient because they can play with pictures…. So they brought pictures [of bleeding seals], and in their own intergroup everyone was shocked…. They were very effective, and through their intergroup they could influence the committee, and they did…. We lost the battle.26 It is thus possible, if difficult, to link LMO activities to specific policy outcomes, meaning that LMOs are not just able to help shape legislators’ minds but also quite capable of directly influencing the content of legislation in some circumstances. Table 7.3 lists instances where our respondents indicated that LMOs’ activities affected the content of legislation, either through amendments or specific language introduced into a given bill. Such influence is difficult to trace, because amendments are usually introduced under the name(s) of one or more LMO members rather than the name of the group itself.27 In the EP, this process can happen informally (e.g., when an MEP introduces an amendment that follows explicitly from a discussion in the intergroup)28 or in a more formalized process, where amendments are sent “to the relevant people, and I say this report is coming up, the intergroup is suggesting it to you. And then [there is] usually at least one person who takes it up.”29 According to the cochair of another intergroup, “Normally…we would try to prevail upon the rapporteur to table those amendments, or our shadows from the respective groups. Failing all of that, you will see my name appear on

amendments, or [my cochair's]—tabled on behalf of the intergroup, basically.”30 Following suggestions from some of our respondents, we traced a series of amendments back to different intergroups in a small number of cases, which are listed in table 7.4. Careful process tracing involving qualitative data may thus help uncover instances of LMO influence on legislative outcomes. An alternative Page 193 → approach is to investigate the impact of LMOs quantitatively. We do so using data on cosponsorship and legislative productivity from the U.S. Congress, given the lack of equivalent data for the EP. We first rely on data on the cosponsorship activities of members of the U.S. House of Representatives to establish if legislators who share connections via the caucus system are more likely to cosponsor one another's legislation.31 If we conceive of cosponsorship as an indicator of either shared priorities (Swers 2005; Fowler 2006a; Cho and Fowler 2010) and/or shared preferences (Koger 2003), finding a positive association between being socially tied in the LMO network and a propensity to cosponsor legislation would point to the political consequentiality of LMOs. A similar conclusion would be warranted if we also found that membership in LMOs positively affects lawmakers’ individual-level productivity. Our second quantitative analysis focuses on this investigation. Page 194 → LMOs and Cosponsorship In our analysis of cosponsorship, we rely on the most state-of-the-art network analysis statistical methods available—exponential random graph models (ERGM), a sophisticated technique that appropriately models the inherent dependence in network data (Handcock et al. 2003, 2011; Hunter et al. 2008; see also Butts et al. 2011). We seek to explain the network created by legislators’ decisions to cosponsor bills as a function of a number of endogenous and exogenous covariates. Endogenous covariates capture the impact of the network structure itself. For example, as we discuss in chapter 2, two individuals who are tied to a third are more likely to also be tied to one another (the principle of triadic closure), a phenomenon for which we must account. In this ERGM, we control for three endogenous structural terms: edges (which act like a constant in a traditional regression), 2-stars (pairs of edges {(i→j), (j→k)}), and triangles (any set of three edges {(i, j), (j, k), (k, i)}). We also control for a series of exogenous covariates, or variables that that we assume affect legislators’ propensity to cosponsor beyond the network structure itself. These variables capture, for example, the effects of shared LMO membership (our key variable of interest), service on the same committee, membership in the same party, coming from the same state, being of the same gender, or being African American. In addition, we control for the tendency of legislators to be of similar seniority (or the covariance of the number of terms each has served in Congress) and similar electoral security (or the covariance of the percentage by which each won the previous election). Appendix B provides technical details on our models. Table 7.5 presents the results from our estimations.32 Most notable from our perspective is that we observe positive and significant coefficients on the caucus connections between legislators in both the 109th and 110th Congresses; in other words, co-membership in legislative caucuses predicts legislators’ tendencies to cosponsor one another's legislation, even when accounting for the impact of network structure and variables such as joint membership in parties and on committees. Control variables and structural terms are in the expected directions to the extent that we have expectations about their performance. Regarding the substantive interpretation of our results, the ERGM coefficients closely resemble logit coefficients, except that the probability of observing any given dyad is conditional on the dyad-wise outcomes of all other dyads (Cranmer and Desmarais 2011, 73). Since we are not interested in the substantive interpretation Page 195 → of these values per se, we do not compute probabilities here. It is sufficient for this demonstration to recognize that the key independent variable of interest (caucus co-membership) is positive and statistically significant in a model that fully accounts for other individual, dyadic, and network-level sources of covariation. Page 196 → Page 197 → Page 198 → LMOs and Legislative Productivity

Our second quantitative test of the impact of LMOs examines the relationship between legislators’ involvement in LMOs and their legislative productivity. Here, we rely on data on the number of bills individual legislators sponsored during each of the three sessions of Congress we study (Adler and Wilkerson 2005–8; Library of Congress 2011). We also examine the number of legislators’ sponsored bills that pass the House chamber and those that become law. We seek to explain these counts using two indicators for legislators’ involvement in LMOs: their total number of LMO ties (or their degree) and a measure of their structural position in the LMO network, eigenvector centrality (Bonacich 1972).33 The advantage of using Page 199 → eigenvector centrality, as opposed to a simpler centrality measure, is that it is not merely based on the number of ego's ties but also takes account of the structure of the entire network in that it examines the total number of ego's connections as well as the connectedness of ego's alters. That is, actors’ centrality is a function of their own connections as well as the connections of adjacent others. We estimate a series of pooled cross-sectional negative binomial regressions in which standard errors are clustered on individual legislators to account for autocorrelation.34 We control for time by including dummies for each Congress as well as a series of factors that may also explain legislative productivity at the individual level: party, gender, the percentage by which legislators won their last elections, seniority, party leadership, committee leadership, and being a racial minority. Of course, as a proportion of all bills, the number of sponsored bills that pass the House or become law is quite small. According to the Résumé of Congressional Activity, the House passed just 8 percent of the 4,122 bills that were introduced in the 109th Congress, and only 3.6 percent of the total became law. In the 110th Congress, the House passed 9 percent of the 3,741 bills introduced, and 3.8 percent became law. And in the 111th Congress, the House passed only 6 percent of the 4,059 bills that were introduced, and 3 percent became law.35 Given the very small number of bills that make it through the process and the indirect impact that we assume caucuses have on the legislative process, we see this is a very conservative test of the political consequentiality of LMOs. Table 7.6 presents the results, showing three models with three different dependent variables: total bills sponsored (an overall measure of productivity), sponsored bills that pass the House chamber, and sponsored bills that became law. The coefficients for legislators’ degree are positive and statistically significant, if substantively small, in two of our three models: While lawmakers’ total number of LMO ties does not affect the number of bills they sponsor (Model 1), the number of ties does correlate positively with both legislators’ total number of sponsored bills that pass the House and sponsored bills that become law (Models 2 and 3). In other words, greater involvement in LMOs slightly increases individual legislative success. In contrast, a legislator's structural position in the LMO network, measured using eigenvector centrality, does not affect legislative productivity or success, as the coefficients (which are positive for the total bills sponsored and negative for sponsored bills that pass the House chamber and become law) Page 200 → do not achieve statistical significance.36 The control variables in these models generally perform as expected. In the end, we take the results of both of our analyses—predicting cosponsorship and legislative success—with more than a few grains of salt even though they point to the political consequentiality of LMOs. We ultimately agree with the evaluation of one respondent in Congress: Caucuses “very rarely accomplish goals as a group. It's very rare that…you'll have a meeting of a caucus where they will actually come up with a legislative program and pass it. That is just not how it works.”37 Page 201 →

How LMOs Matter, Most of the Time: Relationships and Information Despite our findings about the direct influence of LMOs on legislative outcomes, and despite the positive correlations between involvement in caucuses and both cosponsorship and legislative productivity, we consider the primary influence of LMOs to be indirect and diffuse. LMOs and their activities affect discourse, attention, and priorities. These are nontrivial contributions in the legislative arena, however, as one staffer in Congress highlights: “At the end of the day, I don't think any…final decisions are made because of the caucuses, but they

can definitely [have] a lot of…influence, a lot of influence leading up to those things.”38 A member of the European Parliament agrees, describing intergroups as “crucially important…in terms of our work…. They really have a tremendous influence on the way that Parliament works.”39 This influence results from LMOs’ ability to help legislators, both as individuals and as a group, become better educated and more informed about the priorities of key actors inside and outside of the legislatures, about the content of legislation, and about the expected consequences of a policy after it is implemented. Our key findings are quite consistent across both institutions. Intergroups in the European Parliament and caucuses in the U.S. Congress allow legislators to establish relationships with colleagues who share an interest in a given policy area. These relationships tend to be weak, bridging ties that facilitate the efficient flow of policyrelevant information between social groups that would otherwise be more isolated from each other. Moreover, LMOs achieve this beneficial outcome while putting fairly small demands on participants in terms of time and resources. LMOs, as voluntary institutions, thus help mitigate the informational collective action dilemma inherent in legislative politics, where a high demand for information is met with an insufficient supply. These findings imply that there is value even in purely social LMOs that relate to seemingly trivial matters, because relationships can be forged in one setting and become meaningful, from a policy-making perspective, in another. LMOs thus connect legislators with each other; however, they also help build links between legislators and outside advocates and stakeholders who share the lawmakers’ interests in an issue or cause. These outside actors provide legislators with both political information about the positions of Page 202 → key stakeholders and substantive information about the policy area in question. This substantive information ranges from quite general to highly technical and specific to a current policy proposal. Such input is, of course, not unique to LMOs, and outside advocates use LMOs as one of several access points to lawmakers. However, LMOs help increase the quality of the information that enters the legislature by providing an incentive structure for outside advocates to supply legislators with high-utility information—that is, information that is reliable, research-based, and presented in a format that allows policymakers to use it easily in their speeches, press releases, and policy negotiations. This incentive structure is based on the high levels of internal social embeddedness in single-LMO networks, the opportunity to triangulate and verify information supplied by different sources, and the privileged access LMOs grant to outsiders, which legislators can credibly threaten to revoke if outside advocates provide redundant or misleading information. Aside from offering informational value, interest groups and lobbyists eager to establish trusted, long-term relationships with legislators are also willing to cover many of the costs associated with creating and maintaining LMOs, thereby helping to make participation in LMOs inexpensive from the point of view of individual legislators. This low-cost nature is a critical characteristic of successful solutions to institutional collective action dilemmas (Feiock and Scholz 2010). As a whole, our arguments about the benefits of LMOs suggest that these groups help establish internal legislative networks that enhance the efficient flow of policy-relevant information between lawmakers. As social networks that also include outside actors, they provide an incentive structure that increases the quality and utility of the information that enters the legislature, thus allowing legislators to be better informed when making laws. The evidence from our case studies confirms that LMOs play these roles across legislatures. That being said, we observe some notable variance, allowing us to fine-tune some of our conclusions about the existence and functioning of LMOs. Analyzing information from 45 advanced industrial democracies reveals that LMOs exist in all 6 countries with majoritarian electoral systems. This finding suggests that LMOs provide a form of constituency service, giving legislators the opportunity to be involved with issues and causes that their constituents prioritize. LMOs thus provide electoral benefits when the links between voters and lawmakers are strong. This finding confirms that, at least in some contexts, LMOs play a signaling role, Page 203 → where legislators can demonstrate their commitment to an issue or cause that may be important to constituents. One interpretation of this finding is that LMOs simply provide opportunities for legislators to engage in cheap talk—that is, joining LMOs offers an inexpensive way to signal support for a policy even if the lawmaker has no

intention of actively pursuing it. This interpretation does not, however, account for the prevalence and (in some cases) prominence of LMOs in cases where legislators have much less of an incentive to engage in signaling activities (most notably if they are elected in list-proportional representation electoral systems). Moreover, in countries using proportional representation, variables accounting for the presence of electoral institutions that create incentives for politicians to cultivate a “personal vote” do not predict either the presence of LMOs (chapter 3) or the proclivity of members of the European Parliament to join intergroups (chapter 4). These findings shed further doubt on the idea that LMOs serve primarily as signaling devices; if such were the case, we should observe a greater tendency to create and join LMOs in legislatures where members are selected using personal voting. Hence, if we ask, “Why do legislators invest scarce time and resources in forming and maintaining voluntary groups that provide few obvious benefits?,” signaling is an unsatisfactory answer in many, if not most, legislatures we examine in chapter 3. We also find some evidence for the coordination function of LMOs—although, as in the case of signaling, coordination is less prominent in some cases than others. The principal determinant appears to be the relative need for cross-partisan cooperation in a lawmaking institution. In the EP, for example, where multiparty policy coalitions are required to pass legislation, intergroups can serve as arenas for political coordination, such as the search for consensus and compromise or the joint introduction of amendments. Incentives for such activities are less pronounced in Congress, where the majority party dominates the policy-making process. While our data show somewhat limited evidence for the signaling and coordination functions of LMOs, which are also more or less pronounced in some cases than others, our case studies provide consistent support for the informational benefits of LMOs. Joint membership in LMOs allows legislators to build relationships with each other that cut across party and committee lines and thus facilitate the exchange of high-utility information between interested actors inside and outside the legislature. LMOs allow for the diffusion of both political and policy information because their Page 204 → voluntary nature makes participation inexpensive; offers opportunities for legislators to build endogenous, heterogeneous social networks based on policy priorities; and makes them arenas of legislative activity that are less formalized than either parties or committees. This informality allows LMOs to take up issues in a more expeditious fashion since they are not limited by the closed issue space characteristic of committees with fixed jurisdictions. LMOs also benefit because they (and their activities) are genuinely issuefocused. In an institutional space that is structured primarily by partisanship and policy competition, LMOs provide rare opportunities for legislators to learn about and exchange ideas on issues they deem important without party politics getting in the way. The issue focus and informality of LMOs thus allows legislators more freedom to speak their minds and creates a flexibility in discussion where they may feel less constrained by established party positions. The substantive discourse that takes place in LMOs may even offer opportunities for members of a legislative minority, whose support is not necessary for legislation to pass, to influence the content of legislation. Overall, the relative informality of LMOs facilitates the creation of relationships across parties and committees and thus the efficient flow of information across the social boundaries imposed by the formal legislative institutions. Such relationships may in fact benefit these formal institutions as their members who participate in the LMO system become better informed. In this sense, parties and committees and LMOs are not in competition with each other but play complementary roles in legislative politics. Our findings about the informational and coordination functions of LMOs match the results of our cross-national analysis in chapter 3, where we demonstrate that, among countries with nonmajoritarian electoral institutions, those with large legislatures and a high effective number of parties are more likely to have LMOs. Specifically, among countries with proportional representation or mixed electoral systems, the four that feature large legislatures with many parties have LMOs. In contrast, we find LMOs in only about 50 percent of countries with either large legislatures and few parties or small legislatures and many parties. Finally, only one of eight countries with small legislatures and few parties has LMOs. These findings make sense in light of the need for political coordination in LMOs, since they will be more valuable as instruments facilitating collective action in large legislatures with multiple parties. However, our results are also in line with the informational argument about the value of LMOs, Page 205 → because the need to

establish legislative institutions that allow lawmakers who share a common interest in an issue to exchange information is more pronounced in large legislatures where legislators are less likely to meet each other. This need might be less pronounced in small legislatures, which would imply that to the extent that LMOs exist in small chambers, their informational function may be more important than their relational role. In legislatures with many parties, however, cross-party and cross-committee ties are of special value, since they facilitate the flow of political information about the positions of other parties and may help build policy compromise and consensus. Indeed, this hypothesis is confirmed by our case studies, which show a greater incidence of political information exchange in the EP, with its seven party groups, than in the House of Representatives. The existence of LMOs does not necessarily mean that they play a highly relevant role in every legislature, however. LMOs exist in the German Bundestag, for example—as we would expect given that it is a large legislature with a higher than average effective number of parties—but they do not play a substantively significant role in the legislative process. Much variation exists among cases and still requires explanation. This book provides a starting point for such research. Our study—the first comparative analysis of LMOs—looks beyond parties and committees as legislative institutions that can solve lawmakers’ collective action dilemmas and more generally as arenas of political activity in lawmaking institutions. These groups engage in a variety of activities, as our survey in chapter 3 indicates and our case studies confirm. LMOs organize and hold regular meetings, seminars, and colloquiums and issue press releases, reports, and opinions. They provide opportunities for legislators to gain information and become educated on issues that are important to them and offer a venue in which to socialize, network, and build relationships. Respondents to our survey also confirm that LMOs maintain cooperative ties with outside experts, scholars, lobbyists, and interest groups. Finally, LMOs engage in efforts to coordinate legislative activities and seek to influence the legislative agenda. LMOs, in other words, are notable features of legislative politics in a large number of lawmaking bodies. Although LMOs are easily overlooked, our study highlights their prevalence in legislative politics in a large number of established democracies. An exceedingly important feature of LMOs is their cross-partisan nature. Whether partially imposed (where rules governing LMO activities prescribe multiparty membership) or entirely endogenous (where shared Page 206 → policy priorities and voluntary participation alone drive cross-partisan membership), inclusiveness and diversity are among the major benefits of LMOs identified by our respondents. However, the nature of cross-partisanship in LMOs differs between the EP and Congress. Because legislation requires cross-partisan policy coalitions to pass the EP, intergroups may serve as arenas for contestation, negotiation, and coordination. Cross-partisanship in intergroups is thus explicitly political. In contrast, the majority party in the Congress does not require support from across the aisle to pass legislation (unless the partisan majorities in the House of Representatives and the Senate differ); for this reason, many caucuses seek to protect their bipartisan nature by consciously avoiding divisive issues. The almost apolitical nature of bipartisanship may seem like a genuine weakness of the caucus system, but we disagree. Partisanship and polarization have reached unprecedented levels, and the number of caucuses in Congress has grown in near lockstep with increased party polarization in Congress (figure 7.1) (r = 0.93).40 Moreover, most caucuses are truly bipartisan, and the number and proportion of all caucuses that include members of both parties have increased over time (figure 7.2). Across the period of study, only about 25 percent of caucuses are single-party groups, and the vast majority of these have fewer than 10 members each. We also find, however, that the mean weighted difference in party proportions ranges from 4.4 to 11.7 (figure 7.3). The mean weighted difference in party proportions is calculated by determining the absolute difference in the proportion of a caucus's membership that is Republican and Democratic, then weighing this difference by the total size of the caucus and calculating an average for each Congress. The closer this number is to 0, the more equally distributed the parties are in the caucuses. The median of this statistic ranges from 2 to 4 throughout the period under study. Recent Congresses have shown an uptick in the mean weighted difference between the parties, suggesting the presence of party-unbalanced caucuses. As the number and average size of caucuses have grown, therefore, so have party polarization, the number of bipartisan caucuses, and the number of single-party caucuses.

We therefore observe the contradictory forces of party polarization and bipartisanship in the growth of the caucus system, a ripe area of for future research. Do legislators strategically use the caucus system in Congress to help compensate for the barriers placed on them by increased partisan polarization, like an antidote to the decline of comity (Uslaner 1993)? Such a strategy would be consistent with our theory and evidence that Page 207 → shows that the connections provided through LMOs help lawmakers bridge institutional cleavages; however, our initial dip into these data also show evidence of polarization in the caucus-joining habits of legislators. Our descriptive accounts of these patterns invite further research to tease out the causal forces behind these trends.

Our investigation of LMOs also helps shed light on questions of information Page 208 → flow, the use of expertise in the legislative process, and how legislators overcome the informational constraints they face in their everyday work. Legislators constantly require and demand both political and policy information yet face important limitations when it comes to gathering this information. These constraints may result from a dearth of input or, more frequently, from an oversaturation, where there is too much information to process. LMOs not only encourage information exchange across party and committee lines but also help legislators and their staff identify reliable sources of information for particular policy questions. A legislative office knows that members of an LMO leadership network can be approached with specific inquiries relevant to the lawmaking process. LMOs thus help identify sources of internal expertise. Moreover, LMOs encourage the inflow of policy expertise through their institutionalized relationships with particular outside groups, where LMO leaders (or any other member) can seek input when internal expertise does not suffice. It is quite critical that this outside expertise is not supplied intermittently by random outside actors but comes reliably from outsiders who have strong incentives to provide quality information that is of high utility to legislative actors. Hence, our study also informs the existing research on access and the influence of outside advocates in the legislative decision-making process. Specifically, this book shows how LMOs provide privileged points of access Page 209 → to interests and actors outside the legislature and an institutionalized means for outside advocates to build long-standing, trusted relationships with legislators. Indeed, evidence from our case studies demonstrates that interest groups encourage and are very active in initiating the creation of LMOs. In both of our cases, outside actors have approached legislators, offering to help establish and maintain particular LMOs. In turn, LMOs create an incentive structure for outside advocates to provide legislative subsidies (Hall and Deardorff 2006) in the form of high-utility information, both political (about the preferences and positions of key constituents and stakeholders) and substantive (about the content and consequences of legislative proposals). Legislators can readily use this information in public statements, writing legislation, and other activities. We make a general contribution to the study of political institutions by highlighting the role of voluntary institutions with no formal role in the legislative process in solving a particular institutional collective action dilemma—the great demand and insufficient supply of policy-relevant information. A key insight in this regard is that some of the inherent characteristics of parties and committees, the principal formal legislative institutions, may stand in the way of efficient information exchange. Indeed, while these institutions are set up to address particular collective action problems, their structures have unintended consequences that create new problems or impose additional constraints. LMOs help provide opportunities for overcoming and crossing barriers to information exchange created by existing institutions. They help mitigate the consequences of the informational constraints imposed by the formal legislative institutions by structuring the activities of legislative actors in such a way that lawmakers with similar policy priorities come into institutionalized social contact with each other, thus encouraging the creation of relationships between them. These social relationships and the networks they comprise provide low-cost access to high-utility information about issues that are important enough to legislators that they join a voluntary association with no formal role in the legislative process. Less formalized, voluntary organizations thus interact with formal institutions to make legislators more informed as individuals and consequently the legislature more informed as a collective body.

This argument about the value of LMOs is reminiscent of research that examines the role informal institutions may play in mitigating the weaknesses associated with formal institutional structures (e.g., Helmke and Page 210 → Levitsky 2004, 2006; L. L. Tsai 2007).41 Along similar lines, Feiock and Scholz (2010) highlight the idea that flexible, low-cost, self-organizing institutions or social networks can play an important role in overcoming institutional collective action dilemmas. LMOs, however, are neither informal institutions nor mere social networks, even though they provide similar benefits to their participants and the broader structural environment in which they are embedded. In most legislatures, LMOs display distinct institutional structures that set them apart from self-organizing social networks and informal social institutions. LMOs have memberships and membership lists; many are recognized and registered by their legislatures; they may be subject to official rules and procedures governing their existence and operations; some charge membership fees. In other words, not only is there is a general context or infrastructure in which LMOs are situated, but LMOs themselves tend to display a distinct level of institutionalization. Most of the value we ascribe to LMOs is fundamentally relational. We thus contribute to the literature on social networks among elite actors by theorizing how LMO networks, dominated by the presence of weak, bridging ties, help legislators efficiently do their jobs. We also highlight the particular value of cross-partisan ties for legislators, who benefit from relationships not only with political allies but also political opponents. For this reason, ties established in LMOs as voluntary organizations are not strong, homophilous ties, as one would expect in other political contexts, but weak ties that connect actors who are unlike each other. It is notable, however, that we do not infer social network properties from variables such as cosponsorship but analyze a relational structure that legislators themselves describe as social networks. Our discussion of the relationship between formal legislative institutions and LMO networks also touches on more general questions relating to the interplay and relationship between knowledge networks and the broader formal and informal institutional context in which they are embedded. As Phelps, Heidl, and Wadhwa (2012, 37) emphasize, the research on knowledge networks is undercontextualized in this respect. This volume, however, illustrates how the information networks we investigate are affected by the broader legislative structure in which they are placed. First, the information networks we identify are voluntary and largely endogenous, but they are enabled by and closely tied to the semiformal institutional structure that is the LMO system in a given legislature; there is an infrastructure in place that helps shape the structure of networks of individual LMOs and Page 211 → the larger whole they jointly comprise. Second, LMO-based information networks are a counterpoint to parties and committees as formal legislative institutions that impose significant informational constraints on legislators. As such, the design of LMOs and the LMO system is at least in part a response to these shortcomings. It follows that our research recognizes and takes seriously the duality of social group structures (Brass et al. 2004, 808). Individuals are members not only of more than one social group but also of various social networks that may be nested within each other. For example, LMO leaders are simultaneously (and by virtue of this position) members of the LMO leadership network, the single-LMO network, the LMO network as whole, and the legislative network composed of all social and institutional ties among lawmakers. LMO leaders consequently maintain both strong ties (which foster reciprocity and trust) and weak ties (which bridge structural holes in the legislative network and are inexpensive to maintain) with a diverse set of alters that affect not only the leaders’ levels of policy-relevant knowledge but also the overall levels of expertise and informedness in the legislature as a whole. While the evidence presented in this book and the extensive data on LMOs both across a broad sample of democracies and in the context of Congress and EP offer a great deal of insight into LMOs as notable features of legislative politics, our research opens up more avenues for research than it closes. We have only touched on questions relating to the origins of LMOs and LMO systems, their institutional development, and trends over time, all of which deserve further attention. Also, despite offering the first comparative look at LMOs and a great deal of new information and data, we know little about LMOs outside of our case studies. Consideration of additional cases is thus highly warranted, as is a closer examination of the relationship between LMOs and formal legislative institutions such as parties and committees in an effort to enhance our understanding of both legislative institutions and political institutions more generally. We also hope to stimulate further research on the role of

LMOs as points of access and arenas for advocacy and lobbying. LMOs are only one of many ways in which interest groups and lobbyists might seek to influence legislative processes and outcomes, and the topic calls for additional research not just in the context of legislative studies but also relating to the broader literature on lobbying and interest representation. Finally, we remain intrigued by questions relating to the impact of LMOs on both legislative processes and outcomes. Our findings about the different natures of cross-partisanship in the Page 212 → EP and in Congress, for example, leaves us wanting to learn more about how LMOs might affect legislative actors, their interaction, party polarization, and bargaining dynamics. We hope this volume will serve as a starting point for further research on these and many other questions. LMOs create networks of legislators that extend beyond the social boundaries imposed by party and committee lines. The weak ties built through LMOs span structural holes in the legislative network, facilitate the flow of information, and create opportunities for the development of innovative policy solutions. LMOs thus provide legislators with opportunities and venues to learn about policy areas in which they are interested, both in general terms and with regard to specific legislative proposals. Lawmakers use the knowledge and ideas generated in the less formalized settings of LMOs when they return to their parties and committees, where policy is made, and thus influence policy outcomes in a more substantive and meaningful way.

Page 213 →

APPENDIX A Our survey was administered between February 10, 2010, and February 5, 2011.

Survey Questions Round 1: Initial questions. 1. Do LMOs exist in the national legislature of [country]? 2. If yes, what are they called? 3. Can you provide advice, citations, or websites where we might be able to learn more about how LMOs operate in [country]? (e.g., information about LMO registration, membership, etc.) Round 2: If a respondent indicated that LMOs exist and that s/he can provide additional information, we sent a follow-up inquiry that asked a series of contextual questions: 1. Does the [name of legislature] require that LMOs be formally registered with the [name of legislature]? Or is it possible for informal groups to form and exist without formal registration? 2. Who can join the LMOs? Is it only MPs or can outside groups (NGOs, firms, organized interests, etc.) join the LMOs? 3. Are there rules about the partisan makeup of the LMOs? In other words, can they be made up of MPs from only one party, or are they required to be cross-party? If there are requirements about the partisan makeup, can you describe them? 4. Does the [name of legislature] provide funding for the LMOs? If not, where do they get funding (if they have any at all)?Page 214 → 5. Can you provide a brief description of the primary functions of the LMOs? What do they do in general? 6. Can you provide any anecdotes about how the LMOs have gotten involved with parliamentary affairs?

Page 215 →

APPENDIX B Exponential Random Graph Models The ERGM technology is designed to account for the various types of autocorrelation that are present in network models. In this analysis, we explain the network of relationships created by legislators’ decisions to cosponsor bills in part as a function of their relationships in the caucus network. We treat the N × N matrix of legislators in the 109th and 110th Congresses as dichotomous directed adjacency matrices in which each cell (i.j) represents cosponsorships by legislator i on bills sponsored by legislator j. We also generate a 1-mode projection of the 2-mode incidence of legislators’ membership in caucuses in the 109th and 110th Congresses. This creates two N × N adjacency matrices, which we dichotomize so that the matrix indicates pairs of legislators who share some common caucus connections. ERGMs are a statistical approach that is specifically designed for network data and models the inherent autocorrelation and dependencies in the data (see Cranmer and Desmarais 2011). This approach accounts for interdependency in the data by defining a probability distribution for the set of all possible networks for a set of nodes. In the ERGM environment, one can have the highest level of confidence that network dependence has been not only accounted for but explicitly modeled. As Cranmer and Desmarais put it, Researchers can proceed with ERGM analysis based on hypotheses similar to those that would produce regression specifications (i.e., covariate x is expected affect the outcome y), and as much network structure (dependence) as they see fit. Moreover, the ERGM is poised for widespread application in political science and is implemented in a number of software packages. The ERGM is, as we will show, widely applicable to network analysis in political science and is remarkably flexible in its ability to model relational interdependence. (2011, 67) Page 216 → We control for a number of endogenous and exogenous covariates that we assume affect network structure. For the endogenous structure, we control for the overall number of edges (or ties) in the network. This term acts like a constant in a typical regression analysis. We also control for the number of “mixed two-stars,” or “two-paths.” This term adds one statistic to the model equal to the number of pairs that share distinct edges. This is a directed path of length 2. We also control for the number of triangles (any set of three connected edges) in the model. The ERGMs are estimated in the R package ERGM (Handcock et al., 2003, 2011; see also Butts et al. 2011). We dichotomize the networks at the 0 threshold level for the dependent variables, meaning that any pair of legislators who share any tie receives a 1, otherwise, they receive a 0. The dichotomization is necessary for the ERGM analysis. As discussed in the text, we also include a series of exogenous covariates that we assume affect legislators’ propensity to be in the same LMOs: caucus memberships (our key variable of interest), party, shared committee, shared state, gender, shared leadership, shared African American ethnicity, seniority, and the percentage by which members won their last election.

Page 217 →

NOTES Chapter 1 1. Throughout the text, we use the term legislator to refer to members of any legislative assembly. It is an inclusive term akin to lawmaker. 2. Throughout this book, when we are describing relationships between “legislators” or “lawmakers,” we include relationships between members of the legislators’ personal staffs. We thus conceive of legislative staff in the U.S. Congress and parliamentary assistants in the EP as extensions of the legislators themselves. In this sense, legislator may refer to all members of a legislative office, or what Whiteman (1995) refers to as “congressional enterprises” in the U.S. context. 3. Legislators may face an informational deficit either because they do not have enough information to engage successfully in their activities or because they face an informational overload (if they are barraged with information from a wide variety of sources and lack the time and resources efficiently to process the information they receive [Ringe 2010]). 4. Voluntary participation does not necessarily mean that ties will be weak. In fact, membership in voluntary organizations such as churches or social clubs tends to promote homophilous ties (Feld 1982; McPherson 1983; McPherson and Smith-Lovin 1986, 1987; Marsden 1990; McPherson and Ranger-Moore 1991; McPherson, Popielarz, and Drobnic 1992; Popielarz and McPherson 1995). Such connections between similar people tend to be strong ties (see, e.g., Zenger and Lawrence 1989; Marsden 1990; Ibarra 1992, 1995; Krackhardt 1992; Brass 1995; Perry-Smith and Shalley 2003). In legislative bodies, however, members face strong strategic incentives to regularly interact with dissimilar others rather than limit their interactions to colleagues with whom they share many similarities (a point to which we return in chapter 2). These nonhomophilous ties are more likely to be weak. Moreover, where voluntary institutions exist parallel to formal nonvoluntary ones, voluntariness implies a greater likelihood of sporadic involvement for most (but not all) LMO participants and thus an overall greater incidence of weak ties. 5. Most LMOs are cross-partisan organizations; in fact, in some legislatures (such as Westminster and the EP), LMOs are required to include members from multiple parties. In the EP, for example, formally recognized intergroups must have the support of at least three party groups. 6. Personal interview with congressional staffer, Washington, D.C., May 19, 2009. In a sense, we have found through our interviews that the relative success of Page 218 → an LMO relies on what some literatures call “policy entrepreneurs” (whether collective or individuals). 7. This literature also includes a typology of caucuses in the U.S. Congress (see also Dilger 2009), including intraparty, personal interest, industry, regional, state/district, national constituency, and diplomacy. We have found the categorization of LMOs to be sufficiently subjective and ad hoc that it not useful beyond providing a description of the population. 8. See Lazer 2011 and Ward, Stovel, and Sacks 2011 for recent overviews of the study of social networks in political science. Heaney and McClurg (2009) review the literature on social networks in American politics, Siegel (2011) discusses social networks research in comparative politics, and Hafner-Burton, Kahler, and Montgomery (2009) provide an overview of network analysis for international relations.

Chapter 2 1. There are, however, a fair number of descriptive accounts (see Ainsworth and Akins 1997, 408). 2. See “Lobbies in the Israeli Knesset,” at http://www.knesset.gov.il/lexicon/eng/lobby_eng.htm. 3. See “All-Party Groups in the United Kingdom Parliament,” at http://www.parliament.uk/about/mps-andlords/members/apg/. 4. These aspects of political dynamics of British politics are less dominant with the current Conservative–Liberal Democrat coalition government in office, but they remain key aspects of the British political system and have been the norm since 1945. Moreover, large numbers of APGs existed before the coalition government took office in 2010.

5. We say that LMOs “may” help fulfill this informational role in all legislatures not because there are LMOs that play no informational role but because some legislatures have no LMOs. All legislators need information, and in chapter 3 we explore the institutional circumstances that make LMOs more likely to exist. 6. Among the few projects examining social networks in legislatures outside the United States are Bräuniger and Brunner (2009), who examine cosponsorship networks in the German Bundestag, and Ringe, Victor, and Gross (2013), who examine staff networks in the European Parliament. 7. Legislators also often seek procedural information—that is, information about the rules and procedures that can be used to govern the lawmaking process (see Oleszek 2007). While procedural information may be less substantive than policy or political information, lawmakers can nonetheless use their knowledge of parliamentary procedure to make strategic maneuvers to champion or defeat legislation. They may also seek outside counsel on such strategies. 8. We highlight three main instances in which legislative leaders ought to try to control, regulate, interfere with, or even undermine LMOs. First, they might want to impede LMOs’ activities that undermined the formal operation of the legislature—for example, by diverting time and resources away from parties and committees. In Page 219 → other words, if LMOs became too expensive from the point of view of individual legislators, the system would come under threat. Second, legislative leaders should try to rein in LMOs that threaten party cohesion. Third, parties may be concerned about reputational costs associated with LMOs—for example, if LMOs received too much publicity or if their close relationships with particular outside organizations became an issue of public contestation. If one or more of these issues arose, parties might decide to intervene in a legislature's LMO system. This intervention can take different forms, ranging from trying to break up the LMO system (as the House Republican leadership under Newt Gingrich attempted in 1995) to instituting rules regulating LMOs’ meeting times, use of parliamentary facilities, and relationships with outside organizations (as happened in the EP starting in the mid-1990s). Of course, parties and committees also benefit from their members’ participation in LMOs, which allows them to garner information and expertise on specific policy issues; to receive electoral benefits if LMO membership signals that lawmakers support causes that voters care about; and to decrease rank-and-file discontent among members who do not hold the committee assignments of their choice, because LMO membership allows individual lawmakers to pursue pet issues. 9. We are grateful to Yoi Herrera for this example. 10. See chapter 5 for empirical support for these propositions. 11. On information flow in legislatures, see Judge 1981, 1990; Kingdon 1984, 1989; R. D. Arnold 1990; Krehbiel 1991; Mooney 1991; Austen-Smith 1992; Carpenter, Esterling, and Lazer 2004. 12. Extensive research links weak and/or bridging ties to the creation and flow of information, ideas, and knowledge (e.g., Granovetter 1983; Rogers 1995; Constant, Sproull, and Kiesler 1996; Burt 1997, 2004; M. T. Hansen 1999; Fleming, Mingo, and Chen 2007; McFadyen, Semadeni, and Cannella 2009), as well as creativity, innovation, learning, and performance (e.g., Nahapiet and Ghoshal 1998; Baum, Calabrese, and Silverman 2000; Perry-Smith and Shalley 2003; Cross and Cummings 2004; Rodan and Galunic 2004; Perry-Smith 2006). Such ties have been found to matter in a variety of contexts, including legislative politics, where Kirkland (2011) finds that weak (cosponsorship) ties between members of Congress increase legislative success. Other areas of application include finding a job (e.g., Granovetter 1974; Lin, Ensel, and Vaughn 1981; Lin 1982, 1988; de Graaf and Flap 1988; Marsden and Hurlbert 1988; Lai, Lin, and Leung 1998; Yakubovich 2005) and other individual advancement (e.g., Burt 1992, 1997, 2000; Podolny and Baron 1997; Seibert, Kraimer, and Liden 2001; Kavanaugh et al. 2005); crime and terrorism (e.g., CalvoArmengol, Verdier, and Zenou 2007; Patacchini and Zenou 2008; Kennedy and Weiman 2011); and neighborhoods (e.g., Henning and Lieberg 1996). 13. The social networks presented in figure 2.1 and figure 2.2 are hypothetical examples that, for the sake of illustration, greatly simplify actual social realities. For example, we recognize that the LMO Leader as well as LMO Members 1 and 2 also have unique networks comprised of strong ties with colleagues who are not members of the LMO. We omit some of these realities from our illustration to facilitate clarity.Page 220 → 14. Borgatti, Jones, and Everett (1998) highlight the differentiation between “social capital” viewed as a property of a group and that of an individual. 15. In other words, we do not seek to look at the advantages of brokerage across structural holes to

individual actors here, though we hope to do so in the future. 16. For excellent reviews of the extensive literature on knowledge networks, see Brass et al. 2004 and Phelps, Heidl, and Wadhwa 2012 . 17. An extensive literature supports the idea that network closure and strong ties enhance information exchange, knowledge transfer, learning, and innovation (see Ebadi and Utterback 1984; Krackhardt 1992; Abrahamson and Rosenkopf 1997; Uzzi 1997; W. Tsai and Ghoshal 1998; Morgan and Soerensen 1999; Bouty 2000; Kale, Singh, and Perlmutter 2000; Reagans and Zuckerman 2001; W. Tsai 2001; Morrison 2002; Reagans and McEvily 2003; Schulz 2003; Uzzi and Lancaster 2003; Cross and Sproull 2004; Levin and Cross 2004; Granovetter 2005; Obstfeld 2005; J. Singh 2005; Fleming, Mingo, and Chen 2007). 18. For similar arguments in economics, see Tullock 1985; Greif 1989. 19. On the informational value of strong ties, see, e.g., Ebadi and Utterback 1984; Krackhardt 1992; Scott and Bruce 1994; Szulanski 1996; W. Tsai and Ghoshal 1998; M. T. Hansen 1999; Bouty 2000; Kale, Singh, and Perlmutter 2000; Reagans and McEvily 2003; Schulz 2003; Uzzi and Lancaster 2003; Cross and Sproull 2004; Levin and Cross 2004; M. T. Hansen, Mors, and Løvås 2005; Moran 2005; Centola and Macy 2007. 20. Throughout this book, we use LMO network to describe the social network composed of the entire set of all LMO ties in a legislature. We talk about a single-LMO network when we refer specifically to a network composed exclusively of members of a single LMO. These two networks are identical in the hypothetical examples in figures 2.1 and 2.2, which (for presentational purposes) focus on a single LMO. 21. Wollebaek and Selle (2002) find that the effect of active participation and face-to-face contact in voluntary associations, as opposed to passive membership, is negligible at best when it comes to the creation of social capital and that “the relationship between trust and intensity of involvement is tenuous at best” (55). 22. Granovetter (1973, 1361; see also Marsden and Campbell 1984; Krackhardt 1992) conceives the strength of ties to be a function of a combination of four elements: the amount of time spent together, the emotional intensity of a relationship, the intimacy (mutual confiding) of a relationship, and the degree of reciprocation characterizing a relationship. It is all but impossible to capture intensity, intimacy, or reciprocation in the context of legislative politics and with the data available to us. We therefore focus on the likely amount of time spent to establish the relative strength or weakness of LMO ties. 23. Burt (1992), Uzzi (1996, 1997), and Chung, Singh, and Lee (2000) emphasize the benefits of having a mix of weak and strong ties. The utility of internally dense and externally bridging networks are highlighted by Morgan and Soerensen (1999), Reagans and Zuckerman (2001), Schilling and Phelps (2007), and Uzzi and Spiro (2005). As Phelps, Heidl, and Wadhwa (2012, 14) explain, referencing Hulsheger, Page 221 → Anderson, and Salgado (2009), “Greater internal density increases internal knowledge flows and a shared understanding of who knows what in a group, which allows it to utilize the diverse external knowledge inflows more effectively in creating knowledge.” This also relates to the “small world” principle (Watts 1999), which highlights the benefits of local clustering into “dense subnetworks, short paths between all actors, and relatively few ties” (Brass et al. 2004). 24. Our argument about the relationships between actors inside and outside the legislature is an informational one. We do not conceptualize the interactions between outside or inside advocates and the targets of their advocacy efforts in terms of rent-seeking behavior. This viewpoint is in line with much of the recent literature on lobbying, which highlights the role of lobbyists by emphasizing that interest groups and lobbyists play an important informational role in the creation of public policy (Schlozman and Tierney 1986; Potters and Van Winden 1990; Austen-Smith and Wright 1992; Austen-Smith 1993; Ricci 1993; Grossman and Helpman 2001; Broscheid and Coen 2003, 2007; Bouwen 2004a; Esterling 2004; Bouwen and McCown 2007; Mahoney 2008). This literature recognizes that policymakers possess imperfect information about the content and expected consequences of policies (see also Ringe 2010). Lobbyists, or outside advocates, can influence policy making by providing information to relevant decision makers. 25. Outside advocates may provide both political information (for example, about the preferences of key constituencies) and policy information (about the expected consequences of a policy proposal). Moreover, information exchange in LMOs is not one directional, with outside advocates simply supplying political and policy information to legislators. Information—in particular, political information about lawmakers’ priorities and preferences—also flows from legislators to lobbyists.

Chapter 3 1. We limit our sample to legislatures from advanced industrialized countries primarily for practical reasons. As emphasized previously, it is exceedingly difficult to gather reliable information on LMOs across a broad sample of countries, and the availability of data and country expertise tends to be greater when countries are members of international organizations such as the EU, the OECD, or the G-20. 2. Variable: DEMOC; “Polity IV Project” n.d. Malta is not included in this dataset, but as an EU member, it qualifies as an institutionalized democracy. 3. Questions are listed in appendix A. 4. We also received sixty-one responses to our inquiry that lacked substantive value—for example, respondents indicating that they did not know the answers to our questions or no longer considered themselves experts on the legislatures in question. 5. Questions are listed in appendix A. 6. One of the purposes of this chapter is to set the stage for the collection of additional cross-national information on LMOs, an effort that has already begun. Readers with knowledge about LMOs in the legislatures discussed in this chapter Page 222 → and beyond are encouraged to submit relevant information to the authors for inclusion in our database, available at http://bridgingtheinformationgap.wordpress.com. 7. Since we had reliable information on the U.S. Congress, we did not pursue this route here. 8. The ECPRD is a pan-European cooperation network connecting administrations of parliamentary chambers across Europe. 9. LMOs are not the only types of cross-party groups in many legislatures. What sets them apart, however, is that they are issue based. Most important, this characteristic differentiates LMOs from a very common type of cross-party group in several legislatures in our sample—what might be called “parliamentary friendship groups.” Such groups seek to maintain relationships with lawmaking institutions in other countries (for example, by facilitating visits and exchanges), and the legislatures in our sample that have these groups include Australia, Canada, Estonia, Finland, France, Germany, Japan, South Korea, Sweden, and the United Kingdom. Some countries—among them Latvia, Mexico, and Turkey—have no LMOs but do feature parliamentary friendship groups. In our survey and in this chapter, we acknowledge the existence of such groups and do not question their significance; our substantive focus, however, is exclusively on LMOs. 10. A registration system does not necessarily mean that all LMOs actually register with the parliament, since informal LMOs frequently exist. Our criterion with regard to the institutionalization of a given LMO system is that the parliament at least tries to register its LMOs. 11. While we are confident that the legislatures with institutionalized LMOs are categorized correctly, there is some degree of uncertainty in the classification of the noninstitutionalized LMOs in our sample as a consequence of the difficulties associated with finding reliable information on LMOs. Hence, it is possible that a given legislature tries to register its LMOs but does not publicize this information or that an LMO of which the respondents to our survey are unaware charges membership fees. In both cases, we might erroneously categorize a legislature as having noninstitutionalized LMOs rather than institutionalized ones. To maximize our confidence in our categorization, we carefully examined all survey responses for the level of uncertainty in the answers; the idea is that as the uncertainty in the survey responses increases, the likelihood that the LMO system is institutionalized decreases. While imperfect, this approach gives us some additional confidence in our categorization. 12. An alternative model specification might include a simple measure of district magnitude instead of the electoral system dummies. We conducted additional analyses with this variable (results not presented here) that produce an inferior model fit and show district magnitude to be statistically insignificant. 13. In our multivariate analyses, our rule of thumb is that with a small sample size, we can estimate parameters. Accordingly, we should be able to confidently estimate four or five coefficients in our models . This is roughly Page 223 → the same number of parameters we would get (4.5) using Long's (1997) suggestion of about 10 cases per parameter. 14. We also conduct multinomial logit analyses that predict the existence of institutionalized and noninstitutionalized LMOs. These analyses do not show systematic, statistically significant differences in the likelihood of observing either one of those two types of LMOs.

15. See www.publications.parliament.uk/pa/cm/cmallparty/memi01.htm. 16. See www2.parl.gc.ca/IIA/Default.aspx? Lang=E&DCId=4&DTId=1&P=other_interparliamentary_groups. 17. http://164.100.47.134/committee/Cells/FAQs.pdf. 18. The Centre for Legislative Research and Advocacy, based in New Delhi, is “an independent, not-forprofit, non-partisan initiative, which works to support and strengthen Parliament and legislatures so as to realise the values of democratic governance” (www.clraindia.org). 19. http://www.clraindia.org/parliamentary_forum.shtml. 20. http://www.camara.gov.br/internet/deputado/frentes.asp. 21. http://orka.sejm.gov.pl/SQL.nsf/listazesp?OpenAgent&6. 22. Polish president Lech Kaczyński, his wife, and numerous Polish government officials were among the 96 passengers who died in the crash of a Polish Air Force Tupolev Tu-154M aircraft on April 10, 2010, near Smolensk, Russia. 23. Using our various methods of research and data collection, we have unearthed relatively little information about some countries in our sample, including Italy, South Africa, and South Korea. Our method may have missed important information in these countries, and they may in fact have stronger LMO systems than we have discovered. We are well aware of the problem of measurement error, but we have some degree of confidence in our findings. 24. http://www.riigikogu.ee/index.php?id=34605. 25. LMOs in Lithuania are established on the basis of Article 26, Part 4, and Article 42 of the Statute of the Seimas. 26. http://www3.lrs.lt/pls/inter/w5_show?p_r=6117&p_k=z. 27. Parliamentarische Gruppen, intergroupes parlementaires, or intergruppi parlamentari, in the country's three major languages, German, French, and Italian. 28. http://www.parlament.ch/d/organe-mitglieder/bundesversammlung/adressen-websites-fotos/Documents /parlamentarische-gruppen.pdf. 29. Information about Bulgaria in this paragraph and the remainder of this section comes from Respondent 92. 30. See http://bulgaria.usaid.gov/1/news_item.html. 31. Respondent 92. 32. Bulgarian legislators did not have personal staff and were entirely dependent on party, parliamentary, or committee staff for policy information. 33. Respondent 92. 34. Within the context of the local government caucus, these concerns were mitigated by the direct involvement of the parliamentary leaders of the major parties, Page 224 → which signaled to the rank-andfile that it was acceptable or even desirable to join (Respondent 92). 35. See www.publications.parliament.uk/pa/cm/cmallparty/memi01.htm. 36. http://www.knesset.gov.il/lexicon/eng/lobby_eng.htm. The establishment and operation of members’ lobbies are regulated by a resolution adopted by the Knesset House Committee on April 1, 1997. 37. Article 8, Paragraph 7, of the Standing Orders. The formal basis for LMOs in Poland is Article 17, Paragraph 1, of the Act of May 9, 1996 on the Exercise of the Mandate of a Deputy or Senator (consolidated text in Dziennik Ustaw of 2003, No. 221, item 2199 with amendments), which states that deputies in the Sejm may create deputies’ clubs, groups, or groupings according to the principles specified by the Standing Orders (Article 8, Paragraphs 6 and 7). 38. Riigikogu Rules of Procedure Act, Paragraph 43. 39. Parliament Act (ParlA) of December 13, 2002, Chapter 6, Parliamentary Groups, Article 63. 40. Ato da Mesa 69, November 10, 2005, Article 2. 41. For Switzerland, see Parliament Act (ParlA) of December 13, 2002, Chapter 6, Parliamentary Groups, Article 63; for France, see http://www.assemblee-nationale.fr/connaissance/etudes.asp. 42. Ato da Mesa 69, November 10, 2005, Article 4. 43. http://www.knesset.gov.il/lexicon/eng/lobby_eng.htm. 44. For detailed journalistic accounts of contributions to APGs, see Ball 2011a, b. Data on the financing of APGs is available at http://www.guardian.co.uk/news/datablog/2011/feb/24/all-party-parliamentary-group-

lobbying. 45. www.assemblee-nationale.fr/connaissance/etudes.asp. 46. Officers run the group and are responsible for ensuring that it complies with the rules.

Chapter 4 1. Here and throughout we use the term Congress to refer to the U.S. House of Representatives. While the U.S. Senate has a few examples of caucuses, they are much less formal than those found in the House. There are, however, a handful of bicameral caucuses—LMOs in the United States whose membership explicitly includes both senators and representatives. In this volume, we focus our U.S. study exclusively on the House for two reasons, one practical and one theoretical. Practically, data on senators’ membership in caucuses is inconsistent and difficult to come by. Theoretically, the U.S. Senate operates with a strong set of social norms and traditions that encourages collegiality and discourse (see Matthews 1960), making the social networking benefits of caucuses less important in the Senate than in the House. 2. EP elections are carried out according to the national rules of the EU member states, but all use some form of proportional representation. Differences exist with regard to constituency boundaries (national or regional electoral districts), the Page 225 → method of seat allocation (e.g., Droop quota, d'Hondt system), the use of preferential voting, and the existence of electoral thresholds. 3. The constituent 2009–14 parliament had 736 members; the Treaty of Lisbon specifies that the EP will have no more than 751 members. 4. See http://www.europarl.europa.eu/aboutparliament/en/00c9d93c87/Intergroups.html. 5. The Conference of Presidents is composed of the EP president and the chairs of the political groups. 6. See Committee on House Administration, “Congressional Member Organizations,” available at http://cha.house.gov/index.php?option=com_content&view=article&id=333&Itemid=373. 7. A few members of Congress served long enough terms (although not complete terms) that we were able to obtain complete caucus and attribute data. These include Representative Judy Chu in the 111th Congress (winner of a special election to replace Hilda Solis, July 14, 2009), Laura Richardson in the 110th (winner of a special election to replace Juanita Millender-McDonald, August 21, 2007), Paul C. Broun Jr. in the 110th (winner of a special election to replace Charles Norwood, July 17, 2007), and Nicola (Niki) S. Tsongas in the 110th (winner of a special election to replace Martin Meehan, October 16, 2007). 8. Staffers are only imperfect proxies for members of Congress. However, U.S. representatives are much less available for personal interviews than are MEPs, and there is precedent in the literature for using staffers as proxies for members (see Whiteman 1995). 9. Authors’ calculations. This success rate is calculated by counting as accepted all amendments put forward by the EP under the co-decision procedure during the EP's 1999–2004 and 2004–9 terms that were accepted fully, partly, or in principle by the Council of Ministers. Excluded are co-decision dossiers where the bargain between EP and Council was the result of an informal compromise prior to first or second reading (Reh et al. 2013). 10. EP rules of procedure, Article 32, 7th Parliamentary Term—July 2010. 11. For a detailed account of the progression of these attempts, see Chabanet 2006. 12. Rules Governing the Establishment of Intergroups, Decision of the Conference of Presidents of the EP of December 16, 1999, consolidated on May 3, 2004, and amended by the Conference of Presidents on February 14, 2008. EP Doc. PE 339.492/BUR; references in the text are to the 2008 version. 13. Annex III to the rules governing meetings of the political groups. 14. Respondent 55. 15. Respondent 62. 16. Respondent 20. 17. Personal correspondence with S&D Secretariat. 18. Respondents 35, 36, 39, 41, 42, 43, 46, 56, 63, 64, 74. 19. Respondents 36, 39, 42, 43, 44, 45, 46, 56, 62, 63, 64, 74. 20. A copy of this letter to EP president Jerzy Buzek, written by 27 leaders of the current intergroups, is in the possession of Nils Ringe.Page 226 →

21. Respondent 40. 22. These groups were mentioned to the authors in one or more interview(s). 23. http://www.eufores.org. 24. http://www.globe-europe.eu. 25. http://www.kangaroogroup.eu. 26. http://www.spinelligroup.eu. 27. http://leftcaucus.eu. 28. This estimate is lower than the estimate we established using our methodology. It is ultimately impossible to establish the exact number of caucuses in any congress because there are no formal requirements for registration and reporting of membership. 29. Respondent 89. 30. Respondent 83. 31. See Committee on House Administration, “Congressional Member Organizations.” 32. http://www.legistorm.com/. 33. Respondent 77. 34. Congress has more than 80 official advisory boards and commissions that have been established through statute, including, for example, the House of Representatives Page Board, the John F. Kennedy Center for the Performing Arts Board of Trustees, and the Social Security Advisory Board. Some of these groups are made up entirely of members of Congress, while others (especially the advisory boards) include noncongressional participants. These groups exist outside of Congress's formal party and committee structures and are separate from the caucus system because caucuses are not established by statute, have no formal authority, and generally do not draw substantially on congressional resources. For more detail, see Glassman 2009. We have consciously excluded such groups from our study. 35. The information in this paragraph comes directly from author interviews. Respondent numbers are withheld to ensure anonymity. 36. We include the membership information on these single-party caucuses in our quantitative analyses for two reasons. First, some caucuses are explicitly and purposefully single-party (such as the New Democrats or the Republican Study Committee), but other caucuses are unintentionally single-party and may shift party composition over time. It would be arbitrary of us to identify and exclude all such groups from our analyses. Second, excluding such groups would bias our results in favor of our expectations regarding bridging ties across parties. To preclude this bias, we leave all such groups and their members as a part of the quantitative data. 37. For registration rules, see Committee on House Administration, “Congressional Member Organizations.” 38. Respondents 16, 35, 38, 41, 42, 44, 47, 55, 57, 71, 73. 39. Respondent 65. 40. Respondents 35, 55. 41. Respondents 39, 55, 69.Page 227 → 42. Respondents 39, 52, 55, 57. 43. Respondents 16, 18, 37, 38, 39, 40, 52, 55, 58, 60, 65, 67, 68, 71, 72. 44. Respondents 16, 55, 68. 45. To control for the autocorrelation introduced by legislators who served during more than one term in either the EP or Congress, we cluster the errors on individual members. To control for temporal autocorrelation, we include a dummy variable for each term. We use robust Huber-White standard errors to correct for heteroskedasticity. 46. Predicted probabilities were generated while holding other variables at their mean values, except for the party variables, where other parties are held at zero. The values were generated in STATA 10 using Clarify (King, Tomz, and Wittenberg 2000; Tomz, Wittenberg, and King 2003). 47. Differences in predicted probabilities are statistically significant at the .05 level if the confidence intervals do not overlap. 48. Coded as Southern Europe are Cyprus, Greece, Italy, Malta, Portugal, and Spain. The Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovenia, and Slovakia are counted as East-Central Europe. Northwestern Europe includes Austria, Belgium, Denmark, Finland, France, Germany, Ireland,

Luxembourg, the Netherlands, Sweden, and the United Kingdom. 49. Leadership is a dichotomous variable where party and committee chairs as well as members of the EP Bureau are coded 1. This variable is collected from Høyland, Sircar, and Hix 2009. 50. We also ran the analyses individually for EP6 and EP7 and found the following small differences across terms and in comparison to the pooled model: the GUE-NGL coefficient is not statistically significant in EP6; it is in EP7. Since the ECR did not yet exist in EP6, we do not have a result for that term. The leadership variable is not statistically significant in either EP6 or EP7. MEPs from Southern Europe are less likely to join LMOs than their colleagues from Northwestern Europe in EP6 but more likely to do so in EP7 (which explains the insignificant coefficient in the pooled model). Finally, having any LMOs at the national level does not increase the number of groups joined by MEPs in EP6 but does so in EP7. 51. The tenure variable is collected from Høyland, Sircar, and Hix 2009. 52. In some member states, voters may rank the candidates on a party list in order of preference. The preferential vote variable is based on Hix and Hagemann 2008. 53. Respondents 1, 2, 3, 4, 5, 8, 9, 10, 11, 12, 14, 15, 21, 25, 27, 28, 29, 30, 32, 33, 76, 78, 81, 88. 54. Respondents 1, 5. 55. Respondents 5, 11, 12, 14, 21, 24, 30, 34, 76, 81, 84. 56. Respondents 2, 8, 14, 23, 28, 31, 34, 80, 81, 88. 57. Respondents 1, 4, 08, 23, 76. 58. Respondents 2, 3, 11, 14, 24, 27, 30. 59. Predicted probabilities were generated while holding other variables at their mean values. The values were generated in STATA 10 using Clarify (King, Tomz, and Wittenberg 2000; Tomz, Wittenberg, and King 2003).Page 228 → 60. We also estimated a nonpooled series of models and found largely the same effects on the covariates, regardless of time; however, we find a few differences. The electoral percentage variable is statistically insignificant in the 109th and 111th models. Leadership is insignificant in the 109th model, and minority is insignificant in the 111th model. 61. Party leadership data are from U.S. House of Representatives, Office of the Clerk 2010b, where party leadership is defined as Speaker, majority/minority leader, majority/minority whip; Committee assignment data are from Nelson 1993, n.d.; Stewart and Woon 2009. Committee leadership is defined as serving as a committee chair or ranking member for one of the House's standing legislative committees. 62. Data for Asian Americans in Congress are from Tong 2010; data for Hispanic members are from Library of Congress 2010; for women in Congress, see Manning and Shogan 2009; for African Americans in Congress, see Manning and Shogan 2010. 63. Election data are from U.S. House of Representatives, Office of the Clerk, 2010a.

Chapter 5 1. Our qualitative data support the idea that LMO ties are social ties, which justifies our treatment of LMOs as one-mode networks. By this term, we mean that legislators’ memberships in LMOs can be arranged in a traditional incidence matrix where legislators (n) are the observations and LMOs (m) are the variables, so the matrix is n × m. We can then convert that two-mode matrix into a one-mode n × n matrix of legislators where the values in the cells describe their overlapping memberships in LMOs. 2. Respondents 20, 35, 36, 37, 39, 40, 41, 42, 44, 46, 47, 52, 55, 56, 57, 58, 59, 60, 61, 64, 65, 66, 67, 68, 69, 71, 72, 73. 3. Respondent 37. 4. Respondents 20, 35, 39, 40, 41, 46, 47, 52, 55, 59, 65, 66, 69. 5. Respondent 66. 6. Respondent 20. 7. Respondent 39. 8. Respondents 34, 35, 37, 38, 44, 47, 52, 58, 60, 66, 68, 69, 71. 9. Respondent 66. 10. Respondent 60.

11. Respondent 42. 12. Respondent 36. 13. Identification withheld to ensure anonymity. 14. Identification withheld to ensure anonymity. 15. Respondent 56. 16. Respondents 18, 40, 47. 17. Respondent 40. 18. Respondent 18.Page 229 → 19. Respondents 35, 58, 59, 60. 20. Identification withheld to ensure anonymity. 21. Respondent 32. 22. Respondents 2, 3, 5, 8, 13, 14, 31, 32, 76, 81, 83, 84, 87, 88. 23. Respondent 2. 24. Respondent 76. 25. Respondent 15. 26. Respondent 2. 27. Respondent 3. 28. Respondents 2, 3, 10, 25, 79, 84. 29. Respondents 1, 2, 3, 5, 8, 14, 23, 25, 28, 31, 32, 75, 76, 81, 83, 86. 30. Respondent 31. 31. Respondent 32. 32. Respondent 3. 33. Respondent 5. 34. Respondents 2, 6. 35. Respondents 6, 81. 36. Respondent 55. 37. Respondents 16, 17, 18, 20, 35, 40, 44, 46, 61, 65, 67, 69. 38. Respondent 35. 39. Respondents 16, 38, 58. 40. Respondents 18, 34, 39, 43, 45, 46, 47, 55, 56, 58, 59, 66, 67, 69, 70. 41. Respondent 39. 42. All of the respondents discuss this feature except 53, 61, 70, 72. 43. Respondent 36. 44. Respondent 56. 45. Respondent 43. 46. Respondents 16, 17, 19, 36, 41, 43, 45, 56. 47. Respondent 44. 48. Respondents 16, 17, 18, 45. 49. Respondents 16, 17. 50. Respondent 40; also Respondent 47. 51. Respondent 35. 52. Respondent 45. 53. Respondent 20; also Respondent 39. 54. Respondent 43. 55. Respondents 16, 20, 35, 39, 58, 59, 60, 65, 67, 72. 56. Respondent 20. 57. Respondent 65. 58. Respondent 59. 59. Respondent 63. 60. Respondent 34; also Respondent 60. 61. Respondent 47. 62. Respondent 60.Page 230 → 63. Respondents 16, 17, 19, 20, 34, 35, 39, 45, 52, 55, 59, 60, 62, 64, 67, 73.

64. Respondent 60. 65. Respondent 52. 66. Respondent 35. 67. Respondent 59. 68. Respondent 8. 69. Respondents 1, 3, 4, 5, 8, 9, 11, 12, 14, 15, 24, 25, 27, 28, 31, 32, 76, 77, 78, 85, 86. 70. Respondent 34. 71. Respondents 2, 3, 5, 8, 12, 21, 24, 32, 81, 83, 85, 86. 72. Respondent 31. 73. Respondents 3, 8, 31, 32, 89. 74. Respondent 1. 75. Respondents 4, 14, 25, 27, 78, 85. 76. Respondent 9. 77. Respondent 14. 78. Respondent 12. 79. Respondent 12. 80. Respondents 17, 20, 35, 42. 81. Respondent 59. 82. Of the 128 full days that MEPs spent in either Brussels or Strasbourg during 2010 (during plenary session weeks, group weeks, and committee weeks), 56 were devoted to parliamentary committees (authors’ calculations based on 2010 EP legislative calendar, available at: http://www.europarl.europa.eu/pdf/general /cal2010.pdf). 83. Respondents 41, 42, 44, 47, 59, 60, 65, 74. 84. Respondents 39, 40, 42, 52, 57, 58, 66. 85. Respondents 16, 37, 38, 42, 47, 52, 60, 61, 66, 74. 86. Respondents 16, 37, 38, 47, 52, 60, 61, 66. 87. While these numbers are based on estimates from our respondents, we acquired detailed information about the participation of MEPs in seven meetings of the Animal Welfare Intergroup between September 2009 and March 2010. A total of 38 MEPs from 14 member states and 6 party groups (all except the Europe of Freedom and Democracy group) participated in those meetings. Total participation ranged from 8 to 21 MEPs, with a mean of 15.43. 88. Respondents 18, 19, 20, 35, 39, 40, 42, 44, 45, 46, 47, 55, 56, 57, 59, 60, 62, 63, 64, 66, 69, 71, 74. 89. Respondents 34, 35, 36, 39, 42, 45, 46, 47, 52, 55, 63, 65, 69, 74. 90. Respondents 10, 12, 14, 24, 75, 77. 91. Respondents 11, 13, 26, 31, 80. 92. Respondents 2, 76, 79. 93. Respondents 75, 82, 84. 94. Respondent 14. 95. Data and estimate compiled by the authors. 96. Respondents 1, 2, 3, 11, 13, 24, 30, 31, 32, 34, 82, 84, 86.Page 231 → 97. Respondents 2, 3, 4, 5, 8, 9, 10, 11, 12, 21, 22, 30, 31, 34, 75, 76, 79, 84, 85. 98. Respondent 79. 99. Respondent 76. 100. Respondent 14. 101. Respondent 15. 102. Respondents 3, 4, 9, 10, 11, 14, 21, 78, 81. 103. Respondent 35. 104. We include all MEPs of the 2004–9 legislative term who joined the EP before the end of 2004. That is, our list includes those who were members immediately following the June 2004 election as well as a handful of members who joined later in the year. The reasoning here is that members who joined relatively early on in the term still would have had a chance to join intergroups as they are in the process of being created, which usually happens over the first few months of a legislative term. We exclude members who joined after 2004 because membership lists that predate their arrival do not include them.

105. Congresses do not necessarily have exactly 435 voting members because some members serve partial terms because of death, resignation, or special elections. We generally exclude such members from our analysis. 106. We collected two sets of membership lists for the 2004–9 term for three intergroups. The first are the official lists submitted to the EP leadership when the groups requested intergroup status. The second are those lists provided by the intergroups in response to our request. The correlation between the lists is .96. The network analysis in this section relies on the official lists submitted to the Conference of Presidents; however, we also conduct each part of the analysis using the unofficial lists. We do not find any substantively notable differences with regard to our descriptive network measures presented in this chapter, such as network density and average geodesic distance. We do, however, find that several MEPs are not included in the list of the 25 most central actors when using the unofficial membership lists—van den Burg, de Keyser, Deprez, Grossetête, van Lancker, McCarthy, Pleguezuelos Aguilar, Radwan, Jalowiecki, and Savary. In their stead are Demetriou (EPP, Cyprus), Dess (EPP, Germany), Ebner (EPP, Italy), Hassi (Verts /ALE, Finland), Koch (EPP, Germany), Matsakis (ALDE, Cyprus), Nicholson of Winterbourne (ALDE, UK), Olajos (EPP, Hungary), Seeber (EPP, Austria), Titley (PES, UK), and Vatanen (EPP, France). The members who drop out of the top 45 most connected dyads are Morillon, Ulmer, Kratsa-Tsagaropoulou, and Ludford. Included are Breyer, Evans, Berman, and Resetarits, all of whom are among the 20 most central actors in the official data. 107. Respondents 2, 5, 6, 7, 8, 10, 11, 12, 14, 15, 23, 24 , 26, 30, 31, 33, 36, 41, 42, 47, 59, 60, 65, 74, 78, 81, 84. 108. The committee network for EP6 is much denser and features smaller distances between members than does the EP7 committee network, because turnover occurs in committee membership over time (especially halfway through the EP's five-year term), providing opportunities for additional ties to be established. At any one point in time, however, only a set number of connections exist through joint committee membership, which is why comparing the LMO networks to the EP7 Page 232 → committee network (which captures committee membership in the first two years of the new term) is more appropriate here. 109. In reference to figure 2.1, the LMO network is composed of the LMO Leader and LMO Members 1–5, while the legislative network includes all 18 nodes. 110. EP delegations maintain relations and exchange information with parliaments in non-EU countries. 111. The Bureau consists of the president of the EP, the 14 vice presidents, and the 5 quaestors. 112. The College of Quaestors is responsible for administrative and financial matters directly concerning MEPs and their working conditions. 113. The Conference of Presidents consists of the EP president and the chairs of the political groups. 114. To gather this information we relied on Høyland, Sircar, and Hix 2009 and on Simon Hix's EP data, available at http://personal.lse.ac.uk/hix/. 115. In the equivalent analyses for Congress, we randomly draw one legislator from approximately each quartile in the distribution of total caucus memberships. The distribution of intergroup memberships does not lend itself to this procedure. 116. To be precise, William Shuster (R, PA-9) was drawn randomly from the 25.45 percent of members of Congress in the 109th Congress who are members of 1–3 caucuses, Janice Schakowsky (D, IL-9) was drawn from the 24.42 percent who joined 4–7 caucuses, Martin Sabo (D, MN-5) was drawn from the 24.16 percent who joined 8–12 caucuses, and Earl Blumenauer (D, OR-3) was drawn from the 25.96 percent who joined 13–66 caucuses. For the 110th Congress, Jeff Flake (R, AZ-6) was selected randomly from the 26.07 percent of members who joined 1–9 caucuses, Kevin Brady (R, TX-8) was drawn from the 24.64 percent who joined 10–18 caucuses, Ron Lewis (R, KY-2) was drawn from the 24.41 percent who joined 19–27 caucuses, and Edward Brown (R, SC-1) was drawn from the 24.88 percent who joined 28–84 caucuses. Finally, for the 111th Congress, Peter Roskam (R, IL-6) was drawn randomly from the 24.77 percent of members who joined 1–10 caucuses, Geoff Davis (R, KY-4) was drawn from the 25.23 percent who joined 11–23 caucuses, Candace Miller (R, MI-10) was drawn from the 24.7 percent who joined 24–37 caucuses, and Rick Larsen (D, WA-2) was drawn from the 24.3 percent who joined 38–92 caucuses. 117. We do not exclude such single-party caucuses from our network analyses because doing so might bias our empirical findings in favor of our hypotheses. 118. Brown joined more caucuses but has fewer LMO ties than does Lewis because Lewis is a member of

more large caucuses than is Brown. 119. All constraint scores are calculated for networks with valued ties; the results are equivalent, however, when we dichotomize the ties.

Chapter 6 1. Respondents 18, 20, 35, 36, 37, 38, 39, 40, 42, 46, 60, 66, 68, 74. 2. Respondent 66.Page 233 → 3. Respondent 55. 4. Respondent 59. 5. Respondent 41; also Respondent 47. 6. Respondents 66, 67. 7. Respondent 63. 8. Respondent 34. 9. Respondent 64. 10. Respondents 35, 38, 39, 41, 52, 55, 59. 11. Respondent 55. 12. Respondent 52. 13. Respondent 41. 14. Respondent 39; also Respondent 56. 15. Respondent 16; also Respondents 59, 69. 16. Respondent 41. 17. Respondent 19. 18. Respondent 18. 19. Respondents 16, 18, 34, 38, 39, 44, 45, 46, 62, 64. 20. One exception is the Intergroup on Lesbian, Gay, Bisexual, and Transgender (LGBT) Rights, which does not hold public meetings “as a rule” (Respondent 62). 21. Source identifications for these quotations are withheld to ensure anonymity. 22. Respondents 16, 18, 20, 34, 35, 36, 40, 42, 43, 45, 47, 46, 52, 55, 56, 59, 62, 63, 66, 67, 69, 71, 73. 23. Respondent 34; also Respondent 74. 24. Respondent 42; also Respondent 47. 25. Respondent 63. 26. Identification withheld to ensure anonymity. 27. Respondent 1. 28. Respondent 25. 29. Respondent 1. 30. Respondent 3. 31. Respondent 6. 32. Respondent 6. 33. Respondents 31, 75, 84. 34. Respondents 1, 3, 4, 5, 9, 10, 11, 12, 13, 14, 15, 22, 23, 25, 26, 28, 29, 30, 31, 33, 34, 75, 76, 77, 78, 79, 80, 81, 88, 90. 35. Respondent 15. 36. Respondents 3, 4, 8, 11, 13, 15, 23, 25, 34, 76, 78, 88, 90. 37. Respondents 1, 21. 38. Respondent 34. 39. Respondents 2, 10, 11, 12, 13, 14, 24, 26, 31, 75, 76, 77, 78, 79, 80, 82, 84. 40. Respondents 4, 8, 31, 75, 78, 79, 86. 41. Respondents 2, 4, 9, 11, 12, 13, 24, 28, 30, 31, 32, 75, 76, 84, 88. 42. Respondent 24. 43. Respondent 30.Page 234 → 44. Respondent 28.

45. Respondent 21; also Respondents 4, 23. 46. Respondents 2, 3, 4, 6, 8, 9, 11, 12, 28, 29, 32, 34, 75, 76, 78, 81, 84. 47. Respondent 21. 48. Respondent 28. 49. Respondent 31. 50. Respondent 15; also Respondents 23, 24. 51. Respondent 2; also Respondents 6. 52. Respondent 6; also Respondent 81. 53. Respondent 2. 54. Respondent 24. 55. Respondent 27. 56. Respondent 31. 57. Respondent 82. 58. Respondent 10. 59. Respondent 76. 60. Respondent 90. 61. Respondents 8, 10, 23, 25, 33, 90. 62. Respondent 39. 63. Respondent 37. 64. Respondent 55. 65. Identification withheld to ensure anonymity. 66. Respondent 35. 67. Respondent 38. 68. To maintain anonymity, these examples are purely hypothetical. The authors are aware of several comparable real-world examples, however. 69. Respondents 11, 13, 21, 34, 81, 84. 70. Respondent 39; also Respondents 55, 69. 71. Respondents 1, 3, 4, 5, 6, 11, 14, 23, 25, 31, 32, 76, 79, 81. 72. Respondent 15. 73. Respondent 31. 74. Respondent 31. 75. These relationships appear in the intergroups’ financial declarations to the EP or were described by our respondents. 76. Respondent 36. 77. www.atd-quartmonde.org. 78. Respondent 45. 79. www.union-habitat.eu. 80. Respondent 67. 81. Respondents 45, 55, 56, 67, 70, 72. 82. Respondent 34. 83. Respondents 17, 20, 40. 84. Respondent 36. 85. Respondent 56.Page 235 → 86. Respondents 19, 20. 87. Respondent 36. 88. Respondents 37, 52, 61. Examples of groups that provide such services include the Kangaroo Group's 2007 Spring Party, which was “generously sponsored by BP, UPEI, Roche, Deutsche Post World Net, European Smoking Tobacco Association (ESTA), German Savings Banks Association and the Scott-Wilson Partnership” (natrium.be/www.kangaroo2004.be/N/020_list1_events_L.php). 89. Respondents 46, 61. 90. Respondent 55. 91. Respondents 34, 35, 66. 92. Respondents 18, 34, 41, 43, 56.

93. Respondents 37, 38, 40, 56, 59, 61, 63, 64, 66. 94. Respondents 38, 43, 45. 95. Respondents 40, 55, 69. 96. Respondent 62. 97. Respondents 36; also Respondents 39, 40. 98. Respondent 46. 99. Respondent 64. 100. Respondent 67. 101. Respondent 43. 102. Respondents 1, 2, 3, 8, 11, 12, 14, 23, 25, 28, 29, 30, 31, 75, 79, 86, 88. 103. Respondent 10. 104. Respondents 1, 2, 3, 4, 9, 11, 12, 14, 15, 21, 24, 26, 28, 31, 33, 34, 75, 76, 77, 78, 82, 88, 90. 105. Respondent 78. 106. Respondent 79. 107. Respondent 4. 108. Respondents 1, 4, 8, 12, 15, 21, 22, 25, 28, 32, 76, 78, 79, 88, 90. 109. Respondent 4. 110. Respondents 29, 75, 88. 111. Respondent 33. 112. Respondents 4, 22, 29. 113. Respondents 32, 78, 88. 114. Respondents 28, 29. 115. Respondent 29. 116. Respondent 21. 117. Respondents 23, 34, 78, 79, 81. 118. Respondent 87. 119. Respondent 84. 120. Respondent 79; also Respondent 75. 121. Respondents 4, 25, 28, 85. 122. Respondent 29, 82. 123. Respondents 32, 78, 88. 124. Respondent 27.Page 236 → 125. Respondent 27. 126. Respondent 1. 127. Respondent 29. 128. Respondent 4. 129. Respondent 27. 130. Respondents 1, 2, 3, 11, 12, 14, 15, 25, 31, 76. 131. Respondent 6; also Respondents 4, 25. 132. Respondent 22. 133. Respondent 75. 134. Respondent 75; also Respondents 77, 87. 135. Respondent 3. 136. Respondent 43. 137. Respondent 45; also Respondents 16, 18, 19, 20, 44, 56, 59, 60. 138. Respondent 46; also Respondents 43, 56, 67. 139. Respondents 18, 19, 35, 45, 46, 47, 64. 140. Respondent 46. 141. Respondents 19, 46, 47. 142. Respondent 20; also Respondents 38, 40. 143. Respondent 74; also Respondent 47. 144. Respondent 68. 145. Respondents 19, 36, 39, 40, 56, 67, 69, 72.

146. Respondents 17, 18, 19, 34, 40, 47, 55, 56, 58, 61, 64, 71. Not all intergroups distribute position papers from outside organizations. One respondent emphasizes that her group would not do so under any circumstances (Respondent 44). 147. Respondent 19. 148. Respondent 40. 149. Respondents 36, 56, 67, 70. 150. Respondent 67. 151. Respondent 36. 152. Respondent 18; also Respondents 40, 55. 153. Respondent 46. 154. Respondent 72. 155. Respondent 55. 156. Respondent 38. 157. Respondent 18. 158. Respondent 44. 159. The Monti-Kroes package is also known as the Services of General Economic Interest package and seeks to define the conditions under which state aid in the form of public service compensation can be considered compatible with the EU rules (see also http://ec.europa.eu/competition/state_aid/legislation /sgei.html). 160. Respondent 70. 161. Personal correspondence, May 3, 2011. 162. Respondents 34, 36, 37, 47, 52, 56, 57, 62, 63, 64, 71. 163. Respondents 34, 39, 57, 62, 70.Page 237 → 164. Respondents 17, 18, 19, 34, 35, 44. 165. Respondents 39; also Respondents 45, 47. 166. Respondents 18, 35, 39, 59, 63. 167. Respondents 38, 39, 60, 64. 168. Respondent 41. 169. Respondent 39. 170. Respondent 56. 171. Respondent 43. 172. Respondents 16, 17, 18, 20, 40, 44, 46, 61, 65, 67, 69. 173. Respondent 36. 174. Respondent 35. 175. Respondent 56. 176. Respondent 18. 177. Respondent 37. 178. Respondent 59. 179. Respondents 1, 4, 5, 6, 10, 11, 12, 14, 15, 25, 27, 28, 29, 30, 31, 33, 75, 76, 77, 78, 79, 80, 84, 85, 86, 87, 88. 180. Respondent 2; also Respondents 11, 12, 22, 25. 181. Respondent 11. 182. Respondent 1. 183. Identification withheld to ensure anonymity. 184. Respondent 2. 185. Respondents 75, 78, 79. 186. Respondent 12. 187. Respondent 78. 188. Respondent 29. Medical field in question and name of the staffer changed to ensure anonymity. 189. Respondent 28. 190. Respondent 31. 191. Respondent 76. 192. Respondent 27.

193. Respondent 1; also Respondent 84. 194. Respondent 28. 195. Respondent 1. 196. Respondent 13; also Respondents 28, 29. 197. Respondent 15. 198. Respondent 14. 199. Respondent 16. 200. Respondent 40. 201. Respondent 46. 202. Respondent 67. 203. Respondent 60. 204. Respondent 86. 205. Respondent 79.Page 238 → 206. Respondent 36. 207. Respondent 61. 208. Respondent 38. 209. Respondent 41. 210. Respondent 66; also Respondents 16, 20. 211. Respondent 12. 212. Respondent 19; also Respondents 40, 46. 213. Respondents 18, 37, 38, 40, 47, 59, 61, 63, 64, 67, 68. 214. Respondent 1. 215. Respondent 38. 216. Respondents 9, 37, 38, 40, 56, 59, 61, 63, 64, 66, 86. 217. Respondent 40. 218. Respondent 64; also Respondent 63. 219. Respondent 59. 220. Respondent 9. 221. Respondent 86. 222. Respondents 40, 56. 223. Respondent 68. 224. Respondent 56. 225. Respondent 46. 226. Respondent 11; also Respondents 33, 34, 76, 90. 227. Respondent 77; also Respondent 22. 228. Respondent 4. 229. Respondent 36. 230. Respondent 40. 231. Respondent 38. 232. Respondent 63. 233. Respondent 86; also Respondent 28. 234. Respondent 18. 235. Respondent 29. 236. Respondent 46. 237. Respondents 5, 9, 10, 11, 12, 14, 21, 22, 24, 25, 27, 28, 30, 34, 76, 78, 81, 84; see also Miler 2011. 238. Respondents 2, 8, 14, 28, 31, 34, 80, 81, 88. 239. Respondent 23. 240. Respondent 5. 241. Respondents 23, 34, 90. 242. Respondents 1, 4, 5. 243. While only simple majority is needed in the EP's first reading stage, a majority of all MEPs (present or not) is required in second reading. This is the highest hurdle that a piece of legislation may need to overcome.

244. Of course, there may be a need for bipartisanship if different parties hold majorities in House and Senate. Furthermore, the need for bipartisanship may be stronger in the Senate, which is constrained in many instances by supermajoritarian Page 239 → rules, such as the filibuster. Caucuses, however, are not prevalent in the Senate, at least in part because of its smaller size (see chapter 3) and senators’ individual agenda-setting power.

Chapter 7 1. Respondent 62. 2. Respondents 19, 22, 23, 38, 40, 63, 68. 3. Respondent 22. 4. Respondents 35, 38, 39, 47, 68. 5. Respondents 35, 36, 37, 43, 47, 57, 64, 68, 72. 6. Respondent 37 also indicates that his intergroup seeks to influence the Council of Minister's agenda, but given that the Council, unlike the Commission, lacks the power of legislative initiative, it is not surprising that this option was not mentioned more often. 7. Respondents 47, 62. Parliamentary questions in the EP fall into three categories: questions that are “put to the Council or the Commission by a committee, a political group or at least 40 Members with a request that they be placed on the agenda of Parliament” (EP Rules of Procedure, Rule 115); questions “during Question Time with the Council and the Commission” (Rule 116); and questions “for written answer to the President of the European Council, the Council, or the Commission” (Rule 117). Our respondents’ answers do not indicate that intergroups use one or more of this options with particular frequency. 8. Respondents 35, 39, 56, 62. 9. Respondents 20, 35. Initiative reports come in two variants. First, Rule 48 of the EP Rules of procedure describes “own-initiative reports”: “In the areas where the treaties give the European Parliament the right of initiative, its committees may draw up a report on a subject within its remit and present a motion for a resolution to Parliament.” Second, Rule 42 specifies that “on the basis of a report by one of its committees, under Article 192, second paragraph, of the EC Treaty, Parliament, acting by a majority of its Members, may request the Commission to submit any appropriate legislative proposal. The Parliament committee responsible must first ask the Conference of Presidents for authorization. The Commission may agree or refuse to submit the proposal requested.” Our impression is that intergroups primarily submit own-initiative reports. 10. Respondents 35, 42, 55, 62. Rule 123 of the EP Rules of Procedure provides that “a group of up to five MEPs can submit a written declaration by presenting a text to be signed by their colleagues. If the declaration is signed by a majority of the MEPs, it is forwarded to the President, who announces it in plenary. At the end of the part-session, the declaration is forwarded to the institutions named in the text, together with the names of the signatories.” 11. Respondents 3, 9, 10, 24, 25, 81. 12. Respondent 75. 13. Respondent 76.Page 240 → 14. Respondent 14. 15. According to Van Schendelen (2005, 77–78), “The Commission often feels very sympathetic toward any new issue-creation by an intergroup.” 16. Respondent 82. 17. Reference withheld to ensure anonymity. 18. Respondent 24. 19. The existing literature contains a small number of anecdotal examples of intergroup influence. Corbett, Jacobs, and Shackleton (2007, 188), for example, provide a historical example of intergroup influence by highlighting the role of the so-called Crocodile Club (named after the Strasbourg restaurant where it first met) in promoting the EU's single market in the 1980s (see also Chabanet's [2006, 12] discussion of the Kangaroo Club). Corbett, Jacobs, and Shackleton (2007, 190) also single out the Animal Welfare Intergroup as a successful actor that played “a prominent role in a number of major animal welfare initiatives,

including new legislation,” providing a series of examples. Finally, Butler (2008, 277–78) provides a more specific example for the policy influence of intergroups by discussing the role of the Disability Intergroup regarding a legislative proposal concerning the rights of disabled persons and persons with reduced mobility when traveling by air. 20. Identification withheld to ensure anonymity. 21. Identification withheld to ensure anonymity. 22. Identification withheld to ensure anonymity. 23. Identification withheld to ensure anonymity. 24. Identification of the dossier would compromise anonymity of the respondent. 25. Identification withheld to ensure anonymity. 26. Identification withheld to ensure anonymity. 27. Respondents 16, 20, 45, 64, 56, 62. 28. Respondent 64. 29. Respondent 62. 30. Identification withheld to ensure anonymity. 31. We are grateful to Justin Kirkland and James Fowler for supplying cosponsorship data. We do not include analyses for the 111th Congress because complete cosponsorship data are not yet available. 32. We also examined the fit statistics and graphs of these models and found the model to fit the data well. Plots of fit graphs are available at http://bridgingtheinformationgap.wordpress.com. 33. We thank one of our anonymous reviewers for urging us to include the degree variable. 34. This approach is akin to the statistical modeling strategy adopted by other scholars who use egocentered or individual-level analysis even with network measures. See, e.g., Cho and Fowler 2010. 35. http://www.senate.gov/pagelayout/reference/two_column_table/Resumes.htm. Data compiled by authors. 36. We also estimated models in which we substituted normalized betweenness centrality in place of eigenvector centrality, and the results vary. In the model where Page 241 → the dependent variable is bill count, both betweenness centrality and degree are positive and significant. However, in the models for bills that pass the House and become law, none of the centrality measures or degree reach statistical significance. In addition, we estimated another series of models where we only included a measure of normalized degree centrality (and no measure of degree, because it is highly correlated); we found that in the bill count model, the centrality measure is positive and significant, and in the other two models it is positive but statistically insignificant. The lack of robustness across these models does not overly concern us because the theoretical case for inclusion of eigenvector centrality is more compelling than the case for these alternative measures. 37. Respondent 23. 38. Respondent 11. 39. Respondent 16. 40. We measure polarization using Poole and Rosenthal's NOMINATE scores and calculating the average difference between the parties. Data from voteview.com (Poole et al. 2011; see also Poole and Rosenthal 1997, 2007). 41. See Radnitz 2011 for an excellent review of a series of recent books on the interactions between formal institutions and informal politics.Page 242 →

Page 243 →

REFERENCES Abrahamson, Eric, and Lori Rosenkopf. “Social Network Effects on the Extent of Innovation Diffusion: A Computer Simulation.” Organization Science 8 (1997): 289–309. Adler, E. Scott, and John Wilkerson. “Congressional Bills Project: 2005–2008, NSF 00880066 and 00880061.” http://congressionalbills.org. Ainsworth, Scott H. “Regulating Lobbyists and Interest Group Influence.” Journal of Politics 55 (1993): 41–56. Ainsworth, Scott H. “The Role of Legislators in the Determination of Interest Group Influence.” Legislative Studies Quarterly 22 (1997): 517–33. Ainsworth, Scott H., and Frances Akins. “The Informational Role of Caucuses in the U.S. Congress.” American Politics Research 25 (1997): 407–30. Aldrich, Howard E. Organizations Evolving. Thousand Oaks, CA: Sage, 1999. Ancona, Deborah G., and David F. Caldwell. “Demography and Design: Predictors of New Product Team Performance.” Organization Science 3 (1992): 321–41. Arnold, Laura W., Rebecca E. Deen, and Samuel C. Patterson. “Friendships and Votes: The Impact of Interpersonal Ties on Legislative Decision Making.” State and Local Government Review 32 (2000): 142–47. Arnold, R. Douglas. The Logic of Congressional Action. New Haven: Yale University Press, 1990. Asch, Solomon E. “Studies of Independence and Conformity: A Minority of One against a Unanimous Majority.” Psychological Monographs 70 (1956): 258–90. Austen-Smith, David. “Strategic Models of Talk in Political Decision Making.” International Political Science Review 13 (1992): 45–58. Austen-Smith, David. “Information and Influence: Lobbying for Agendas and Votes.” American Journal of Political Science 37 (1993): 799–832. Austen-Smith, David, and John R. Wright. “Competitive Lobbying for a Legislator's Vote.” Social Choice and Welfare 9 (1992): 229–57. Austen-Smith, David, and John R. Wright. “Counteractive Lobbying.” American Journal of Political Science 38 (1994): 25–44. Ball, James. 2011a. “Coalition Urged to Act over Lobbyists Who Use Party Groups ‘to Buy Influence.’” Guardian, February 24, 2011. Ball, James. 2011b. “Parliamentary groups flourish as David Cameron fails to deliver reforms.” Guardian, February 24, 2011. Barabasi, Albert-László , and Réka Albert. “Emergence of Scaling in Random Networks.” Science 286 (1999): 509–12. Page 244 → Bauer, Raymond A., Ithiel de Sola Pool, and Lewis A. Dexter. American Business and Public Policy: The Politics of Foreign Trade. New York: Atherton, 1963.

Baughman, John. Common Ground: Committee Politics in the U.S. House of Representatives. Palo Alto, CA: Stanford University Press, 2006. Baum, Joel A. C., Tony Calabrese, and Brian S. Silverman. “Don't Go It Alone: Alliance Network Composition and Startups’ Performance in Canadian Biotechnology.” Strategic Management Journal 21 (2000): 267–94. Baumgartner, Frank, Jeffrey M. Berry, Marie Hojnacki, David C. Kimball, and Beth L. Leech. Lobbying and Policy Change: Who Wins, Who Loses, and Why. Chicago: University of Chicago Press, 2009. Baumgartner, Tom, and Tom R. Burns. “The Structuring of International Economic Relations.” International Studies Quarterly 19 (1975): 126–59. Beazley, Christopher, Michael Gahler, Satu Hassi, Toomas Henrik Ilves, Ģirts Valdis Kristivskis, Henrik Lax, and Alexander Stubb. Europe's Strategy for the Baltic Sea Region. November 2005. http://www.eu300.eu/uploads /media/Europes-Strategy-for-the-Baltic-Sea-Region_1_.pdf. Benedetto, Giacomo. “The European Parliament: Consensus and Coordination for Enhanced Powers.” In PolicyMaking Processes and the European Constitution: A Comparative Study of Member States and Accession Countries, edited by Thomas König and Simon Hug. London: Routledge, 2006. Berelson, Bernard R., Paul F. Lazarsfeld, and William N. McPhee. Voting: A Study of Opinion Formation in a Presidential Election. Chicago: University of Chicago Press, 1954. Bernhagen, Patrick, and Thomas Bräuninger. “Structural Power and Public Policy: A Signaling Model of Business Lobbying in Democratic Capitalism.” Political Studies 53 (2005): 43–64. Berry, Jeffrey M. The Interest Group Society. New York: Harper Collins, 1989. Beyers, Jan. “Voice and Access: Political Practices of European Interest Associations.” European Union Politics 5 (2004): 211–40. Boessen, Sandra, and Hans Maarse. “A Ban on Tobacco Advertising: The Role of Interest Groups.” In Lobbying the European Union: Institutions, Actors, and Issues, edited by David Coen and Jeremy Richardson. Oxford: Oxford University Press, 2009. Bogue, Allan G., and Mark P. Marlaire. “Of Mess and Men: The Boardinghouse and Congressional Voting, 1821–1842.” American Journal of Political Science 19 (1975): 207–30. Bonacich, Philip. “Factoring and Weighing Approaches to Clique Identification.” Journal of Mathematical Sociology 2 (1972): 113–20. Borgatti, Stephen P., Martin G. Everett, and Linton C. Freeman. 2002. Ucinet for Windows: Software for Social Network Analysis. Harvard, MA: Analytic Technologies. http://www.analytictech.com/downloaduc6.htm. Borgatti, Stephen P., Candace Jones, and Martin G. Everett. “Network Measures of Social Capital.” Connections 21 (1998): 27–36. Bouty, Isabelle. “Interpersonal and Interaction Influences on Informal Resource Page 245 → Exchanges between R&D Researchers across Organizational Boundaries.” Academy of Management Journal 43 (2000): 50–65. Bouwen, Pieter. “Exchanging Access Goods for Access: A Comparative Study of Business Lobbying in the European Union Institutions.” European Journal of Political Research 43 (2004a): 337–69. Bouwen, Pieter. “The Logic of Access to the European Parliament: Business Lobbying in the Committee on Economic and Monetary Affairs.” Journal of Common Market Studies 42 (2004b): 473–95.

Bouwen, Pieter, and Margaret McCown. “Lobbying versus Litigation: Political and Legal Strategies of Interest Representation in the European Union.” Journal of European Public Policy 14 (2007): 422–43. Brass, Daniel J. “A Social Network Perspective on Human Resources Personnel.” Research in Personnel and Human Resources Management 13 (1995): 39–79. Brass, Daniel J., Joseph Galaskiewicz, Henrich R. Greve, and Wenpin Tsai. “Taking Stock of Networks and Organizations: A Multilevel Perspective.” Academy of Management Journal 47 (2004): 795–817. Bratton, Kathleen A., and Stella M. Rouse. “Networks in the Legislative Arena: How Group Dynamics Affect Cosponsorship.” Legislative Studies Quarterly 36 (2011): 423–60. Bräuninger, Thomas, and Martin Brunner. “Cosponsorship Networks in the German Bundestag: Evidence for Personal or Party Behavior?” Paper presented at the 67th annual national conference of the Midwest Political Science Association, Chicago, April 2–5, 2009. Broscheid, Andreas, and David Coen. “Insider and Outsider Lobbying of the European Commission: An Informational Model of Forum Politics.” European Union Politics 4 (2003): 165–89. Broscheid, Andreas, and David Coen. “Lobbying Activity and for a Creation in the EU: Empirically Exploring the Nature of the Policy Good.” Journal of European Public Policy 14 (2007): 346–65. Browne, William P., and Won K. Paik. “Beyond the Domain: Recasting Network Politics in the Post Reform Congress.” American Journal of Political Science 37 (1993): 1054–78. Bullock, Charles S., III, and Burdett Loomis. “The Changing Congressional Career.” In Congress Reconsidered, 3rd ed., edited by Lawrence Dodd and Bruce Oppenheimer. Washington, D.C.: Congressional Quarterly Press, 1985. Burger, Timothy. “No Action after Three Studies; Roberts Calls LSOs ‘an Accident’ Waiting to Happen.” Roll Call, June 4, 1992. Burkett, Tracy, and John Skvoretz. “Political Support Networks among U.S. Senators: Stability and Change from 1973–1990.” Unpublished paper, College of Charleston, 2001. http://web1.cas.usf.edu/MAIN/include/0-1201-000 /polsupportnet.pdf. Burt, Ronald S. Toward a Structural Theory of Action. New York: Academic Press, 1982. Burt, Ronald S. Structural Holes. Cambridge: Harvard University Press, 1992. Page 246 → Burt, Ronald S. “The Contingent Value of Social Capital.” Administrative Science Quarterly 42 (1997): 339–65. Burt, Ronald S. “The Network Structure of Social Capital.” Research in Organizational Behavior 22 (2000): 345–423. Burt, Ronald S. “The Social Capital of Structural Holes.” In The New Economic Sociology: Developments in an Emerging Field, edited by Mauro F. MGuillén, Randall Collins, Paula England, and Marshall Meyer. New York: Sage, 2002. Burt, Ronald S. “Structural Holes and Good Ideas.” American Journal of Sociology 110 (2004): 349–99. Butler, Israel De Jesús. “Non-Governmental Organisation Participation in the EU Law-Making Process: The Example of Social Non-Governmental Organisations at the Commission, Parliament, and Council.” European Law Journal 14 (2008): 558–82.

Butts, Carter T., Mark S. Handcock, David R. Hunter, and Martina Morris. “Network: Classes for Relational Data.” R package version 1.6 (January 28, 2011). Irvine, CA. http://statnet.org/ Caldeira, Gregory A., John A. Clark, and Samuel C. Patterson. “Political Respect in the Legislature.” Legislative Studies Quarterly 18 (1993): 3–28. Caldeira, Gregory A., and Samuel C. Patterson. “Political Friendship in the Legislature.” Journal of Politics 49 (1987): 953–75. Caldeira, Gregory A., and Samuel C. Patterson. “Contours of Friendship and Respect in the Legislature.” American Politics Quarterly 16 (1988): 466–85. Caldwell, Charles F. “Government by Caucus: Informal Legislative Groups in an Era of Congressional Reform.” Journal of Law and Politics 5 (1989): 625–55. Calvo-Armengol, Antoni, Thierry A. Verdier, and Yves Zenou. “Strong and Weak Ties in Employment and Crime.” Journal of Public Economics 91 (2007): 203–33. Canon, David T. “Redistricting and the Congressional Black Caucus.” American Politics Quarterly 23 (1995): 159–89. Carey, John M., and Simon Hix. “Electoral Sweet Spot Data.” 2010. Carey data archive: www.dartmouth.edu /~jcarey. Carman, Christopher J. “Norms, Incentives, and Institutions: On the Development of Cross-Party Groups in the Scottish Parliament.” Unpublished manuscript, 2005. Carpenter, Daniel P., Kevin M. Esterling, and David M. J. Lazer. “The Strength of Weak Ties in Lobbying Networks.” Journal of Theoretical Politics 10 (1998): 417–44. Carpenter, Daniel P., Kevin M. Esterling, and David M. J. Lazer. “Information and Contact Making in Policy Networks: A Model with Evidence from the U.S. Health Policy Domain.” Rationality and Society 15 (2003): 411–40. Carpenter, Daniel P., Kevin M. Esterling, and David M. J. Lazer. “Friends, Brokers, and Transitivity: Who Informs Whom in Washington Politics?” Journal of Politics 66 (2004): 224–46. Centola, Damon, and Michael Macy. “Complex Contagions and the Weakness of Long Ties.” American Journal of Sociology 113 (2007): 702–34. Page 247 → Chabanet, Didier. “The Regulation of Interest Groups in the EU.” CONNEX Research Group 4 Work Package A1 Report, October 9, 2006. Cho, Wendy K. Tam, and James H. Fowler. “Legislative Success in a Small World: Social Network Analysis and the Dynamics of Congressional Legislation.” Journal of Politics 72 (2010): 124–35. Chung, Seungwha, Harbir Singh, and Kyungmook Lee. “Complementarity, Status Similarity, and Social Capital as Drivers of Alliance Formation.” Strategic Management Journal 21 (2000): 1–23. Coen, David. “Empirical and Theoretical Studies in EU Lobbying.” Journal of European Public Policy 14 (2007): 333–45. Coen, David, and Jeremy Richardson. Lobbying the European Union: Institutions, Actors, and Issues. Oxford: Oxford University Press, 2009.

Cohen, Lauren, and Christopher Malloy. Friends in High Places. Working Paper 16437. Cambridge, MA: National Bureau of Economic Research, 2010. Cohen, Robert. “Terror Edges Congress Closer to a Virtual World.” Newark (NJ) Star-Ledger, May 20, 2002. Coleman, James S. “Social Capital in the Creation of Human Capital.” American Journal of Sociology 94 (1988): S95–S120. Coleman, James S. Foundations of Social Theory. Cambridge: Harvard University Press, 1990. Commission of the European Communities. European Union Strategy for the Baltic Sea Region. June 10, 2009. ec.europa.eu/regional_policy/sources/docoffic/official/communic/baltic/com_baltic_en.pdf. Constant, David, Lee Sproull, and Sara Kiesler. “The Kindness of Strangers: The Usefulness of Electronic Weak Ties for Technical Advice.” Organization Science 7 (1996): 119–35. Cooper, Joseph. The Origins of Standing Committees and the Development of the Modern House. Houston: Rice University Publications, 1970. Corbett, Richard, Francis Jacobs, and Michael Shackleton. The European Parliament. 7th ed. London: Harper, 2007. Cox, Gary W., and Mathew D. McCubbins. Legislative Leviathan: Party Government in the House. Berkeley: University of California Press, 1993. Cox, Gary W., and Mathew D. McCubbins. Setting the Agenda: Responsible Party Government in the U.S. House of Representatives. Cambridge: Cambridge University Press, 2005. Cranmer, Skyler J., and Bruce A. Desmarais. “Inferential Network Analysis with Exponential Random Graph Models.” Political Analysis 19 (2011): 66–86. Crisp, Brian F., Kristin Kanthak, and Jenny Leijonhufvud. “The Reputations Legislators Build: With Whom Should Representatives Collaborate?” American Political Science Review 98 (2004): 703–16. Crombez, Christophe. “Information, Lobbying, and the Legislative Process in the European Union.” European Union Politics 3 (2002): 7–32. Cross, Rob, and Jonathon N. Cummings. “Tie and Network Correlates of Individual Performance in KnowledgeIntensive Work.” Academy of Management Journal 47 (2004): 928–37. Page 248 → Cross, Rob, and Lee Sproull. “More Than an Answer: Information Relationships for Actionable Knowledge.” Organization Science 15 (2004): 446–62. Davidson, Roger, and Walter Oleszek. Congress and Its Members. Washington, DC: Congressional Quarterly Press, 1981. Deering, Christopher J., and Steven S. Smith. Committees in Congress. 3rd ed. Washington, DC: Congressional Quarterly Press, 1997. de Graaf, Nan Dirk, and Hendrik Derk Flap. “‘With a Little Help from My Friends’: Social Resources as an Explanation of Occupational Status and Income in West Germany, the Netherlands, and the United States.” Social Forces 67 (1988): 452–72. Dilger, Robert Jay. Congressional Member Organizations: Their Purpose and Activities, History, and Formation.

Report R40683. Washington, DC: Congressional Research Service, 2009. Downs, Anthony. An Economic Theory of Democracy. New York: Harper and Row, 1957. Dreier, David. “‘Virtual Congress’ Would Weaken Deliberative Process.” Roll Call, December 20, 2001. Easley, David, and Jon Kleinberg. Networks, Crowds, and Markets: Reasoning about a Highly Connected World. Cambridge: Cambridge University Press, 2010. Ebadi, Yar M., and James M. Utterback. “The Effects of Communication on Technological Innovation.” Management Science 30 (1984): 572–85. Esterling, Kevin M. The Political Economy of Expertise: Information and Efficiency in American National Politics. Ann Arbor: University of Michigan Press, 2004. Eulau, Heinz. “Bases of Authority in Legislative Bodies: A Comparative Analysis.” Administrative Science Quarterly 7 (1962): 309–21. European Centre for Parliamentary Research and Documentation (ECPRD). Request 1075, “Legal Rules Concerning the Founding of Deputies’ Groupings” (2008). European Parliament. Report on a Baltic Sea Region Strategy for the Northern Dimension (2006/2171(INI)).” October 10, 2006. Feiock, Richard C., and John T. Scholz, eds. Self-Organizing Federalism: Collaborative Mechanisms to Mitigate Institutional Collective Action Dilemmas. Cambridge: Cambridge University Press, 2010. Feld, Scott L. “The Focused Organization of Social Ties.” American Journal of Sociology 86 (1981): 1015–35. Feld, Scott L. “Structural Determinants of Similarity among Associates.” American Sociological Review 47 (1982): 797–801. Fenno, Richard. The Power of the Purse: Appropriations Politics in Congress. Boston: Little, Brown, 1966. Fenno, Richard. Congressmen in Committees. Boston: Little, Brown, 1973. Fiellin, Alan. “The Functions of Informal Groups in Legislative Institutions.” Journal of Politics 24 (1962): 72–91. Fish, Steven M., and Matthew Kroenig. The Handbook of National Legislatures: A Global Survey. Cambridge: Cambridge University Press, 2009. Fleming, Lee, Santiago Mingo, and David Chen. “Collaborative Brokerage, Generative Creativity, and Creative Success.” Administrative Science Quarterly 52 (2007): 443–75. Page 249 → Fowler, James H. “Connecting the Congress: A Study of Cosponsorship Networks. Political Analysis 14 (2006a): 456–87. Fowler, James H. “Habitual Voting and Behavioral Turnout.” Journal of Politics 68 (2006b): 335–44. Fowler, James H. “Legislative Cosponsorship Networks in the U.S. House and Senate.” Social Networks 28 (2006c): 454–65. Fowler, James H., and Sangick Jeon. “The Authority of Supreme Court Precedent.” Social Networks 30 (2008):

16–30. Fowler, James H., Timothy R. Johnson, James F. Spriggs II, Sangick Jeon, and Paul J. Wahlbeck. “Network Analysis and the Law: Measuring the Legal Importance of Precedents at the U.S. Supreme Court.” Political Analysis 15 (2007): 324–46. Frank, Ove. “Network Sampling and Model Fitting.” In Models and Methods in Social Network Analysis, edited by Peter J. Carrington, John Scott, and Stanley Wasserman. New York: Cambridge University Press, 2005. Freeman, Linton C. “Centrality in Social Networks: Conceptual Clarification.” Social Networks 1 (1979): 215–39. Friedkin, Noah. “A Test of Structural Features of Granovetter's Strength of Weak Ties Theory.” Social Networks 2 (1980): 411–22. Gais, Thomas L., Mark A. Peterson, and Jack L. Walker. “Interest Groups, Iron Triangles, and Representative Institutions in American National Government.” British Journal of Political Science 14 (2004): 161–86. Galtung, Johan. “A Structural Theory of Imperialism.” Journal of Peace Research 8 (1971): 81–117. Gile, Krista, and Mark S. Handcock. Model-Based Assessment of the Impact of Missing Data on Inference for Networks. Working Paper 66. Seattle: University of Washington Center for Statistics and the Social Sciences, 2006. Gilligan, Thomas W., and Keith Krehbiel. “Collective Decisionmaking and Standing Committees: An Informational Rationale for Restrictive Amendment Procedures.” Journal of Law, Economics, and Organization 3 (1987): 287–335. Gilligan, Thomas W., and Keith Krehbiel. “Asymmetric Information and Legislative Rules with a Heterogeneous Committee.” American Journal of Political Science 33 (1989): 459–90. Gimpel, James G., Frances E. Lee, and Shanna Pearson-Werkowitz. “The Check Is in the Mail: Interdistrict Funding Flows in Congressional Elections.” American Journal of Political Science 52 (2008): 373–94. Glassman, Matthew Eric. Congressional Membership and Appointment Authority to Advisory Commissions, Boards, and Groups.” Report RL 33313, 7-5700. Washington, DC: Congressional Research Service, 2009. Granovetter, Mark. “The Strength of Weak Ties.” American Journal of Sociology 78 (1973): 1360–80. Granovetter, Mark. Getting a Job. Cambridge: Harvard University Press, 1974. Granovetter, Mark. “The Strength of Weak Ties: A Network Theory Revisited.” Sociological Theory 1 (1983): 201–33. Granovetter, Mark. “Economic Action and Social Structure: The Problem of Embeddedness.” American Journal of Sociology 91 (1985): 481–510. Page 250 → Granovetter, Mark. “Problems of Explanation in Economic Sociology.” In Networks and Organization, edited by Nitin Nohria and Robert G. Eccles. Cambridge: Harvard Business School Press, 1992. Granovetter, Mark. “The Impact of Social Structure on Economic Outcomes.” Journal of Economic Perspectives 19 (2005): 33–50. Grant, Wyn. “Pressure Politics: The Changing World of Pressure Groups.” Parliamentary Affairs 57 (2004): 408–19.

Grant, Wyn. “Pressure Groups: A Politics of Collective Consumption?” Parliamentary Affairs 58 (2005): 366–79. Greif, Avner. “Reputation and Coalition in Medieval Trade: Evidence on the Maghribi Traders.” Journal of Economic History 49 (1989): 857–82. Gross, Justin H., and Cosma Shalizi. “Cosponsorship in the U.S. Senate: A Multilevel Approach to Detecting the Subtle Influence of Social Relational Factors on Legislative Behavior.” Unpublished working paper, Department of Statistics, Carnegie Mellon University, 2007. Grossman, Gene M., and Elhanan Helpman. Special Interest Politics. Cambridge: MIT Press, 2001. Grossmann, Matt, and Casey B. K. Dominguez. “Party Coalitions and Interest Group Networks.” American Politics Research 37 (2009): 767–800. Hafner-Burton, Emilie M., Miles Kahler, and Alexander H. Montgomery. “Network Analysis for International Relations.” International Organization 63 (2009): 559–92. Hall, Richard L., and Alan V. Deardorff. “Lobbying as Legislative Subsidy.” American Political Science Review 100 (2006): 69–84. Hammond, Susan Webb. “Congressional Caucuses and Party Leaders in the House of Representatives.” Political Science Quarterly 106 (1991): 277–94. Hammond, Susan Webb. Congressional Caucuses in National Policy Making. Baltimore: Johns Hopkins University Press, 1998. Hammond, Susan Webb, Daniel P. Mulhollan, and Arthur G. Stevens Jr. “Congressional Caucuses: Legislators as Lobbyists.” In Interest Group Politics, edited by Allan Cigler and Burdett A. Loomis. Washington, DC: Congressional Quarterly Press, 1983. Hammond, Susan Webb, Daniel P. Mulhollan, and Arthur G. Stevens Jr. “Informal Congressional Caucuses and Agenda Setting.” Political Research Quarterly 38 (1985): 583–605. Handcock, Mark S., and Krista Gile. Modeling Networks with Sampled or Missing Data. Working Paper 75. Seattle: University of Washington Center for Statistics and the Social Sciences,, 2007. Handcock, Mark S., David R. Hunter, Carter T. Butts, Steven M. Goodreau, Pavel N. Krivitsky, and Martina Morris. “ERGM: A Package to Fit, Simulate and Diagnose Exponential-Family Models for Networks.” Version 2.4-2. http://statnetproject.org; http://CRAN.R-project.org/package=ergm. 2011. Handcock, Mark S., David R. Hunter, Carter T. Butts, Steven M. Goodreau, and Martina Morris. “Statnet: Software Tools for the Statistical Modeling of Network Page 251 → Data.” Version 2.0. http://statnetproject.org; http://CRAN.R-project.org/package=statnet. 2003. Hanneman, Robert A., and Mark Riddle. Introduction to Social Network Methods. Riverside: University of California, Riverside, 2005. http://faculty.ucr.edu/~hanneman/. Hansen, John Mark. “Choosing Sides: The Creation of an Agricultural Policy Network in Congress.” Studies in American Political Development 2 (1987): 183–229. Hansen, John Mark. Gaining Access: Congress and the Farm Lobby, 1919–1981. Chicago: University of Chicago Press, 1991. Hansen, Morten T. “The Search-Transfer Problem: The Role of Weak Ties in Sharing Knowledge across Organization Subunits.” Administrative Science Quarterly 44 (1999): 82–111.

Hansen, Morten T., Marie Louise Mors, and Bjørn Løvås. “Knowledge Sharing in Organizations: Multiple Networks, Multiple Phases.” Academy of Management Journal 48 (2005): 776–93. Hansen, Morten T., Joel M. Podolny, and Jeffrey Pfeffer. “So Many Ties, So Little Time: A Task Contingency Perspective on the Value of Social Capital in Organizations.” Research in the Sociology of Organizations 18 (2000): 21–57. Heaney, Michael T. “Brokering Health Policy: Coalitions, Parties, and Interest Group Influence.” Journal of Health Politics, Policy, and Law 31 (2006): 887–944. Heaney, Michael T., and Scott D. McClurg. “Social Networks and American Politics: Introduction to the Special Issue.” American Politics Research 37 (2009): 727–41. Heaney, Michael T., and Fabio Rojas. “Partisans, Nonpartisans, and the Antiwar Movement in the United States.” American Politics Research 35 (2007): 431–64. Heclo, Hugh. “Issue Networks and the Executive Establishment.” In The New American Political System, edited by Anthony King. Washington, DC: American Enterprise Institute, 1978. Heinz, John P., Edward O. Laumann, Robert H. Salisbury, and Robert L. Nelson. “Inner Circles or Hollow Cores? Elite Networks in National Policy Systems.” Journal of Politics 52 (1990): 356–90. Helmke, Gretchen, and Steven Levitsky. “Informal Institutions and Comparative Politics: A Research Agenda.” Perspectives on Politics 2 (2004): 725–40. Helmke, Gretchen, and Steven Levitsky, eds. Informal Institutions and Democracy: Lessons from Latin America. Baltimore: Johns Hopkins University Press, 2006. Henning, Cecilia, and Mats Lieberg. “Strong Ties or Weak Ties? Neighbourhood Networks in a New Perspective.” Scandinavian Housing and Planning Research 13 (1996): 3–26. Hix, Simon, and Sara Hagemann. “Could Changing the Electoral Rules Fix European Parliament Elections?” Unpublished manuscript, 2008. Hix, Simon, Abdul Noury, and Gerard Roland. Democratic Politics in the European Parliament. New York: Cambridge University Press, 2007. Houghton, Ben. “Residential Mobility, Community Mobility, and Voter Turnout.” Political Behavior 22 (2000): 109–20. Page 252 → “House of Commons: Guide to the Rules on All-Party Groups.” April 2010. http://www.parliament.uk/documents /upload/PCFSGroupsRules.pdf. Høyland, Bjørn, Indraneel Sircar, and Simon Hix. “An Automated Database of the European Parliament.” European Union Politics 10 (2009): 143–52. Huckfeldt, Robert. “Political Participation and the Neighborhood Social Context.” American Journal of Political Science 23 (1979): 579–92. Huckfeldt, Robert. “Social Contexts, Social Networks, and Urban Neighborhoods: Environmental Constraints on Friendship Choice.” American Journal of Sociology 89 (1983): 651–69. Huckfeldt, Robert. “The Social Communication of Political Expertise.” American Journal of Political Science 45 (2001): 425–38.

Huckfeldt, Robert. “Unanimity, Discord, and the Communication of Public Opinion.” American Journal of Political Science 51 (2007): 978–95. Huckfeldt, Robert, Paul Allen Beck, Russell Dalton, and Jeffrey Levine. “Political Environment, Cohesive Social Groups, and the Communication of Public Opinion.” American Journal of Political Science 39 (1995): 1025–54. Huckfeldt, Robert, Paul E. Johnson, and John Sprague. Political Disagreement: The Survival of Diverse Opinions within Communication Networks. New York: Cambridge University Press, 2004. Huckfeldt, Robert, and John Sprague. “Networks in Context: The Social Flow of Political Information.” American Political Science Review 81 (1987): 1197–1216. Hulsheger, Ute R., Neil Anderson, and Jesus F. Salgado. “Team-Level Predictors of Innovation at Work: A Comprehensive Meta-Analysis Spanning Three Decades of Research.” Journal of Applied Psychology 94 (2009): 1128–45. Hunter, David R., Mark S. Handcock, Carter T. Butts, Steven M. Goodreau, and Martina Morris. “ERGM: A Package to Fit, Simulate, and Diagnose Exponential-Family Models for Networks.” Journal of Statistical Software 24 (2008). http://www.jstatsoft.org/v24/i03/paper. Ibarra, Herminia. “Homophily and Differential Returns: Sex Differences in Network Structure and Access in an Advertising Firm.” Administrative Science Quarterly 37 (1992): 422–47. Ibarra, Herminia. “Race, Opportunity, and Diversity of Social Circles in Managerial Networks.” Academy of Management Journal 38 (1995): 673–703. Jackson, Matthew O. Social and Economic Networks. Princeton: Princeton University Press, 2010. Jenkins-Smith, Hank, and Paul Sabatier. “Evaluating the Advocacy Coalition Framework.” Journal of Public Policy 14 (1994): 175–203. Joenniemi, Pertti. The EU Strategy for the Baltic Sea Region: A Catalyst for What? DIIS Brief. Copenhagen: Danish Institute for International Studies, 2009. Johnson, Joel W., and Jessica S. Wallack. 2007. “Database of Electoral Systems and the Personal Vote.” http://hdl.handle.net/1902.1/17901 V1. Judge, David. Backbench Specialisation in the House of Commons. London: Heinemann, 1981. Judge, David. Parliament and Industry. Brookfield, VT: Dartmouth, 1990. Kale, Prashant, Harbir Singh, and Howard Perlmutter. “Learning and Protection of Page 253 → Proprietary Assets in Strategic Alliances: Building Relational Capital.” Strategic Management Journal 21 (2000): 217–37. Kavanaugh, Andrea L., Debbie Denise Reese, John M. Carroll, and Mary Beth Rosson. “Weak Ties in Networked Communities.” Information Society 21 (2005): 119–31. Keller, Amy. “E-Congress: Possible? Yes. Likely? No.” Roll Call, November 5, 2001. Kennedy, Jonathan, and Gabriel Weiman. “The Strength of Weak Terrorist Ties.” Terrorism and Political Violence 23 (2011): 201–12. King, Gary, Michael Tomz, and Jason Wittenberg. “Making the Most of Statistical Analyses: Improving Interpretation and Presentation.” American Journal of Political Science 44 (2000): 347–61. Kingdon, John W. Agendas, Alternatives, and Public Policies. Boston: Little, Brown, 1984.

Kingdon, John W. Congressmen's Voting Decisions. 3rd ed. Cambridge: Harper and Row, 1989. Kirkland, Justin H. “The Relational Determinants of Legislative Outcomes: Strong and Weak Ties between Legislators.” Journal of Politics 73 (2011): 887–98. Klofstad, Casey A. Civic Talk: Peers, Politics, and the Future of Democracy. Philadelphia: Temple University Press, 2011. Knoke, David. Change and Continuity in American Politics: The Social Bases of Political Parties. Baltimore: Johns Hopkins University Press, 1976. Knoke, David, and Song Yang. Social Network Analysis. 2nd ed. New York: Sage, 2008. Koger, Gregory. “Position Taking and Cosponsorship in the U.S. House.” Legislative Studies Quarterly 28 (2003): 225–46. Koger, Gregory, Seth E. Masket, and Hans Noel. “Partisan Webs: Information Exchange and Party Networks.” British Journal of Political Science 39 (2009): 633–53. Kohler-Koch, Beate. “Organized Interests in the EC and European Parliament.” 1997. http://eiop.or.at/eiop/texte /1997-009a.htm. Krackhardt, David. “The Strength of Strong Ties: The Importance of Philos in Organizations.” In Networks and Organizations: Structure, Form, and Action, edited by Nitin Nohria and Robert Eccles. Boston: Harvard Business School Press, 1992. Krackhardt, David, and Robert N. Stern. “Informal Networks and Organizational Crisis: An Experimental Simulation.” Social Psychology Quarterly 51 (1988): 123–40. Krehbiel, Keith. Information and Legislative Organization. Ann Arbor: University of Michigan Press, 1991. Lai, Gina, Nan Lin, and Shu-Yin Leung. “Network Resources, Contact Resources, and Status Attainment.” Social Networks 20 (1998): 159–78. Lake, Robert La Due, and Robert Huckfeldt. “Social Capital, Social Networks, and Political Participation.” Political Psychology 19 (1998): 567–84. Laumann, Edward O., and David Knoke. The Organizational State: Social Choice in National Policy Domains. Madison: University of Wisconsin Press, 1987. Page 254 → Laumann, Edward O., and Franz Urban Pappi. Networks of Collective Action: A Perspective on Community Influence Systems. New York: Academic Press, 1976. Lazarsfeld, Paul Felix, Bernard Berelson, and Hazel Gaudet. The People's Choice. New York: Duell, Sloan, and Pearce, 1944. Lazer, David. “The Free Trade Epidemic in the 1860s and Other Outbreaks of Economic Discrimination.” World Politics 51 (1999): 447–83. Lazer, David. “Networks in Political Science: Back to the Future.” PS: Political Science and Politics 44 (2011): 61–68. Leana, Carrie R., and Harry J. Van Buren III. “Organizational Social Capital and Employment Practices.” Academy of Management Review 24 (1999): 538–55.

Levin, David Z., and Rob Cross. “The Strength of Weak Ties You Can Trust: The Mediating Role of Trust in Effective Knowledge Transfer.” Management Science 50 (2004): 1477–90. Levy, Arthur B., and Susan Stoudinger. “Sources of Voting Cues for the Congressional Black Caucus.” Journal of Black Studies 7 (1976): 29–46. Library of Congress. “Hispanic Americans in Congress, 1822–1995.” 2010. http://www.loc.gov/rr/hispanic /congress/chron.html. Library of Congress. “Legislative Data, Bills and Resolutions, 111th Congress.” 2011. http://thomas.loc.gov/home /LegislativeData.php?&n=Browse&c=111. Lin, Nan. “Social Resources and Instrumental Action.” In Structure and Network Analysis, edited by Peter Marsden and Nan Lin. Beverly Hills, CA: Sage, 1982. Lin, Nan. “Social Resources and Social Mobility: A Structural Theory of Status Attainment.” In Social Mobility and Social Structure, edited by Ronald Breiger. Cambridge: Cambridge University Press, 1988. Lin, Nan, Walter M. Ensel, and John C. Vaughn. “Social Resources and Strength of Ties: Structural Factors in Occupational Status Attainment.” American Sociological Review 46 (1981): 393–405. Long, J. Scott. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage, 1997. Loomis, Burdett A. “Congressional Caucuses and the Politics of Representation.” In Congress Revisited, 2nd ed., edited by Lawrence C. Dodd and Bruce I. Oppenheimer. Washington, DC: Congressional Quarterly Press, 1981. Love, Alice A. “Already, LSOs Start to Go Private Route.” Roll Call, December 12, 1994. Love, Alice A. “Like Other Defunct LSOs. Black Caucus Forms CMO.” Roll Call, March 2, 1995. Lupia, Arthur, and Mathew D. McCubbins. The Democratic Dilemma: Can Citizens Learn What They Need to Know? Cambridge: Cambridge University Press, 1998. Mahoney, Christine. Brussels versus the Beltway: Advocacy in the United States and the European Union. Washington, DC: Georgetown University Press, 2008. Maloney, William A., Grant Jordan, and Andrew M. McLaughlin. “Interest Groups and Public Policy: The Insider /Outsider Model Revisited.” Journal of Public Policy 14 (1994): 17–38. Page 255 → Mann, Thomas, and Norman Ornstein. The Broken Branch: How Congress Is Failing America and How to Get It Back on Track. Oxford: Oxford University Press, 2006. Mann, Thomas, and Marco Reuter. “Die außenpolitische Funktion der Tibet Intergroup.” In Das Europäische Parlament: Institution, Vision, und Wirklichkeit, edited by Doris Dialer, Eva Lichtenberger, and Heinrich Neisser. Innsbruck: Innsbruck University Press, 2010. Manning, Jennifer E., and Colleen J. Shogan. Women in the United States Congress: 1917–2009. Report RL30261. Washington, DC: Congressional Research Service, 2009. Manning, Jennifer E., and Colleen J. Shogan. African American Members of the United States Congress: 1870–2009.” Report RL30378. Washington, DC: Congressional Research Service, 2010. Maoz, Zeev. Networks of Nations: The Evolution, Structure, and Impact of International Networks, 1816–2001.

Cambridge: Cambridge University Press, 2010. Marsden, Peter V. “Network Data and Measurement.” Annual Review of Sociology 16 (1990): 435–63. Marsden, Peter V., and Karen E. Campbell. “Measuring Tie Strength.” Social Forces 63 (1984): 482–501. Marsden, Peter V., and Jeanne S. Hurlbert. “Social Resources and Mobility Outcomes: A Replication and Extension.” Social Forces 66 (1988): 1038–59. Marshall, David. “Who to Lobby and When: Institutional Determinants of Interest Group Strategies in European Parliament Committees.” European Union Politics 11 (2010): 553–75. Masket, Seth E. “Where You Sit Is Where You Stand: The Impact of Seating Proximity on Legislative CueTaking.” Quarterly Journal of Political Science 3 (2008): 301–11. Matthews, Donald. U.S. Senators and Their World. Chapel Hill: University of North Carolina Press, 1960. McClurg, Scott D. “Social Networks and Political Participation: The Role of Social Interaction in Explaining Political Participation.” Political Research Quarterly 56 (2003): 449–64. McClurg, Scott D. “The Electoral Relevance of Political Talk: Examining Disagreement and Expertise Effects in Social Networks on Political Participation.” American Journal of Political Science 50 (2006): 737–54. McCormick, James M., and Neil J. Mitchell. “Commitments, Transnational Interests, and Congress: Who Joins the Congressional Human Rights Caucus?” Political Research Quarterly 60 (2007): 579–92. McFadyen, M. Ann, Matthew Semadeni, and Albert A. Cannella. “Value of Strong Ties to Disconnected Others: Examining Knowledge Creation in Biomedicine.” Organization Science 20 (2009): 552–64. McPherson, J. Miller. “An Ecology of Affiliation.” American Sociological Review 48 (1983): 519–32. McPherson, J. Miller, Pamela Popielarz, and Sonja Drobnic. “Social Networks and Organizational Dynamics.” American Sociological Review 57 (1992): 153–70. Page 256 → McPherson, J. Miller, and James Ranger-Moore. “Evolution on a Dancing Landscape: Organizations and Networks in Dynamic Blau Space.” Social Forces 70 (1991): 19–42. McPherson, J. Miller, and Lynn Smith-Lovin. “Sex Segregation in Voluntary Associations.” American Sociological Review 51 (1986): 61–79. McPherson, J. Miller, and Lynn Smith-Lovin. “Homophily in Voluntary Organizations: Status Distance and the Composition of Face-to-Face Groups.” American Sociological Review 52 (1987): 370–79. Milbrath, Lester. The Washington Lobbyists. Chicago: Rand McNally, 1963. Miler, Kristina. “The Constituency Motivations of Caucus Membership.” American Politics Research 39 (2011): 885–920. Miller, Cheryl M. “Agenda-Setting by State Legislative Black Caucuses: Policy Priorities and Factors of Success.” Policy Studies Review 9 (1990): 339–54. Mooney, Christopher Z. “Information Sources in State Legislative Decision Making.” Legislative Studies Quarterly 16 (1991): 445–55.

Moran, Peter. “Structural vs. Relational Embeddedness: Social Capital and Managerial Performance.” Strategic Management Journal 26 (2005): 1129–51. Morgan, Stephen L., and Aage B. Soerensen. “Parental Networks, Social Closure, and Mathematics Learning: A Test of Coleman's Social Capital Explanation of School Effects.” American Sociological Review 64 (1999): 661–81. Morrison, Elizabeth W. “Newcomers’ Relationships: The Role of Social Network Ties during Socialization.” Academy of Management Journal 45 (2002): 1149–60. Mutz, Diana C. “The Consequences of Cross-Cutting Networks for Political Participation.” American Journal of Political Science 46 (2002a): 838–55. Mutz, Diana C. “Cross-Cutting Social Networks: Testing Democratic Theory in Practice.” American Political Science Review 96 (2002b): 111–26. Mutz, Diana C. Hearing the Other Side: Deliberative versus Participatory Democracy. New York: Cambridge University Press, 2006. Mutz, Diana C., and Paul S. Martin. 2001. “Facilitating Communication across Lines of Political Difference: The Role of Mass Media.” American Political Science Review 95 (2001): 97–114. Nahapiet, Janine, and Sumantra Ghoshal. “Social Capital, Intellectual Capital, and the Organizational Advantage.” Academy of Management Review 23 (1998): 242–66. Navarrete Yánez, Bernardo. Los Problemas de Articulacion de Intereses en las Regionenes de Chile (1989–2009). Paper presented at “Sentando Más Actores a la Mesa II: Nuevos Actores para la Descentralización,” January 7–8, 2010, Programa Ciudadanía y Gestión Pública Universidad de Los Lagos.http://www.innovacionciudadana.cl. Nelson, Garrison. Committees in the U.S. Congress, 1947–1992. Vol. 1, Committee Jurisdictions and Member Rosters. Washington D.C.: Congressional Quarterly Press, 1993. Nelson, Garrison. N.d. “Committees in the U.S. Congress, 1947–1992, House Committees Page 257 → 80th–97th Congresses and House Committees 98th–102nd Congresses.” http://web.mit.edu/~17.251/www/data_page.html#1. Nickerson, David W. “Is Voting Contagious? Evidence from Two Field Experiments.” American Political Science Review 102 (2008): 49–57. Obstfeld, David. “Social Networks, the Tertius Lunges Orientation, and Involvement.” Administrative Science Quarterly 50 (2005): 100–130. Oleszek, Walter J. Congressional Procedures and the Policy Process. 7th ed. Washington, DC: Congressional Quarterly Press, 2007. Patacchini, Eleonora, and Yeves Zenou. “The Strength of Weak Ties in Crime.” European Economic Review 52 (2008): 209–36. Patterson, Samuel C. “Patterns of Interpersonal Relations in a State Legislative Group: The Wisconsin Assembly.” Public Opinion Quarterly 23 (1959): 101–9. Peoples, Clayton D. “Interlegislator Relations and Policy Making: A Sociological Study of Roll-Call Voting in a State Legislature.” Sociological Forum 23 (2008): 455–80. Peoples, Clayton D. “Contributor Influence in Congress: Social Ties and PAC Effects on U.S. House Policymaking.” Sociological Quarterly 51 (2010): 649–77.

Perry-Smith, Jill E. “Social Yet Creative: The Role of Social Relationships in Facilitating Individual Creativity.” Academy of Management Journal 49 (2006): 85–101. Perry-Smith, Jill E., and Christina E. Shalley. “The Social Side of Creativity: A Static and Dynamic Social Network Perspective.” Academy of Management Review 28 (2003): 89–106. Phelps, Corey, Ralph Heidl, and Anu Wadhwa. “Knowledge, Networks, and Knowledge Networks: A Review and Research Agenda.” Journal of Management. 2012. http://jom.sagepub.com/content/early/2012/01/04 /0149206311432640.abstract. Podolny, Joel M., and James N. Baron. “Resources and Relationships: Social Networks and Mobility in the Workplace.” American Sociological Review 62 (1997): 673–93. “Polity IV Project: Political Regime Characteristics and Transitions, 1800–2009.” N.d. Monty G. Marshall, Director, Monty G. Marshall and Keith Jaggers, Principal Investigators. Center for Systemic Peace and Colorado State University, Ted Robert Gurr, Founder, University of Maryland. http://www.systemicpeace.org/polity /polity4.htm. Poole, Keith T., and Howard Rosenthal. Congress: A Political-Economic History of Roll Call Voting. Oxford: Oxford University Press, 1997. Poole, Keith T., and Howard Rosenthal. Ideology and Congress. New Brunswick, NJ: Transaction, 2007. Poole, Keith T., Howard Rosenthal, Royce Carroll, Jeff Lewis, James Lo, and Nolan McCarty. “DW NOMINATE Scores with Bootstrapped Standard Errors.” 2011. voteview.com/dwnominate.asp. Popielarz, Pamela, and J. Miller McPherson. “On the Edge or In Between: Niche Position, Niche Overlap, and the Duration of Voluntary Memberships.” American Journal of Sociology 101 (1995): 698–720. Page 258 → Porter, Mason A., Peter J. Mucha, M. E. J. Newman, and A. J. Friend. “Community Structure in the United States House of Representatives.” Physica A: Statistical Mechanics and Its Applications 386 (2007): 414–38. Porter, Mason A., Peter J. Mucha, M. E. J. Newman, and Casey M. Warmbrand. “A Network Analysis of Committees in the U.S. House of Representatives.” Proceedings of the National Academy of Sciences of the United States of America 120 (2005): 7057–62. Portes, Alejandro. “Social Capital: Its Origins and Applications in Modern Technology.” Annual Review of Sociology 24 (1998): 1–24. Potters, Jan, and Franz Van Winden. “Modelling Political Pressure as Transmission of Information.” European Journal of Political Economy 6 (1990): 61–88. Radnitz, Scott. “Informal Politics and the State.” Comparative Politics 43 (2011): 351–71. Reagans, Ray, and Bill McEvily. “Network Structure and Knowledge Transfer: The Effects of Cohesion and Range.” Administrative Science Quarterly 48 (2003): 240–67. Reagans, Ray, and Ezra W. Zuckerman. “Networks, Diversity and Productivity: The Social Capital of Corporate R&D Teams.” Organization Science 12 (2001): 502–17. Reh, Christine, Adrienne Heritier, Edoardo Bressanelli, and Christel Koop. “The Informal Politics of Legislation: Explaining Secluded Decision Making in the European Union.” Comparative Political Studies (2013) (forthcoming).

Ricci, David M. The Transformation of American Politics: The New Washington and the Rise of Think Tanks. New Haven: Yale University Press, 1993. Richardson, Sula P. Caucuses and Legislative Service Organizations of the 103rd Congress: An Informational Directory. Washington, DC: Congressional Research Service, 1993. Ringe, Nils. Who Decides, and How? Preferences, Uncertainty, and Policy Choice in the European Parliament. Oxford: Oxford University Press, 2010. Ringe, Nils, Jennifer N. Victor, and Justin H. Gross. “Keeping Your Friends Close and Your Enemies Closer: Information Networks in Legislative Politics.” British Journal of Political Science 43, no. 3 (2013): 601–28. Ripley, Randall. Congress: Process and Policy. New York: Norton, 1983. Rodan, Simon, and Charles Galunic. “More Than Network Structure: How Knowledge Heterogeneity Influences Managerial Performance and Innovativeness.” Strategic Management Journal 25 (2004): 541–56. Rodriguez, Paul M. 1993. “Groups Don't Play by Congress’ Rules.” Washington Times, June 30, 1993. Rogers, Everett M. Diffusion of Innovations. 4th ed. New York: Free Press, 1995. Rogowski, Jon C., and Betsy Sinclair. “Estimating the Causal Effects of Social Interaction with Endogenous Networks.” Working paper, University of Chicago, 2011. Rohde, David. Parties and Leaders in the Postreform House. Chicago: University of Chicago Press, 1991. Page 259 → Rolfe, Meredith. Voting Together: A Social Theory of Voting. Cambridge: Cambridge University Press, 2012. Rosenthal, Elizabeth A. “Social Networks and Team Performance.” Ph.D. diss., University of Chicago, 1996. Routt, Garland C. “Interpersonal Relationships and the Legislative Process.” Annals of the American Academy of Political and Social Science 195 (1938): 129–36. Royce, Ed. “Royce Remarks on Release of Africa Policy Advisory Report, ‘Rising U.S. Stakes in Africa,’” July 8, 2004. http://royce.house.gov/News/DocumentSingle.aspx?DocumentID=16453. Sabatier, Paul. “An Advocacy Coalition Model of Policy Change and the Role of Policy-Oriented Learning Therein.” Policy Sciences 21 (1988): 129–68. Schilling, Melissa A., and Corey C. Phelps. “Interfirm Collaboration Networks: The Impact of Large-Scale Network Structure on Firm Innovation.” Management Science 53 (2007): 1113–26. Schlozman, Kay Lehman, and John T. Tierney. Organized Interests and American Democracy. New York: Harper and Row, 1986. Schneider, Mark, John Scholz, Mark Lubell, Denisa Mindruta, and Matthew Edwardsen. “Building Consensual Institutions: Networks and the National Estuary Program.” American Journal of Political Science 47 (2003): 143–58. Scholz, John T., Ramiro Berardo, and Brad Kile. “Do Networks Solve Collective Action Problems? Credibility, Search, and Collaboration.” Journal of Politics 70 (2008): 393–406. Scholz, John T., and Cheng-Lung Wang. “Cooptation or Transformation? Local Policy Networks and Federal Regulatory Enforcement.” American Journal of Political Science 50 (2006): 81–97.

Schulz, Martin. “Pathways of Relevance: Exploring Inflows of Knowledge into Subunits of Multinational Corporations.” Organization Science 14 (2003): 440–59. Scott, Susanne G., and Reginald A. Bruce. “Determinants of Innovative Behavior: A Path Model of Individual Innovation in the Workplace.” Academy of Management Journal 37 (1994): 580–607. Seibert, Scott E., Maria L. Kraimer, and Robert C. Liden. “A Social Capital Theory of Career Success.” Academy of Management Journal 44 (2001): 219–37. Shepsle, Kenneth A., and Mark S. Bonchek. Analyzing Politics: Rationality, Behavior, and Institutions. New York: Norton, 1997. Sherlock, Stephen. “Indonesia's Regional Assembly: Democracy, Representation, and the Regions: A Report on the Dewan Perwakilan Daerah (DPD).” 2005. http://www.cdi.anu.edu.au/. Siegel, David A. “Social Networks in Comparative Perspective.” PS: Political Science and Politics 44 (2011): 51–54. Sinclair, Betsy. The Social Citizen Peer Networks and Political Behavior. Chicago: University of Chicago Press, 2012. Singh, Jasjit. “Collaborative Networks as Determinants of Knowledge Diffusion Patterns.” Management Science 51 (2005): 756–70. Page 260 → Singh, Robert. “The Rise and Fall of Legislative Service Organisations in the United States Congress.” Journal of Legislative Studies 2 (1996): 79–102. Stevens, Arthur G., Jr., Arthur H. Miller, and Thomas E. Mann. “Mobilization of Liberal Strength in the House, 1955–1970: The Democratic Study Group.” American Political Science Review 68 (1974): 667–81. Stevens, Arthur G., Jr., Daniel P. Mulhollan, and Paul S. Rundquist. “U.S. Congressional Structure and Representation: The Role of Informal Groups.” Legislative Studies Quarterly 6 (1981): 415–37. Stewart, Charles, III. Analyzing Congress. New York: Norton, 2001. Stewart, Charles, III, and Jonathan Woon. 2009. “Congressional Committee Assignments, 103rd to 111th Congresses, 1993–2009: House Committee Assignment Data.” 2009. http://web.mit.edu/~17.251/www /data_page.html#1. Swers, Michele L. “Connecting Descriptive and Substantive Representation: An Analysis of Sex Differences in Cosponsorship Activity.” Legislative Studies Quarterly 30 (2005): 407–33. Szulanski, Gabriel. “Exploring Internal Stickiness: Impediments to the Transfer of Best Practice within the Firm.” Strategic Management Journal 17 (1996): 27–43. Tanasescu, Irina. The European Commission and Interest Groups: Towards a Deliberative Interpretation of Stakeholder Involvement in EU Policy-Making. Brussels: ASP-VUB Press, 2009. Thornton, Patricia H. “The Sociology of Entrepreneurship.” Annual Review of Sociology 25 (1999): 19–56. Tomz, Michael, Jason Wittenberg, and Gary King. “CLARIFY: Software for Interpreting and Presenting Statistical Results.” Journal of Statistical Software 8 (2003): 1–30. Tong, Lorraine H. Asian Pacific Americans in the United States Congress. Report 97-398. Washington, DC:

Congressional Research Service, 2010. Tsai, Lily L. Accountability without Democracy: Solidary Groups and Public Goods Provision in Rural China. New York: Cambridge University Press, 2007. Tsai, Wenpin. “Knowledge Transfer in Intraorganizational Networks: Effects of Network Position and Absorptive Capacity on Business Unit Innovation and Performance.” Academy of Management Journal 44 (2001): 996–1004. Tsai, Wenpin, and Sumantra Ghoshal. “Social Capital and Value Creation: The Role of Intrafirm Networks.” Academy of Management Journal 41 (1998): 464–73. Tullock, Gordon. “Adam Smith and the Prisoners’ Dilemma.” Quarterly Journal of Economics 100 (1985): 1073–81. U.S. House of Representatives. Committee on House Administration. Ad Hoc Subcommittee on Legislative Service Organizations. Legislative Service Organizations: Report of the Ad Hoc Subcommittee on Legislative Service Organizations of the Committee on House Administration, House of Representatives. 97th Cong., 2nd sess., June 24, 1982. Washington, DC: U.S. Government Printing Office, 1982. U.S. House of Representatives. Office of the Clerk. “Election Information: Election Statistics.” 2010a. http://clerk.house.gov/member_info/electionInfo/index.html (files for 2004, 2006, 2008). Page 261 → U.S. House of Representatives. Office of the Clerk. “Leadership Information. | Historical Leadership Information.” 2010b. http://clerk.house.gov/member_info/leadership.html. Uslaner, Eric. The Decline of Comity in Congress. Ann Arbor: University of Michigan Press, 1993. Uzzi, Brian. “The Sources and Consequences of Embeddedness for Economic Performance of Organizations: The Network Effect.” American Sociological Review 61 (1996): 674–98. Uzzi, Brian. “Social Structure and Competition in Interfirm Networks: The Paradox of Embeddedness.” Administrative Science Quarterly 42 (1997): 35–47. Uzzi, Brian, and Ryon Lancaster. “Relational Embeddedness and Learning: The Case of Bank Loan Managers and Their Clients.” Management Science 49 (2003): 383–99. Uzzi, Brian, and Jarrett Spiro. “Collaboration and Creativity: The Small World Problem.” American Journal of Sociology 111 (2005): 447–504. Van Schendelen, Rinus. Machiavelli in Brussels: The Art of Lobbying the EU. Amsterdam: Amsterdam University Press, 2005. Vega, Arturo. “Congressional Informal Groups as Representative Responsiveness.” American Review of Politics 14 (1993): 355–73. Victor, Jennifer N., and Gregory Koger. “The Strength of Lobbying Ties: The Network of Lobbyists and Legislators.” Unpublished manuscript, 2011. Victor, Jennifer N., and Nils Ringe. “The Social Utility of Informal Institutions: Caucuses as Networks in the 110th U.S. House of Representatives.” American Politics Research 37 (2009): 742–66. Wallerstein, Immanuel M. The Modern World-System. Vol. 1, Capitalist Agriculture and the Origins of the European World-Economy in the Sixteenth Century. New York: Academic Press, 1974.

Ward, Michael D., Katherine Stovel, and Audrey Sacks. “Network Analysis and Political Science.” Annual Review of Political Science 14 (2011): 245–64. Wasserman, Stanley, and Katherine Faust. Social Network Analysis: Methods and Applications. New York: Cambridge University Press, 1994. Watts, Duncan J. “Networks, Dynamic, and the Small World Phenomenon.” American Journal of Sociology 105 (1999): 493–527. Watts, Duncan J., and Steven H. Strogatz. “Collective Dynamics of ‘Small-World’ Networks.” Nature 393 (1998): 440–42. Whiteman, David. Communication in Congress: Members, Staff, and the Search for Information. Lawrence: University Press of Kansas, 1995. Wolfe, Eugene L. “Japanese Electoral and Political Reform: Role of the Young Turks.” Asian Survey 35 (1995): 1059–74. Wollebaek, Dag, and Per Selle. 2002. “Does Participation in Voluntary Associations Contribute to Social Capital? The Impact of Intensity, Scope, and Type.” Nonprofit and Voluntary Sector Quarterly 31 (2002): 32–61. Wright, John R. Interest Groups and Congress: Lobbying, Contributions, and Influence. Boston: Allyn and Bacon, 1996. Yakubovich, Valery. “Weak Ties, Information, and Influence: How Workers Find Page 262 → Jobs in a Local Russian Labor Market.” American Sociological Review 70 (2005): 408–21. Yeoh, Kok Kheng. 2001. Toward an Index of Ethnic Fractionalization. FEA Working Paper 2001-3. cmsad.um.edu.my/images/fep/doc/2001%20Pdf/FEA-WP-2001-003.pdf. Young, James Sterling. The Washington Community, 1800–1828. New York: Columbia University Press, 1966. Zenger, Todd R., and Barbara S. Lawrence. 1989. “Organizational Demography: The Differential Effects of Age and Tenure Distributions on Technical Communication.” Academy of Management Journal 32 (1989): 353–76. Websites All-Party Groups in the United Kingdom Parliament http://www.parliament.uk/about/mps-and-lords/members/apg/ Centre for Legislative Research and Advocacy: Parliamentary Forums http://www.clraindia.org/parliamentary_forum.shtml EUFORES http://www.eufores.org European Commission: Services of General Economic Interest (SGEI) http://ec.europa.eu/competition/state_aid/legislation/sgei.html Frentes Parlamentares in the Brazilian Câmara dos Deputados http://www.camara.gov.br/internet/deputado/frentes.asp

GLOBE Europe http://www.globe-europe.eu Kangaroo Group http://www.kangaroogroup.eu Intergroups in the European Parliament http://www.europarl.europa.eu/aboutparliament/en/00c9d93c87/Intergroups.html LEFT Caucus http://leftcaucus.eu Les groupes d'études in the French Assemblée Nationale http://www.assemblee-nationale.fr/connaissance/etudes.asp Lobbies in the Israeli Knesset http://www.knesset.gov.il/lexicon/eng/lobby_eng.htm Parlamentarische Gruppen in the Swiss Bundesversammlung http://www.parlament.ch/d/organe-mitglieder/bundesversammlung/adressen-websites-fotos/Documents /parlamentarische-gruppen.pdf Parlamentinės grupės in the Lithanian Seimas http://www3.lrs.lt/pls/inter/w5_show?p_r=6117&p_k=2 Parliamentary Forums in the Parliament of India http://164.100.47.134/committee/Cells/FAQs.pdf Register of All-Party Groups in the United Kingdom Parliament http://www.publications.parliament.uk/pa/cm/cmallparty/memi01.htm Page 263 → The Spinelli Group http://www.spinelligroup.eu Ühendused in the Estonian Riigikogu http://www.riigikogu.ee/index.php?id=34605 United States Senate: Résumé of Congressional Activity http://www.senate.gov/pagelayout/reference/two_column_table/Resumes.htm USAID to Organize a U.S. Study Tour for Bulgarian Members of Parliament http://bulgaria.usaid.gov/1/news_item.html

Zespoły in the Polish Sejm http://orka.sejm.gov.pl/SQL.nsf/listazesp?OpenAgent&6 Page 264 →

Page 265 →

INDEX access points, LMOs as, 9, 11, 22, 44, 202, 208, 211 Ad Hoc Subcommittee on Legislative Service Organizations (U.S. Congress), 96–97 Affordable Care Act (U.S.), 188, 193 See also Pro-life Caucus AGE Platform, 159–60, 171 See also Intergroup on Ageing and Intergenerational Solidarity agenda setting. See under LMO activities Ainsworth, Scott H., 9, 20–22, 24, 30, 42, 44, 47, 218 Alliance of Liberals and Democrats for Europe (ALDE) (EP), 82, 85–86, 91–92, 101–3, 129, 131–32, 134, 195, 231 American Recovery and Reinvestment Act (U.S.), 188, 193 See also Congressional Steel Caucus Amnesty International, 64, 165 See also Tom Lantos Human Rights Commission Animal Welfare Intergroup (EP), 159, 161, 169–71, 178, 191–93, 230, 240 See also Eurogroup for Animals ATD Fourth World, 161 See also Intergroup on Extreme Poverty and Human Rights Audy, Jean-Pierre, 91, 129, 131–32 Australia. See under LMOs in Baltic Europe Intergroup (EP), 88, 187–89 See also Baltic Sea Strategy for the Northern Dimension; European Union Strategy for the Baltic Sea Region Baltic Sea Strategy for the Northern Dimension (EU), 187–89 See also Baltic Europe Intergroup; European Union Strategy for the Baltic Sea Region bargaining, 2, 7, 13, 47, 83, 117–19, 169, 191, 202, 206, 212 Bates, Jim, 96 Bauer, Edit, 91, 129, 131–32, 134 Belgium. See under LMOs in

bicameral (legislature), 20, 54–57, 82, 92, 94, 224 See also unicameral (legislature) Bike Caucus. See Congressional Bike Caucus Biomedical Research Caucus (U.S. Congress), 3, 107 Bipartisan Congressional Task Force on Alzheimer's Disease (U.S. Congress), 80, 108 bipartisanship. See cross-partisanship Black Caucus. See Congressional Black Caucus Blue Dog Democrats (U.S. Congress), 66, 99–100 Brady, Kevin, 139–40, 232 Brazil. See under LMOs in bridging ability. See bridgingness bridging connections. See bridgingness bridging nature. See bridgingness bridging power. See bridgingness bridging ties. See bridgingness bridgingness (of LMO ties), 5–6, 11, 16, 32–33, 36–37, 39–40, 46, 48, 76, 112, 125, 135–37, 139, 141–42, 145–46, 201, 210, 219, 226 brokerage, 36, 135, 220 Brown, Edward, 139–40, 232 Bulgaria, 68–69, 223 Bureau (of EP, EP party groups, or intergroups), 86, 117–18, 122–23, 136, 149–52, 157, 160, 171, 183, 227, 232 Burt, Ronald S., 5, 18–19, 32, 35–36, 76, 113, 141, 144–45, 219–20 See also under network measures Bushill-Matthews, Philip, 91, 150–51 Page 266 → “Buy American” amendment (U.S.). See American Recovery and Reinvestment Act Buzek, Jerzy, 87, 225 campaign contributions, 8, 23 Canada. See under LMOs in Capitol Hill, 93, 96, 116, 154, 166

caucus Biomedical Research Caucus, 3, 107 bipartisan caucus (see cross-partisanship) Bipartisan Congressional Task Force on Alzheimer's Disease, 80, 108 Chowder and Marching Society, 94 Congressional Arts Caucus, 3, 107–8 Congressional Bike Caucus, 79, 108, 154, 165 Congressional Black Caucus, 3, 94, 96–98, 107, 124 Congressional Brain Injury Task Force, 108, 165 Congressional Caucus for Women's Issues, 3, 94–95, 107 Congressional Caucus on India and Indian Americans, 107–8, 124, 165 Congressional Caucus on Parkinson's Disease, 31, 165, 193 Congressional Caucus to Combat Alzheimer's Disease, 80 Congressional Coalition on Adoption, 16, 107–8, 165, 186 Congressional Coalition on Population and Development, 107 Congressional Diabetes Caucus, 108, 165, 186–87 Congressional Fire Services Caucus, 107–8, 153, 165 Congressional Hispanic Caucus, 3, 94, 107, 124 Congressional Human Rights Caucus, 16, 107–8 congressional member organizations (CMOs), 75, 78, 95, 97 Congressional Progressive Caucus, 107, 124 Congressional Steel Caucus, 16, 107, 165, 188, 193 Congressional Traumatic Brain Injury Caucus, 154, 188, 193 Congressional Western Caucus, 139 Democratic Study Group, 94, 97 House Wednesday Group, 94 International Conservation Caucus, 61, 153, 165, 188, 193 Korean Caucus, 186–87 Members of Congress for Peace through Law, 94

Northeast-Midwest Congressional Coalition, 80 Older Americans Caucus, 3, 107 party-unbalanced caucus, 206 Pro-Life Caucus, 61, 188, 193 Rural Health Care Coalition, 3, 107 single-party caucus, 99–100, 226, 232 See also legislative service organization centrality. See under network measures cheap talk, 203 See also under legislative member organizations Chile. See under LMOs in Chowder and Marching Society (U.S. Congress), 94 coalition building. See under LMO activities codecision procedure. See ordinary legislative procedure Coen, David, 43–44, 84, 221 collective action dilemma/problem, 3–4, 12, 18, 24, 29, 43, 47, 76, 121, 156, 184, 201, 209 committee, 2–7, 8–9, 11–13, 16, 18–19, 21, 23, 25–31, 34, 37–38, 40–41, 46–48, 61, 64, 71, 75–77, 101, 109, 112–13, 115, 125, 128–29, 142, 144–47, 185, 192, 203–5, 208–9, 211–12, 216, 218–19, 223, 230 committee assignments, 23, 27, 30, 40, 78, 112, 128, 219, 228 committee jurisdiction, 4, 19, 28, 46, 76, 119 committee network, 27, 128–29, 231–32 committee staff, 25, 175, 188, 223 in Congress, 1, 10, 12, 16, 21, 75, 77–78, 93–97, 100, 103–5, 116, 120, 124, 128–29, 138–39, 144–45, 155–56, 165–67, 175, 188, 194, 198–200, 225–26 in the EP, 77–78, 83, 85, 88, 101, 111, 114, 117–19, 123, 128–29, 136, 139, 143, 145, 149–51, 160–61, 170–71, 187–88, 195–96, 227, 231–32, 239 cross-committee ties, 205 See also cross-party ties Committee on House Administration (U.S. Congress), 78, 95–97, 100, 225–26 Page 267 → Conference of Committee Chairs (EP), 136

Conference of Delegation Chairs (EP), 136 Conference of Presidents (EP), 78, 86, 136, 225, 231–32, 239 Congo Basin Forest Partnership Act (U.S.), 188, 193 See also International Conservation Caucus Congress Democratic Party, 93, 121 members of Congress, 1, 8, 10, 20, 75, 80, 93–94, 96, 98, 100, 103, 105, 109, 124, 126–28, 134, 139, 144, 164, 166, 173, 175, 181–82, 200, 219, 225–26, 232 Republican Party, 93, 96, 121 Senate, 93, 166, 188, 206, 224, 238–40 virtual Congress, 1 Congressional Arts Caucus (U.S. Congress), 3, 107–8 Congressional Bike Caucus (U.S. Congress), 79, 108, 154, 165 Bike to Work Week, 154 Congressional Black Caucus (U.S. Congress), 3, 94, 96–98, 107, 124 Congressional Black Caucus Foundation (U.S. Congress), 96, 98 Congressional Brain Injury Task Force (U.S. Congress), 108, 165 Congressional Caucus for Women's Issues (U.S. Congress), 3, 94–95, 107 Congressional Caucus on India and Indian Americans (U.S. Congress), 107–8, 124, 165 Congressional Caucus on Parkinson's Disease (U.S. Congress), 31, 165, 193 Congressional Caucus to Combat Alzheimer's Disease (U.S. Congress), 80 Congressional Coalition on Adoption (U.S. Congress), 16, 107–8, 165, 186 Congressional Coalition on Population and Development (U.S. Congress), 107 Congressional Diabetes Caucus (U.S. Congress), 108, 165, 186–87 Congressional Fire Services Caucus (U.S. Congress), 107–8, 153, 165 Congressional Fire Services Institute (U.S. Congress), 96, 98, 173, 186 Congressional Hispanic Caucus (U.S. Congress), 3, 94, 107, 124 Congressional Human Rights Caucus (U.S. Congress), 16, 107–8 congressional member organizations (CMOs). See caucus(U.S. Congressional Progressive Caucus (U.S. Congress), 107, 124

Congressional Research Service (U.S. Congress), 175 Congressional Sportsmen's Caucus Foundation (U.S. Congress), 98 Congressional Steel Caucus (U.S. Congress), 16, 107, 165, 188, 193 Congressional Traumatic Brain Injury Caucus (U.S. Congress), 154, 188, 193 Congressional Western Caucus (U.S. Congress), 139 Congressional Yellow Book (U.S.), 78–79 consensus building. See under LMO activities constituency service, 4, 9, 20–21, 25, 30, 36, 40, 49, 56–57, 64, 69, 74, 77, 84, 92–93, 95, 101, 103, 109, 127, 152, 155, 157, 161, 167–68, 173–74, 181–83, 202–3, 209, 218, 221, 224 constituent organization (see outside advocate) constraint (Burt's). See under network measures Conyers, John, 144 coordination dilemma/problem, 2 coordination (function of LMOs). See legislative member organizations Corbett, Richard, 83–84, 86, 123, 240 Council of Ministers (EU), 81–82, 225 cross-partisanship, 72 Congress, 13, 48, 120–21, 183, 206–7, 211–12, 238 EP, 13, 49, 117–19, 182, 206, 211–12 cross-party ties, 8, 11, 19, 33–34, 37, 40, 46, 76, 112, 117, 136, 138, 143, 145, 185, 205, 210, 226 density, network. See under network measures Democratic Party (U.S.), 93, 121 Democratic Study Group (U.S. Congress), 94, 97 Dingell, John, 144 Disability Intergroup (EP), 149, 171, 193, 240 See also European Disability Forum Dreier, David, 1 ego network. See network eigenvector centrality. See under network measures Page 268 →

electoral system, 12, 15, 49, 51, 54–56, 59, 70, 75, 222 majoritarian electoral system, 12, 21, 51, 54–60, 62, 69, 74, 75, 77, 93, 104, 126, 181–82, 202 modified electoral system, 54–55, 57–58, 60, 62, 64, 66, 68, 204 nonmajoritarian electoral system, 63–64, 66, 68, 204 proportional representation, 21, 49, 54–58, 60, 62, 64, 66, 68, 75, 126, 182, 203–4, 224 embeddedness, 36–37, 43, 77, 147, 177, 202 Employment and Social Affairs Committee (EP), 195 endogenous network. See under network English, Phil, 106, 130, 133, 135 Environmental and Energy Study Conference (U.S. Congress), 96 ERGM. See exponential random graph models Esterling, Kevin M., 5, 14, 25, 36, 43–44, 46, 219, 221 Estonia. See under LMOs in ethics, 96, 99, 164, 166 See also gift rules; House Ethics Committee ethnic fractionalization, 54–55, 57 Eufores (EP), 87, 226 Eurogroup for Animals, 159–61, 171 See also Animal Welfare Intergroup Europe of Freedom and Democracy group (EFD) (EP), 82, 85, 102, 230 European Center for Parliamentary Research (ECPRD), 53, 64–67, 71, 222 European Commission, 82, 168, 170, 178, 186–87, 196–97, 239, 240 European Conservatives and Reformists (ECR) (EP), 82, 85–86, 101–3, 195, 227 European Disability Forum, 160, 168, 171 See also Disability Intergroup European Parliament (EP) Alliance of Liberals and Democrats for Europe (ALDE), 82, 85–86, 91–92, 101–3, 129, 131–32, 134, 195, 231 Europe of Freedom and Democracy group (EFD), 82, 85, 102, 230 European Conservatives and Reformists (ECR), 82, 85–86, 101–3, 195, 227

European People's Party (EPP), 48, 81, 85–87, 89, 91–92, 101–3, 114, 18, 129–32, 134, 137–38, 140, 149–50, 182, 231 European United Left–Nordic Green Left (GUE-NGL), 82, 85, 92, 101–3, 149, 227 Greens–European Free Alliance (Greens-EFA), 82, 85–86, 91–92, 101–3, 149, 195–97 Independence/Democracy group (IND/DEM), 82, 85, 102 Members of the European Parliament (MEPs), 3, 75, 80, 82, 84, 86–87, 89–90, 98, 101–5, 109, 113–15, 117–19, 122–23, 126, 128–30, 136–38, 142–43, 148–51, 161–63, 168–72, 176–79, 181–83, 186, 192, 203, 225, 227, 230–32, 238–39 Party of European Socialists (PES), 48, 82, 85–86, 91, 101–3, 131–32, 134, 231 Progressive Alliance of Socialists and Democrats (S&D), 48, 82, 85–87, 89, 91–92, 118, 131–32, 134, 182, 195–97, 225 Rules of Procedure, 70, 83, 136, 224–25, 239 Union for Europe of the Nations (UEN), 82, 85, 91, 102, 134 European People's Party (EPP) (EP), 48, 81, 85–87, 89, 91–92, 101–3, 114, 18, 129–32, 134, 137–38, 140, 149–50, 182, 231 European Trade Union Confederation, 159–60, 171, 177, 191 See also Trade Union Coordination Intergroup European Union Strategy for the Baltic Sea Region, 187–89 European United Left–Nordic Green Left (GUE-NGL) (EP), 82, 85, 92, 101–3, 149, 227 European Works Council Directive (EU), 152 See also Trade Union Coordination Intergroup Europe's Strategy for the Baltic Sea Region. See European Union Strategy for the Baltic Sea Region exogenous network. See under network exponential random graph models (ERGM), 194, 198, 215–16 Federation of Associations for Hunting and Conservation in the EU, 159–60 See also Hunting Intergroup Feiock, Richard C., 4, 6, 18, 24, 29–31, 40, 202, 210 Page 269 → Finland. See under LMOs in Flake, Jeff, 139–40, 232 focal legislator (analysis), 113, 137–41, 145 formal legislative institutions. See committee; party

Fourth World European Committee. See Intergroup on Extreme Poverty and Human Rights Fowler, James H., 14, 23, 193, 240 France. See under LMOs in G-20. See Group of 20 geodesic distance. See under network measures Germany. See under LMOs in gift rules, 72, 97, 164 See also ethics; House Ethics Committee Gingrich, Newt, 219 Globalisation Intergroup (EP), 88 GLOBE Network (EP). See under issue groups Government Accountability Office (U.S.), 175 government-opposition dynamic, 21, 83 Granovetter, Mark, 4–5, 19, 32–33, 36, 43, 76, 122, 183, 219–20 Greens–European Free Alliance (Greens-EFA) (EP), 82, 85–86, 91–92, 101–3, 149, 195–97 Griesbeck, Natalie, 91, 92, 129, 131–32, 134 Grosch, Mathieu, 137–38, 140 Grossetête, Françoise, 91, 137–38, 140, 231 Group of 20, 51, 221 Hammond, Susan Webb, 8–9, 20–22, 94–95 Handbook of National Legislatures, 52 Handcock, Mark S., 15, 194, 216 heterogeneous (set of network participants), 4, 19, 26–29, 31, 34, 47, 90, 204 See also homogenous (set of network participants) high-quality information. See information high-utility information. See information Hispanic Caucus. See Congressional Hispanic Caucus Hix, Simon, 55, 83, 227, 232 homogeneous (set of network participants), 26–28, 39, 55–56, 100

See also heterogeneous (set of network participants) House Administration Committee. See Committee on House Administration House Ethics Committee, 166 See also ethics; gift rules House of Representatives. See Congress House Wednesday Group (U.S. Congress), 94 Huckfeldt, Robert, 14, 39–40 Human Rights Commission. See Tom Lantos Human Rights Commission Hungary. See under LMOs in Hunting Intergroup (EP), 159, 192 See also Federation of Associations for Hunting and Conservation in the EU hybrid system. See semipresidential (system) Independence/Democracy group (IND/DEM) (EP), 82, 85, 102 India. See under LMOs in India Caucus. See Congressional Caucus on India and Indian Americans Indonesia. See under LMOs in informality of LMOs. See legislative member organizations information high-quality information, 18, 27, 43, 47, 148, 169, 174–75, 180, 202, 208 high-utility information, 7–8, 20, 43–47, 49, 77, 112, 147, 159, 169, 175, 177, 180, 184, 202–3, 208–9, 220 information acquisition, 9, 20 informational benefits, 6, 15, 18, 24, 26, 29, 30, 39, 46, 48–49, 76, 109–10, 112, 121, 153, 156–57, 182, 184, 203 informational deficit, 4, 25, 29, 217 information exchange, 4–5, 7, 9–10, 12–13, 16, 18–22, 25, 27–29, 32, 36, 38, 40, 48–49, 56–57, 62–63, 74, 77, 87, 94, 110, 112, 119, 128, 145–49, 152, 155, 159, 168, 172, 175, 180, 182–85, 203–5, 208–9, 220–21, 232 information networks, 10, 76, 147, 210–11 information provision, 9, 42, 175–76 informational resources (available to legislator), 24–25, 30 low-utility information, 43–44, 46 policy information, 7, 12, 25, 42, 76, 147–49, 155, 158, 170–71, 175, 184, 203, 208, 221, 223

political information, 4, 8–9, 13, 16, 19, 25, 42, 48–49, 146–50, 155, 175, 180, 182–83, 201, 205, 218, 221 See also coordination (function of LMOs); signaling (function of LMOs) Page 270 → inside advocate, 17, 42, 48, 77, 147, 158–59, 161, 164, 166, 168, 184, 221 See also outside advocate institutionalized LMOs. See legislative member organizations interest groups. See outside advocate intergroup Animal Welfare Intergroup, 159, 161, 169–71, 178, 191–93, 230, 240 Baltic Europe Intergroup, 88, 187–89 Disability Intergroup, 149, 171, 193, 240 Globalisation Intergroup, 88 Hunting Intergroup, 159, 192 Intergroup of Elected Local and Regional Representatives, 84 Intergroup on Ageing and Intergenerational Solidarity, 3, 88, 159–60, 171, 193, 195 Intergroup on Climate Change, Biodiversity and Sustainable Development, 88, 90, 178, 195 Intergroup on Extreme Poverty and Human Rights, 3, 88, 160–61, 171 Intergroup on Financial Services, 84 Intergroup on Lesbian, Gay, Bisexual, and Transgender Rights, 88, 160, 162, 193, 195–97, 233 New Media Intergroup, 3, 88 Peace Initiatives Intergroup, 88, 160, 162 Public Services Intergroup, 3, 88, 160–61, 170 Rules Governing the Establishment of Intergroups, 84–85, 225 Sky and Space Intergroup, 3, 88, 90, 160 SME “Small and Medium-Sized Enterprise” Intergroup, 88, 90, 117, 160, 170, 193 Social Economy Intergroup, 3, 87–89 Tibet Intergroup, 3, 88, 159–60 Trade Union Coordination Intergroup, 118, 150–51, 159, 170–71, 177, 191, 193 Ways of Saint James Intergroup, 88, 118

See also issue groups Intergroup of Elected Local and Regional Representatives (EP), 84 Intergroup on Ageing and Intergenerational Solidarity (EP), 3, 88, 159–60, 171, 193, 195 See also AGE Platform Intergroup on Climate Change, Biodiversity and Sustainable Development (EP), 88, 90, 178, 195 Intergroup on Extreme Poverty and Human Rights (EP), 3, 88, 160–61, 171 See also ATD Fourth World Intergroup on Financial Services (EP), 84 Intergroup on Lesbian, Gay, Bisexual, and Transgender Rights (EP), 88, 160, 162, 193, 195–97, 233 International Conservation Caucus (U.S. Congress), 61, 153, 165, 188, 193 Israel. See under LMOs in issue groups (EP), 47, 84, 87, 89, 162 Eufores, 87, 226 GLOBE Network, 87, 226 Kangaroo Group, 87, 235, 240 LEFT Caucus, 87 Spinelli Group, 87 issue networks, 47 issue scope, 19, 25–31, 47, 76, 178 Italy. See under LMOs in Jacobs, Francis, 83–84, 86, 123, 240 Japan. See under LMOs in Judge, David, 9, 20, 22, 219 Kaczyński, Lech, 223 Kangaroo Group (EP). See under issue groups Krehbiel, Keith, 2, 27, 31, 44, 219 Korean Caucus (U.S. Congress), 186–87 Korea Free Trade Agreement (U.S.), 186–87 See also Korean Caucus

Lantos, Tom, 99 See also Tom Lantos Human Rights Commission Larsen, Richard, 106–7, 129, 133, 140, 232 lawmakers. See legislators Lazer, David M. J., 5, 14, 36, 218–19 LEFT Caucus (EP). See under issue groups legislative member organizations (LMOs) coordination (function of LMOs), 3, 5, 9–10, 13, 20–21, 23–24, 38–39, 48–49, 56–57, 63, 73–74, 77, 84, 94, 110, 117, 119, 155, 159, 171, 181–83, 203–4, 206 LMO leadership network (see under network) LMO network (see under network) informality of LMOs, 60–61, 65, 68–71, 116, 145–46, 150–51, 155, 171, 204, 209–10, 213, 222 information (function of LMOs), 9, 16, 20–22, Page 271 → 24, 48, 74, 77, 89, 109–10, 148, 152–53, 172, 181–82, 184, 202–5 institutionalized LMOs, 11, 24, 45–46, 52–53, 59, 66, 147, 159, 175, 177, 184, 208–9, 222 low-cost (nature of LMOs), 12, 19, 24, 40, 43, 76, 145, 156, 184, 202, 209–10 noninstitutionalized LMO, 52–53, 60, 72, 222–23 signaling (function of LMOs), 9–10, 13, 21–22, 38, 48–49, 57, 74, 77, 110, 182–83, 202–3 (see also cheap talk) single-LMO network (see under network) See also caucus; intergroup; LMO activities; LMOs in legislative research services, 25, 69 legislative service organization (LSO), 95–97 See also caucus legislative subsidies, 6, 17, 42, 49, 77, 110, 147, 158–59, 162, 164, 166, 184, 209 legislators, 1–13, 15–20, 22–35, 37–44, 46–50, 53, 56, 58, 61–62, 66–67, 69–70, 72–81, 83, 86, 90, 93, 100–102, 104, 106, 109–13, 115–16, 122–23, 125–30, 134–39, 141–47, 149–50, 156–57, 159, 166–69, 170–71, 174–77, 180–82, 184, 186, 192–94, 198–99, 201–12, 215–19, 221, 223, 227–28, 232 Lesbian and Gay Intergroup. See Intergroup on Lesbian, Gay, Bisexual, and Transgender Rights Lewis, Ron, 139–40, 232 Lithuania. See under LMOs in LMO activities, 5, 8, 30, 38, 41, 73, 112, 147, 186, 192, 205

agenda maintenance/setting, 9, 20–21, 93, 186 coalition building, 9, 20–21, 29–30, 69, 117, 182–83 consensus building, 1, 21, 60, 73, 117, 163, 177, 203, 205 coordination (see under legislative member organizations) information acquisition/exchange/provision (see information) events, 7, 23, 30, 38, 41–43, 60, 67, 72, 86, 99–100, 116, 122–25, 153–55, 161, 164, 166–68, 172, 176, 178–79, 184 signaling (of policy priorities to constituents) (see under legislative member organizations) tracking legislation, 150, 168 LMOs in Australia, 52–53, 59–61 Belgium, 52–53, 60, 62, 68, 70, 227 Brazil, 52–53, 60, 62–63, 71–73 Canada, 52, 59–61, 71, 73 Chile, 52–53, 60, 62, 68 Estonia, 52, 60, 62, 66, 73, 227 European Union (see intergroup) Finland, 52, 60, 62, 66, 71–73, 227 France, 52, 59–60, 71–73, 224, 227 Germany, 52–53, 60, 62, 64, 189, 227 Hungary, 52, 60, 62, 65, 70, 189, 227 India, 52–53, 59–62, 73 Indonesia, 52–53, 60, 62, 64, 70 Israel, 21, 52, 60, 62, 66, 68, 70–73 Italy, 52–53, 60, 62, 65, 70–73, 223, 227 Japan, 52, 60, 62, 64–65, 71–73 Lithuania, 52, 60, 62, 67, 223, 227 New Zealand, 52–53, 60, 62, 68, 70 Norway, 52, 60, 62, 68, 72–73 Poland, 52, 60, 62, 64–65, 70, 189, 224, 227

South Africa, 52, 60, 62, 65, 223 South Korea, 52–53, 60, 62, 65–66, 70, 223 Sweden, 52, 60, 62–63, 227 Switzerland, 52, 60, 62, 67, 70–73, 224 United Kingdom, 52–53, 59–61, 71, 218, 227 United States (see caucus) lobbying, 14, 42, 47, 84, 86, 97, 153, 167, 176, 211, 221, 224 lobbyists, 6, 8, 11, 22, 39, 44, 60, 72–73, 158, 162, 170, 175, 177, 202, 205, 211, 221 (see also outside advocate) LoBiondo, Frank, 106, 130, 133, 135 logrolling, 48, 183 low-quality information. See information low-utility information. See under information L'Union Sociale pour l'Habitat, 161 See also Public Services Intergroup majoritarian electoral system. See under electoral system Mann, Thomas, 91, 129, 131–32, 134 maximum flow, 129–31, 133 McDermot, James, 106–7, 129, 133, 135 McNulty, Michael, 106, 129, 133, 135 McPherson, J. Miller, 39, 217 Members of Congress for Peace through Law (U.S. Congress), 94 Page 272 → modified electoral system. See under electoral system Monti-Kroes package (EU), 170, 236 See also Public Services Intergroup Moore, Dennis, 106–7, 129, 133, 135 Morillon, Philippe, 91, 129, 131–32, 231 Mulhollan, Daniel P., 9, 20–22, 94 multiparty policy coalitions, 13, 69, 117, 147, 203, 206

National Earthquake Hazards Reduction Program (U.S.), 186–87 negotiation. See bargaining network density (see under network measures) ego network, 113, 137–39, 141–42, 145 endogenous network, 26, 29, 31, 76, 204–5, 210 exogenous network, 26–28 geodesic distance (see under network measures) institutional ties network, 136, 138–39, 141–45, 211 legislative network, 5–6, 19, 35, 37, 39, 45, 76, 135, 142, 202, 211–12, 232 LMO leadership network, 38–39, 46, 118, 208, 211 LMO network, 4, 5, 7, 10, 13, 16, 29, 31–35, 37–39, 40, 42, 45–48, 76–77, 109, 112, 125–30, 134–35, 142, 145, 147, 184, 193, 198–99, 202, 210–11, 220, 231–32 maximum flow (see under network measures) single-LMO network, 43, 147, 202, 211, 220 network measures, 231, 240 betweenness centrality, 130, 134–35, 240–41 constraint (Burt's), 113, 141, 144–45 (see also Burt, Ronald S.) density, 36, 126–27, 144, 221, 231 eigenvector centrality, 198–99, 240–41 geodesic distance (between legislators), 127–29, 231 maximum flow, 129–31, 133 New Media Intergroup (EP), 3, 88 New Zealand. See under LMOs in nongovernmental organizations (NGOs), 65, 67, 73, 84, 86, 178–79, 213 noninstitutionalized LMOs. See under legislative member organizations nonmajoritarian electoral system. See under electoral system Northeast-Midwest Congressional Coalition (U.S. Congress), 80 See also Upper Mississippi River Task Force Norway. See under LMOs in

Obamacare. See Affordable Care Act obligatory participation (in networks), 26, 28, 38, 41 OECD. See Organisation for Economic Co-Operation and Development Older Americans Caucus (U.S. Congress), 3, 107 O'Neill, Tip, 94 ordinary legislative procedure (EU), 82, 225 Organisation for Economic Co-Operation and Development (OECD), 51, 221 outside advocacy organization. See outside advocate outside advocate, 6–8, 11, 14–15, 17, 19, 22, 40–47, 49, 56, 61, 72–73, 76–77, 80, 84, 86, 103, 110, 113–15, 117–18, 147, 158–59, 161–64, 166–69, 170–71, 173–77, 179–80, 184, 201–2, 205, 208–9, 211, 221 See also inside advocate; lobbying: lobbyists parliamentarians. See legislators parliamentary (system), 3, 48, 50, 52–55, 57, 60–65, 67–68, 70–73, 83–85, 136, 186, 214, 217–19, 222–24, 230 See also presidential (system); semipresidential (system) party, 3–4, 6, 12–13, 18, 27, 29, 115, 204, 209–11 party affiliation, 30, 70, 82, 104, 109, 112, 139 party discipline, 21, 93 party list, 21, 49, 182, 227 party networks, 14, 28 party organizations, 2, 27, 80 partisan polarization (in the U.S. Congress), 11, 97, 206–7, 212, 241 Party of European Socialists (PES) (EP), 48, 82, 85–86, 91, 101–3, 131–32, 134, 231 See also Progressive Alliance of Socialists and Democrats (S&D) Patterson, Samuel C., 14, 23 Payne, Donald, 106–7, 129, 133 Peace Initiatives Intergroup (EP), 88, 160, 162 Page 273 → Pelosi, Nancy, 99 personal staff, 23, 25, 80, 98, 123–24, 217, 223 personal voting, 54–57, 203

Poland. See under LMOs in policy entrepreneurs, 218 policy formulation, 20 policy information. See under information political information. See under information political intelligence. See information Polity IV Dataset, 51, 221 presidential (system), 54–57 See also parliamentary (system); semipresidential (system) privileged access (to LMOs), 6–7, 19, 41, 44–46, 48–49, 77, 147, 158–59, 163, 166, 175, 177, 179, 184, 202, 208 privileged relationship. See privileged access Pro-Life Caucus (U.S. Congress), 61, 188, 193 Progressive Alliance of Socialists and Democrats (S&D) (EP), 48, 82, 85–87, 89, 91–92, 118, 131–32, 134, 182, 195–97, 225 See also Party of European Socialists (PES) Progressive Caucus. See Congressional Progressive Caucus proportional representation. See under electoral system Public Services Intergroup (EP), 3, 88, 160–61, 170 See also L'Union Sociale pour l'Habitat Quaestors (EP), 136, 232 rapporteur (EP), 170, 172, 192 reciprocity, 2, 36, 38–39, 211 Republican Party (U.S.), 93, 96, 121 Republican Study Committee (U.S. Congress), 16, 100, 139, 226 reputational costs, 7, 37, 43, 219 Résumé of Congressional Activity (U.S.), 199 Richardson, Laura, 225 Roberts, Pat, 96 Rural Health Care Coalition (U.S. Congress), 3, 107 Scholz, John T., 4, 6, 14, 18, 24, 29–31, 40, 202, 210

Scottish Parliament, 9 semipresidential (system), 54–55 See also parliamentary (system); presidential (system) Senate. See under Congress seniority, 2, 105, 194, 199, 216 Services Directive (EU), 180, 191, 193 Services of General Economic Interest package (EU). See Monti-Kroes package Shackleton, Michael, 83–84, 86, 123, 240 signaling (function of LMOs). See under legislative member organizations single-member district plurality electoral system. See electoral system Sky and Space Intergroup (EP), 3, 88, 90, 160 SME “Small and Medium-Sized Enterprise” Intergroup (EP), 88, 90, 117, 160, 170, 193 Social Economy Intergroup (EP), 3, 87–89 social fragmentation, 33, 94 social network analysis, 13–14, 16, 19, 22–23, 78, 112, 125 South Africa. See under LMOs in South Korea. See under LMOs in Speaker's Task Force for a Drug-Free America (U.S. Congress), 80 Spinelli Group (EP). See under issue groups Stevens, Arthur G., Jr., 9, 20–22, 24, 94 Strasbourg, 81, 85–86, 160, 162, 230, 240 Striffler, Michele, 91, 129, 131–32, 134 strong ties, 5, 19, 32–36, 38–39, 45–46, 76, 111–12, 122–24, 145, 183, 211, 217, 219–20 structural holes, 5–6, 18–19, 32, 35–37, 39, 46, 76, 97, 143, 211–12, 220 Stupak Amendment (U.S.), 188 See also Affordable Care Act; Pro-life Caucus Sturdy, Robert, 137–38, 140 Sweden. See under LMOs in Switzerland. See under LMOs in

Thomas, Bill, 96 Tibet Intergroup (EP), 3, 88, 159–60 Tom Lantos Human Rights Commission (U.S. Congress), 99, 108, 165 See also Amnesty International Toubon, Jacques, 91, 129, 131–32, 134, 138, 140 Trade Union Coordination Intergroup (EP), 118, 150–51, 159, 170–71, 177, 191, 193 See also European Trade Union Confederation transaction costs, 6, 18–19, 24–25, 29, 33 transitivity, 32–33 Page 274 → Treaty of Lisbon, 82, 225 triadic closure, 32, 36, 194 Tuesday to Thursday club (U.S. Congress), 1 Ulmer, Thomas, 92, 129, 131–32, 231 Upper Mississippi River Task Force (U.S. Congress), 80 See also Northeast-Midwest Congressional Coalition United States Agency for International Development (USAID), 68–69, 80 U.S. Congress. See Congress U.S. House of Representatives. See Congress U.S. Senate. See Congress. unicameral (legislature), 20, 54–55 See also bicameral (legislature) Union for Europe of the Nations (UEN) (EP), 82, 85, 91, 102, 134 United Kingdom. See under LMOs in USAID. See United States Agency for International Development Uzzi, Brian, 36, 44, 220 Van Hollen, Christopher, 129 voluntary institutions/organizations, 3, 5, 6, 8, 11–12, 16, 18, 20, 24, 29, 33, 41, 43, 46, 201, 209–10, 217 Walsh, Jim, 96

Waxman, Henry, 106–7, 129, 133 Ways of Saint James Intergroup (EP), 88, 118 weak ties, 4–5, 8, 19, 32–35, 37–38, 46–47, 76, 111–12, 122, 125, 183, 210–12, 217 Working Time Directive (EU), 170 See also Trade Union Coordination Intergroup Yellow Book. See Congressional Yellow Book