Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

One-party deliberations in the U.S. House of Representatives Bendix, William Claus 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2012_fall_bendix_william.pdf [ 1.04MB ]
Metadata
JSON: 24-1.0073053.json
JSON-LD: 24-1.0073053-ld.json
RDF/XML (Pretty): 24-1.0073053-rdf.xml
RDF/JSON: 24-1.0073053-rdf.json
Turtle: 24-1.0073053-turtle.txt
N-Triples: 24-1.0073053-rdf-ntriples.txt
Original Record: 24-1.0073053-source.json
Full Text
24-1.0073053-fulltext.txt
Citation
24-1.0073053.ris

Full Text

ONE-PARTY DELIBERATIONS IN THE U.S. HOUSE OF REPRESENTATIVES by William Claus Bendix  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES (Political Science)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) August 2012  ©William Claus Bendix, 2012  Abstract  Over the last two decades, the majority party in the U.S. House of Representatives has increasingly bypassed committee deliberations and restricted floor debate to prevent the minority party from shaping provisions in bills. In response to this rising partisanship, scholars have attempted to measure the direct influence that parties have on legislative outcomes, often by examining roll-call votes. However, they have done relatively little work on the decline of the committee system and the increasing control that majority-party leaders exert on legislative deliberations. In this project, I examine the way in which increasingly cohesive parties draft bills at the prefloor and floor stages of the legislative process. I ask two questions. First, when does the House majority party seize full control of bill development and exclude the minority party from decision-making? Second, what are the policy consequences of the House majority party barring the minority party from formal legislative deliberations? To answer the first question, I construct multiple measures of legislative actions and procedures for major bills developed between 1983 and 2008 in the House. I find that the majority party controls deliberations when its majority status is threatened or its policy goals face strong opposition. I also find that the majority limits discussion on bills designed to promote its electoral brand. These party-brand bills involve tax and welfare policies as well as moral issues. To answer the second question, I create new measures of bill defects and bill extremity to assess how legislative processes affect bill quality. I determine that truncated deliberation tends to produce substantively problematic legislation. Thus, the findings in this project reinforce concerns over and provide evidence of the negative effects of one-party lawmaking.  ii  Table of Contents Abstract ........................................................................................................................................... ii Table of Contents ........................................................................................................................... iii List of Tables .................................................................................................................................. v Acknowledgements ....................................................................................................................... vii 1. Introduction ................................................................................................................................. 1 2. When the Majority Rules: Parties and the Control of Deliberative Processes ........................... 7 2. 1 Introduction .......................................................................................................................... 7 2.2 Variations in Deliberative Processes ................................................................................... 10 2.3 Leadership Influence on Legislation ................................................................................... 11 2.4 Explaining Majority Control and Minority Involvement .................................................... 13 2.4.1 Partisan Polarization ..................................................................................................... 16 2.4.2 Electoral Pressures ........................................................................................................ 17 2.4.3 Unified Government ..................................................................................................... 18 2.4.4 Republican Control ....................................................................................................... 19 2.5 Data and Variables .............................................................................................................. 20 2.5.1 Dependent Variables..................................................................................................... 21 2.5.2 Independent Variables .................................................................................................. 23 2.6 Results ................................................................................................................................. 25 2.7 Discussion ........................................................................................................................... 33 2.8 Conclusion........................................................................................................................... 36 3.Bill Variation and Committee Leadership: Additional Explanations of One-Party Control ..... 38 3.1 Introduction ......................................................................................................................... 38 3.2 Agreement and Disagreement between the Parties ............................................................. 41 3.3 Explaining Policy Control ................................................................................................... 45 3.3.1 Party-Brand Hypothesis ................................................................................................ 47 3.3.2 Intraparty-Conflict Hypothesis ..................................................................................... 51 3.3.3 Committee-Leadership Hypotheses .............................................................................. 52 iii  3.4 Data and Coding .................................................................................................................. 54 3.4.1 Dependent Variables..................................................................................................... 55 3.4.2 Bill Variables ................................................................................................................ 56 3.4.3 Committee Variables .................................................................................................... 58 3.5 Results ................................................................................................................................. 60 3.6 Discussion ........................................................................................................................... 70 3.7 Are Markups a "Formality"? ............................................................................................... 73 4. Deliberative Deficit: The Consequences of One-Party Lawmaking......................................... 75 4.1 Introduction ......................................................................................................................... 75 4.2 Concerns over One-Party Deliberations .............................................................................. 78 4.2.1 Limitations of the Literature ......................................................................................... 79 4.3 The Effectiveness of Group Deliberations .......................................................................... 80 4.4 Data and Methods................................................................................................................ 83 4.4.1 Measuring Legislative Defects ..................................................................................... 83 4.4.2 Measuring Legislative Extremity ................................................................................. 87 4.4.3 Independent Variables .................................................................................................. 88 4.5 Results ................................................................................................................................. 90 4.6 Tracing the Effects of Hearings on Bills ............................................................................. 98 4.6.1Democratic Bills ............................................................................................................ 99 4.6.2 Republican Bills ......................................................................................................... 100 4.7 Leadership Strategies in Markups ..................................................................................... 102 4.8 Discussion and Conclusion ............................................................................................... 106 5. Conclusion .............................................................................................................................. 109 5.1 Future Directions ............................................................................................................... 112 References ................................................................................................................................... 116 Appendices .................................................................................................................................. 126 Appendix A: Estimates of Committee Factors and Bill Types on Hearings and Markups ..... 126 Appendix B: Supplementary Data and Analyses on Bill Quality ........................................... 132 Appendix C: Descriptive Statistics ......................................................................................... 137  iv  List of Tables  Table 2.1. Logit Results for the Effects of Political Conditions on the Holding of Hearings ..... 27 Table 2.2. OLS Results for the Effects of Political Conditions on the Ideological Balance of Committee Witnesses.................................................................................................................... 29 Table 2.3. Logit Results for the Effects of Political Conditions on the Holding of Markups ..... 31 Table 2.4. Logit Results for the Effects of Political Conditions on the Use of Open Floor Rules ....................................................................................................................................................... 32 Table 3.1. Logit Results for the Effects of Bill Type on the Holding of Hearings ....................... 61 Table 3.2. Negative Binomial Results for the Effects of Bill Type on the Number of Hearings . 63 Table 3.3. Logit Results for the Effects of Bill Type on the Holding of Markups ....................... 65 Table 3.4. Negative Binomial Results for the Effects of Bill Type on the Number of Markups . 66 Table 3.5. Negative Binomial Results for the Effects of Committee Factors on the Number of Hearings ........................................................................................................................................ 68 Table 3.6. Negative Binomial Results for the Effects of Committee Factors on the Number of Markups ........................................................................................................................................ 69 Table 4.1. ZINB Results for the Effects of Legislative Deliberations on All-Sources Editorial Criticism...................................................................................................................................... 126 Table 4.2. ZINB Results for the Effects of Legislative Deliberations on Nonpartisan Editorial Criticism ....................................................................................................................................... 93 Table 4.3. OLS Results for the Effects of Legislative Deliberations on Partisan Editorial Divergence .................................................................................................................................... 95 Table 4.4. OLS Results for the Effects of Legislative Deliberations on Bill Extremity............... 96 Table A1. Logit Results for the Effects of Committee Factors on Hearings Being Held ........... 126 Table A2. Logit Results for the Effects of Committee Factors on Markups Being Held ........... 127 Table A3. Logit Results for the Effects of Bill Type on the Holding of Hearings ..................... 128 v  Table A4. Negative Binomial Results for the Effects of Bill Type on the Number of Hearings ..................................................................................................................................................... 129 Table A5. Logit Results for the Effects of Bill Type on the Holding of Markups .................... 130 Table A6. Negative Binomial Results for the Effects of Bill Type on the Number of Markups ..................................................................................................................................................... 131 Table B1. Editorial Endorsements of Presidential Candidates by Party, 1992-2008 ................. 132 Table B2. Frequency Distribution of All-Sources Editorial Criticism ....................................... 134 Table B3. Frequency Distribution of Nonpartisan Editorial Criticism ....................................... 135 Table B4. OLS Results for the Effects of Legislative Deliberations on Bill Extremity ............. 136 Table C1. Descriptive Statistics for Dependent and Independent Variables .............................. 137 Table C2. Percentage of Bills that Undergo Truncated Deliberations at Each Legislative Stage for Each Congress ....................................................................................................................... 138 Table C3. Data for Partisan Polarization and Electoral Risk Variables by Congress................. 139  vi  Acknowledgements  I started this project with an interest in political rhetoric and floor debate in the partisan Congress. However, with advice from Paul Quirk, I expanded my focus to include both the prefloor and floor stages of congressional deliberations. The scope of this topic was more ambitious than what I had initially contemplated, and I am grateful to Paul for encouraging me to think big. I am also grateful to him for his guidance and support as my supervisor through the inevitably difficult process of completing the dissertation. His suggestions were always constructive, and his demand for clear, precise prose forced me to strengthen arguments and to improve my writing. Perhaps most important, he helped me develop confidence in my ideas and my abilities as a researcher. I cannot thank him enough for all his encouragement over the years. As well, I am grateful to my committee members, Fred Cutler and Alan Jacobs, who were generous with their time and eager to discuss research challenges. With his door always open, Fred welcomed me into his office at a moment's notice and let me interrupt his day, in some cases, for hours at a time. Alan was equally accommodating and, in our many conversations, he pushed me to think more carefully about research design. Both Fred and Alan made me a better scholar. I must also thank Gyung-Ho Jeong, who, upon coming to the University of British Columbia last year, acted almost as a fourth committee member. He read drafts of the dissertation and offered practical feedback as the project neared completion. In addition to the faculty members at UBC, I am indebted to James Downey of the University of Waterloo. He hired me as his research assistant when I was still an undergraduate student, and he inspired me to pursue graduate studies.  vii  I was lucky to make a number of close friends at UBC, many of whom provided moral support, discussed research ideas, and read drafts of my dissertation chapters. Nathan Allen, Shane Barter, Tim Came, and John McAndrews deserve my thanks. I appreciate all the help and advice that they gave me on matters big and small as I worked to finish this project. Without financial assistance, I could not have completed this dissertation. I want to acknowledge the Faculty of Graduate Studies at UBC for providing me with the Tuition Award. I also want to acknowledge the generous support that I received from the Cordula and Gunter Paetzold Graduate Fellowship. The Fellowship provided two years of funding and allowed me in that time to concentrate fully on putting together the data sets crucial to this project. It is not easy to live with someone who is writing a dissertation. But my wife, Patsy, demonstrated infinite patience and understanding over the years, and took care of all the important things that I overlooked as I collected, coded, and dreamed about data. Above all, I am amazed by and grateful for her willingness to crisscross the continent so that I could pursue academic goals. I owe much to my son, Liam, as well. He always found a way to make me laugh, and he performed the essential duty of making sure that I never took my work or myself too seriously. Both Patsy and Liam sacrificed a great deal for me; the least I can do now is dedicate this project to them.  viii  To Patsy and Liam  ix  1. Introduction  To develop and select effective policy solutions, Congress conducts legislative deliberations in committees and on the floor. Committees undertake extensive information gathering, discussion, and negotiation before recommending bills to the whole House or Senate. Members keep committee assignments indefinitely and thus gain expertise that helps them develop workable legislation, satisfy conflicting interests, and avoid adverse consequences. Moreover, because committees include both Republican and Democratic members, committee debate exposes legislative proposals to a wide range of perspectives (Fenno 1973; Krehbiel 1991; Quirk 2005; Rohde 2005; Baughman 2006). Floor deliberation is also important for the development of legislation (Mucciaroni and Quirk 2006). Procedural rules give lawmakers from both parties opportunity to improve committee versions of bills by offering, discussing, and voting on amendments during the floor debate. However, congressional observers have raised serious concerns that the parties have undermined the formal deliberative processes in Congress, especially in the House of Representatives (Cohen 1999; Mann and Ornstein 2006; Rohde 2005). Over the last two decades, increasingly intense partisan conflict has led majority-party leaders in the House to limit committee deliberations or to bypass them altogether. They take such steps to diminish the influence that their partisan rivals have on the development of legislation. Committee chairs have sometimes shut down hearings prematurely to stop minority-party witnesses from testifying (Ahuja 2008, 33); chairs have also forced committees to vote on draft legislation that minorityparty members have been prevented from reading (Hetherington 2009, 413). Even more striking, the Speaker and other senior party leaders have often excluded the committees entirely from the  1  decision-making process, having key proposals drafted by ad hoc party task forces. Republican House Speaker Newt Gingrich used this strategy to develop bills based on the "Contract with America" (Sinclair 1997), while Democratic House Speaker Nancy Pelosi used it to develop and pass her "100-Hour Agenda" (Pearson and Schickler 2009). Both parties, when in the majority, have also undermined floor deliberations by placing restrictive rules on most bills to prevent floor amendments from being offered and debated. Aldrich and Rohde (1997, 2000a, 2000b, 2009) have presented a leading explanation for the increased use of these partisan strategies in Congress. In their theory of conditional party government, they argue that as each party takes on increasingly homogenous preferences, rankand-file members delegate greater legislative power to their leaders. Majority caucus members, in other words, allow the Speaker and other senior party members to exercise increasing control over bill development to ensure that the minority, with its divergent preferences, has no influence on bill provisions. Multiple studies on floor votes confirm that the two parties have become increasingly cohesive and polarized (e.g., Han and Brady 2007; McCarty, Poole, and Rosenthal 2006; Theriault 2008), and case studies on prefloor deliberations detail how majority-party leaders draft complex and important measures in backroom sessions (Mann and Ornstein 2006; Sinclair 1997, 2006). However, researchers have not determined why majority-party leaders take control of bill development in some cases but allow committees to draft bills in other cases. Little systematic data on prefloor legislative activity has been collected and analyzed. Moreover, researchers have not given enough attention to the internal workings of committees in the partisan era (Evans 2011, 397). It is unclear to what extent a highly cohesive majority will consider proposals or adopt amendments from the minority party during panel sessions. And  2  while concerns over increasing partisanship are commonly voiced, scholars have not determined what effect party-directed deliberation has on legislative quality. Therefore, in this project, I ask two questions concerning one-party legislating in the House. First, when does the majority party seize total control of bill development and bar the minority party from the legislative process? Do certain political factors, policy issues, or committee features encourage one-party lawmaking? Second, what are the policy consequences of the majority dominating bill deliberation and preventing the minority from participating in decision-making? Are bills drafted in closed-door sessions more likely to contain significant mistakes than bills written in committee rooms? I address these questions in three chapters, designed to stand on their own (for eventual publication as articles). This introduction, then, is an overview of the dissertation for nonspecialist readers (and is not intended for publication). To answer my research questions, I compile an original data set on the development of House bills, with an emphasis on procedures, to measure the level of one- and two-party lawmaking. The data are collected on156 major bills developed over the last three decades, and are used to determine the extent to which members of the minority party participate in both committee deliberations and floor debate. I develop nine different measures of legislative deliberations, many of them new to the congressional literature. In developing these measures, I assume that the absence of formal legislative steps indicates one-party control. So, if the minority party has opportunity to invite witnesses in hearings and to offer amendments in markup and on the floor, then it has significant opportunity to shape provisions in bills. Conversely, if hearings and markups are omitted and floor debate is closed, then the minority has no significant opportunity to influence legislative content.  3  This project proceeds as follows. In Chapter 2, I look at the question of when the majority dominates bill development. In particular, I examine which political factors are correlated with omitted hearings, skipped markups, and restrictive floor rules. I find that increasing partisan polarization and electoral competition lead to higher levels of one-party lawmaking. These findings, as I argue, reinforce the two leading explanations of party influence in Congress: the theory of conditional party government and cartel theory. I also find that the majority grants the minority greater opportunity to participate in bill drafting under conditions of unified government than under conditions of divided government. I identify several reasons why. But mainly I argue that when the House majority faces a divided government, it likely has no interest in compromising early in the legislative process when it will have to compromise later in the conference committee. Overall, the results on political factors indicate that majority-party leaders tend to control legislative deliberations when their party’s majority status is under threat or their party’s policy goals face significant opposition. I continue in Chapter 3 to focus on the question of when the majority dominates. Here again, I examine why some bills undergo one-party processes and others undergo two-party procedures. I look to see if certain types of legislation and different types of committees affect deliberative processes. I find that bills designed to promote the majority party's brand, such as tax measures and moral-values legislation, tend to undergo limited panel sessions. Because the majority party has fixed preferences on these issues, it has little incentive to collect new policy information and to explore counter proposals. If the majority deviates from its long-held policy positions, it risks harming its brand and alienating its core supporters. I also find that bills that initially divide the majority party tend to undergo mixed committee deliberations. On one hand, they often undergo extensive hearings but, on the other, they often skip markups. I argue that  4  majority-party leaders circumvent bill-writing sessions on these measures to prevent committee members from offering destructive amendments and exacerbating majority divisions. As for committee features, I show that chairs with extreme preferences hold fewer panel sessions than chairs with moderate preferences. These results suggest that extreme chairs act as party-brand managers, driving priority bills quickly through their committees to reduce discussion and to limit amendments. In Chapter 4, I identify the policy consequences of one-party deliberations by measuring the level of negative criticism that bills receive in unsigned editorials from a broad selection of partisan and nonpartisan newspapers. I use the level of negative editorial criticism as an indicator of policy mistakes or other problematic features of bills. What I find is that legislation undergoing truncated deliberation receives significantly higher levels of criticism than legislation undergoing regular-order proceedings. These results, I argue, suggest two things: that one-party bills tend to be more defective than two-party bills, and that process matters. When bills undergo one-party deliberations, lawmakers likely consider a narrow range of arguments and evidence and, in turn, fail to consider important counterpoints. Lacking diverse perspectives, they are prone to making serious mistakes in the design of legislation. To explore additional consequences of partisan lawmaking, I examine in Chapter 4 whether one-party processes tend to produce ideologically extreme bills. Extreme bills are problematic, I argue, because they often confer significant benefits to the majority's base at a cost to the rest of the public. To estimate bill extremity, I develop a new measure based on DWNOMINATE and roll-call data. Surprisingly, I determine that truncated deliberation does not have a consistent effect on a bill's ideological position. I find that the presence of hearings is correlated with bill moderation but that, contrary to expectations, the presence of markups is  5  associated with bill extremity. Further investigation shows that moderate bills often divide the majority caucus and encourage majority-party leaders to minimize internal disputes by omitting markup sessions. This finding matches my earlier finding that bills dividing the majority party tend to skip bill-writing sessions. Put together, these results show that not all legislative stages are equally important to the development of effective policy. In fact, I contend that hearings often have a much bigger impact on, and provide the minority a much greater opportunity to influence, bill content than markups do. Finally, in Chapter 5, I discuss the main contributions of this project and suggest several new research directions on congressional deliberations. I argue that scholars need to re-examine the role that markups play in the House today and, in particular, they need to consider whether bill-writing sessions have become empty formalities.  6  2. When the Majority Rules: Parties and the Control of Deliberative Processes  2. 1 Introduction When Republicans assumed control of the House in 1995, Speaker Newt Gingrich drafted several important bills behind closed doors to ensure that Democrats had no input. Gingrich used ad hoc Republican policy groups rather than the committees of jurisdiction to develop a Medicare reform bill, an immigration bill, and several appropriations bills (Sinclair 1997, 17). Democrats objected to the use of these party policy groups and promised to restore regular order once they regained majority status. However, as soon as Democrats took back the House in 2007, Speaker Nancy Pelosi hurried through six bills—on matters ranging from prescription-drug prices to oil-industry subsidies—in just six days. All six bills were drafted outside committee by Democrats and came to the floor under closed rules (Pearson and Schickler 2009, 166-167). Clearly, majority-party leaders exert tremendous control over the legislative path that bills travel. They can decide to schedule hearings or not, to hold markups or not, and to place open or restrictive rules on floor debate. We know that regular order is still often maintained. But we also know that the majority has blocked the minority from bill development with increasing frequency over the last three decades (Binder et al. 2009; Quirk 2005; Rohde 1991, 2005; Sinclair 1997, 2005, 2006). What we do not know is for what reasons the majority decides the minority should be excluded from the legislative process. Is the majority party primarily motivated to control bill development for ideological or electoral reasons? Do majority-party leaders develop varying legislative strategies in response to the partisan composition of Congress and the White House? Is one party more likely to centralize power and exclude the minority than 7  the other party, or are Republicans and Democrats equally prone to exercising control over bill writing? So far, scholars have not provided clear answers to these questions. While we recognize that the majority party dominates legislative deliberations to an ever-greater extent, we have not determined which political conditions encourage the majority to choose domination over cooperation. In fact, not only do we lack a full understanding of why decision-making processes have changed and varied over recent decades, but we also lack an explanation of why they vary within a single Congress. One problem is that scholars have tended to ignore legislative deliberation as an area of concern (for a review, see Quirk and Bendix 2011). Considerable research focuses on how interest groups, public opinion, and electoral pressures affect floor votes, but little work considers how internal congressional processes—such as information gathering by committees— shape decision-making. Few studies, in other words, move beyond roll-call data and attempt to systematically examine the way in which bills are developed in either the House or the Senate. Moreover, the few studies on congressional deliberation often focus on only one stage of the legislative process, such as hearings (Esterling 2004, 2011) or floor debates (Mucciaroni and Quirk 2006; Steiner et al. 2004). No study has attempted a systematic examination of congressional deliberation at all stages—from hearings, to markups, to floor actions. And no study has attempted to determine why the majority chooses at times to expand and at other times to curtail bill deliberations. In this chapter, I attempt to identify the factors that encourage majority-party leaders to seize control of bill development and, in turn, prevent the minority party from participating in House deliberations. Specifically, I consider the question: Under what conditions do closed, oneparty deliberations occur, and under what conditions do open, two-party deliberations take place?  8  To answer this question, I collect systematic data on legislative actions and procedures for bills developed between 1983 and 2008. With these data, I create multiple measures of deliberative processes and conduct statistical analyses to determine which political conditions are strongly associated with one-party control of bill development. As is expected, I find that increasing partisan polarization encourages majority-party leaders to take control of the lawmaking process. I also find that electoral competition plays an important role in fostering one-party deliberation at all legislative stages. More surprising, I find that the minority party has greater opportunity to participate in committee deliberations when government is unified than when it is divided. Also surprising, I find no significant difference in how the two parties conduct prefloor decisionmaking, but I do find that Republican majority leaders are more likely to allow open floor debate than Democratic majority leaders. The contributions I make in this chapter are thus twofold: I document with a set of new measures what scholars have long claimed are the main causes of party-directed lawmaking, but I also challenge core assumptions about the sources of partisan conflict and bipartisan collaboration. This chapter is organized as follows. First, I discuss how deliberative processes in the House have changed over the last three decades, and I review previous research on party-directed lawmaking. I focus especially on research pertaining to party task forces. Next, I consider what the costs and benefits are for the majority party to exclude its partisan rivals from bill development, in order to explain why regular-order proceedings are ever maintained in the partisan Congress. After outlining these costs and benefits, I present testable hypotheses on party-dominated lawmaking. I then describe my methodology for building an original dataset on deliberative processes, and I discuss the results of the data analysis. Finally, I suggest new research directions.  9  2.2 Variations in Deliberative Processes Relatively little research has been conducted on legislative deliberation in Congress. What we do know is that decision-making processes in the House have evolved and varied significantly over the last thirty years. In the 1970s and 1980s, the decision-making process could be understood as a continuum that ran "from very centralized to very decentralized" (Smith 1985, 205). Centralized deliberation was often used to accelerate the passage of major bills, and included senior House members from both parties as well as senators and executive-branch negotiators (Sinclair 1997, 70-72). Decentralized lawmaking involved a division of labor where members on the committee of jurisdiction contributed most to bill development, but other legislators from either party could also influence policy outcomes.1 By the early 1990s, the decision-making process in the House had changed dramatically (Rohde 1991). The issue was no longer whether bills underwent centralized or decentralized processes but, rather, the degree to which majority-party leaders dominated bill deliberations. This situation persists. If deliberation in the House can be understood as running along a continuum, it now runs from closed, one-party control to open, two-party participation. On priority legislation, the Speaker, the Majority Leader, and their staff consult committee chairs as they determine a bill's scope and decide the extent of committee involvement in bill development (Cox and McCubbins 2005, 18-19; Lipinski 2009, 342; Sinclair 2006, 149-150). If majority-party leaders choose a one-party process, they will omit all committee deliberations, draft bills behind closed doors, and place severe restrictions on floor activity. To conduct bill drafting in these cases, the Speaker will establish a small "design team" 1  Smith (1985) distinguishes decentralized deliberation from "collegial" deliberation, where every member participates equally in policy decisions. Collegial deliberation has usually been impractical in the House given the number of legislators and the workload of the chamber.  10  composed of the relevant committee and subcommittee chairs (Sinclair 1997, 17), or the Speaker will convene a party task force, somewhat larger than a design team, that consists of loyal majority-party members from multiple committees (Baughman 2006, 163). If, on the other hand, majority-party leaders allow a bill to undergo some level of two-party deliberation, minorityparty members will have opportunity to participate in at least one legislative stage of bill development. And if leaders allow an open two-party process, minority-party members will have opportunity to participate in all formal legislative stages: hearings, markups, and floor debates. Scholars have measured the frequency with which majority-party leaders employ oneand two-party strategies. Sinclair (2006, 148) reports that 23 percent of major bills developed between 1961 and 1982 underwent partisan procedures, whereas 51 percent of major bills drafted between 1993 and 2004 underwent some level of party-dominated processes. Binder et al. (2009) collect data on committee deliberations over the last two decades and find that, while legislative output has not declined, the number of committee markups has. This decline, however, has not always been consistent. The 110th Congress, for example, had more markups than the 109th Congress (Binder et al. 2009, 39). Overall, the data demonstrate that party-directed deliberation has generally increased over the decades. Yet, the data also show that, even in years of intense partisanship, the majority has often allowed the minority to participate in bill drafting. 2.3 Leadership Influence on Legislation A large literature attempts to measure the direct influence that majority-party leaders have on legislative outcomes. Most of this research examines floor votes and ignores prefloor decision-making (for a review, see Smith 2007, ch. 4). Floor votes provide limited insight into party influence, however, because majority-party leaders make the most important policy decisions well before bills reach the floor (Cox and McCubbins 2005; Sinclair 2002). Yet, what 11  we know about leadership action at the prefloor stage is rather modest and comes from a handful of studies on party task forces. To date, Sinclair has conducted the most extensive research on one-party lawmaking. In a series of case studies, she finds that Democratic House leaders in the 1980s and Republican House leaders in the 1990s used party task forces for two simultaneous purposes: to control the drafting of key legislation, and to instill party loyalty in junior members by involving them in the policy process (1981, 1983, 1995, 1997, 1999). The two parties, Sinclair explains, have also used ad hoc policy groups in different ways. Early Democratic task forces were rather large, sometimes having over forty members, and functioned as an extension of the whip system. The members drafted bills in private and then pressured colleagues on the floor to support these bills (1983, ch. 5). Early Republican task forces were comparatively small and allowed favored lobbyists to help write task-force legislation (Sinclair 1997, ch. 8; 1999). Building on Sinclair's work, scholars have examined voting records and ideological scores of task-force members. They conclude that both parties stack task forces with loyalists and ideologues to ensure these groups fulfill the leaders' priorities (Baughman 2006; Garand 1988; Garand and Clayton 1986). Scholars also note that majority-party leaders sometimes marginalize the minority party without having to organize party policy groups. For example, leaders can package substantive measures into large omnibus spending bills and then usher these bills to the floor just before the end of a session, ensuring that most members do not read or consider key provisions. By rushing an omnibus bill, majority-party leaders are able to pass items that would not otherwise survive a floor vote (Krutz 2001, 32-35; Mann and Ornstein 2006, 173); and by folding so many important measures into one giant bill, they are able to "accomplish much of a session's significant legislating in a single stroke" (Quirk 2005, 333).  12  In sum, the existing research demonstrates that majority-party leaders have direct influence over legislative outcomes by using partisan strategies at the prefloor stage. However, the literature has not identified the specific political conditions that have fostered the rise of task forces and the decline of committees. Put another way, no quantitative study has been conducted to explain variation in one- and two-party deliberation at the prefloor stage. Moreover, existing research has not explained why, within the same Congress, some bills undergo partisan processes and others do not. And previous studies have not determined whether factors that encourage oneparty control at early legislative stages (e.g., hearings) are the same factors that encourage oneparty control at later stages (e.g., floor debate). Thus, we know how majority-party leaders exert influence on legislative outcomes, but we have yet to determine when they exert their influence. 2.4 Explaining Majority Control and Minority Involvement Majority-party leaders operate as agents for the collective principals in the majority caucus (Sinclair 1995, 1999). Theories of congressional parties seek to explain, among other things, why rank-and-file members yield legislative control to their party's leaders. A common explanation is that lawmakers give up procedural powers and bill-writing duties to leaders whose preferences align with their own (Aldrich and Rohde 2000a; Rohde 1991; Sinclair 1995). That is, parties function to help coordinate strategies among legislators who have "common interests," especially common ideological interests (Aldrich 1995, ch. 7). Lawmakers recognize that they will see more policy objectives fulfilled if they defer to their leaders than if they defy them. So far, theories of congressional parties have not considered why majority-party leaders seize control of bill writing in some but not all cases. By taking control of bill deliberations, leaders exclude the minority party from all decision-making and draft bills designed to satisfy the majority caucus. If excluding the minority were a costless activity, or if the benefits of exclusion 13  always outweighed the costs, then the majority would have no incentive to involve the minority in bill development under any circumstances. Why, then, are two-party deliberations ever held? There are three primary reasons. First, the exclusion of the minority party also means the exclusion of committees, and thus majority-party members, from bill deliberations. Committee work provides its members with achievements that they can highlight during election campaigns; it allows them to influence policy; and it helps bolster their reputations within Congress and Washington (Fenno 1973; Lipinski 2009). Majority-party lawmakers lose these benefits if leaders omit committees from deliberations and draft measures behind closed doors. Rank-and-file members have strong incentives, therefore, to uphold their committee's turf in order to maintain benefits—especially electoral benefits—associated with committee work (Adler 2002). Members, and even chairs themselves, can publicly complain about leadership intrusions if their committee is omitted repeatedly from major legislative decision-making (Owens 1997, 254). Second, the exclusion of the minority party (and by extension committees) increases the likelihood of defective legislation being produced. Lawmakers face tremendous risks as they develop legislative solutions to existing policy problems (Arnold 1990; Fenno 1973: Krehbiel 1991). Committee deliberations conducted by members with specialized policy knowledge function primarily to reduce policy uncertainty and to help lawmakers develop bills that avoid embarrassing or unintended results (Krehbiel 1991; Quirk 2005). Hearings play an especially important role because they allow both parties to present and criticize policy evidence during proceedings. House rules stipulate that committee members from the minority party must have opportunity to invite and question witnesses (Brown and Johnson 2003, 278). So, even when chairs stack hearings with experts who support the majority's position, Republicans must still  14  hear from Democratic witnesses and Democrats must still hear from Republican witnesses. Homogenous groups, by contrast, often fail to seek out diverse arguments and information, and adopt flawed positions as a result (Janis 1982; Sunstein 2007). Because majority-party leaders lack the multifaceted expertise of committees, they are especially prone to overlooking major policy concerns when they draft bills behind closed doors. Third, the minority party can voice serious concerns over transparency and accountability when a bill circumvents regular order. Both parties recognize that institutional norms encourage significant participation by the minority in hearings, markups, and floor debates; both parties also recognize that violations of these norms often spark significant objections from the minority and its supporters (Hamilton 2004). Committee markups and open floor debates enable the minority to discuss bills in detail and to offer multiple amendments (Quirk 2005). If majorityparty leaders shutdown markups and severely limit floor action, the minority has no opportunity to shape legislative language.2 It can then raise doubts in the public's mind about the content of bills that have been rushed through the House. The majority can avoid such controversy by upholding regular order. In all likelihood, these factors that encourage two-party deliberations are relatively constant across policy issues. All majority-party members are averse to losing committee duties and the associated electoral benefits, regardless of which committees they sit on. The majority party has no motivation to produce defective legislation and wants to avoid severe policy mistakes in all cases. And the minority party can object to violations of institutional norms  2  Gilligan and Krehbiel (1987; see also Krehbiel 1991) contend that the Rules Committee does not restrict floor amendments for partisan reasons. Rather, it restricts amendments so that a committee knows that its bill will not undergo drastic change on the floor. Assured that its work will remain intact, the committee thus has strong incentive to provide the parent chamber with reliable, specialized information. However, many scholars provide significant counterevidence to demonstrate that majority-party influence accounts for the use of restrictive floor rules (see Cox and McCubbins 2005; Dion and Huber 1996; Roberts 2010; Sinclair 1994).  15  whenever they occur, no matter which policy issue is under consideration. While the majority recognizes the benefits of accommodating the minority, it likely sacrifices these benefits whenever a compelling political interest arises. Variations in deliberative procedures, then, are probably best understood by exploring conditions in which the majority sees greater benefit in bill control than in bill compromise. The next sections present testable hypotheses of why House majority-party leaders dominate lawmaking: 2.4.1 Partisan Polarization Aldrich and Rohde (1997, 2000a, 2000b, 2009), in their theory of conditional party government, explain how ideological cohesion within parties affects the level of power that congressional leaders acquire. They note that members of the majority party delegate legislative decisions to leaders whose preferences are similar to theirs. A party with relatively homogenous preferences will therefore defer frequently to its leaders and grant them significant legislative control. Aldrich and Rohde also note that the ideological composition of the minority party plays an equal role in determining the power of majority-party leaders. As both parties become increasingly unified and hold increasingly divergent preferences, the majority delegates greater decision-making authority to the leadership. With high ideological agreement within the caucus, few majority-party lawmakers worry about being coerced into supporting bills they oppose and most expect to derive benefits from having other party members coerced into supporting important legislation. Furthermore, powerful majority-party leaders ensure that the minority cannot succeed in either passing its own measures or blocking the majority's. Concern within the majority caucus over leaders abusing their power is offset by the guarantee that strong leaders will deny the minority party legislative victories. Consequently, majority-party lawmakers defer to their leaders when facing an opposition that holds consistently contrary policy views. 16  Hypothesis 1: As House Democrats and House Republicans become increasingly polarized, the majority party will increasingly control bill deliberations. 2.4.2 Electoral Pressures The literature on congressional parties assumes a strong relationship between electoral competition and the power of majority-party leaders. But this literature does not explain whether, and to what extent, shifts in public support for the majority party affect the level of leadership power. In their procedural cartel theory, for example, Cox and McCubbins (1993, 2005) contend that majority-party members have a fixed electoral motive for delegating authority to leaders. To win votes, legislative parties develop brand names that have mass appeal. These brand names provide electoral benefit to party members because voters are more familiar with party labels than they are with individual candidates. Because the majority party builds its reputation through its legislative record, majority-party members have a strong electoral incentive to see their party enjoy legislative victories. However, because parties are large, complex organizations comprised of members with competing concerns, the majority caucus finds it difficult to coordinate a votemaximizing strategy. To solve this coordination problem, rank-and-file members grant agendasetting and procedural powers to majority-party leaders. They allow leaders to decide which bills reach the floor and under what circumstances. In turn, leaders block bills that harm and champion bills that enhance the party label. The more efficiently the majority works as a legislative cartel, the greater the electoral payoff. Similarly, Lee (2009) argues that majority-party members have a collective interest to remain in the majority because it enhances their ability to pursue individual professional goals. She contends that many disputes between the two parties are thus driven by "reflexive partisanship" rather than by ideological conflict (2009, 4). That is, they oppose and attack each 17  other's proposals simply to maintain their distinctive party brands. She even contends that if the parties had similar preferences, the majority would still have electoral incentive to use abusive rhetoric and tough procedural maneuvers against the minority. Neither Cox and McCubbins nor Lee discusses how leadership power varies when electoral pressures fluctuate. However, a plausible implication of their theories is that as electoral pressures increase on the majority cartel, majority-party members have even greater incentive to delegate legislative authority to their party leaders. Presumably, a record of accomplishment becomes more important as poll numbers tighten. The majority can boost its legislative accomplishments most efficiently by driving bills through the House beyond minority-party influence. Hypothesis 2: Increases in one-party bill deliberations will occur when polls show decreases in voter support for the majority party. 2.4.3 Unified Government Several scholars note that, in times of unified government, the majority party often coordinates to fulfill the president's promises. Keefe and Hetherington argue that in unified periods the president functions as the "legislator in chief" and majority-party leaders become his reliable allies (2003, 158). Sinclair (2006, 238-241) concurs and indentifies two reasons that lawmakers of the president's party push the president's proposals. First, they tend to agree with the president on key policy goals and, second, they fear public disputes with the White House will make the majority party look inept and disorganized. Because the majority has both a desire and a need to enact the president's program, majority-party leaders dominate bill writing in unified periods and rank-and-file members support these leadership interventions.  18  Evidence also suggests that two-party cooperation most likely occurs in divided government. Trubowitz and Mellow (2005), examining roll calls between 1889 and 2002, identify all cases in which a majority of one party voted with a majority of the other party and determine that bipartisan voting is more prevalent in divided periods than in unified periods. Voters in divided periods, they posit, have sent lawmakers "mixed signals about which party is preferred" and, in response, Republicans and Democrats decide that the best way to win reelection is to be seen cooperating with each other (2005, 445). Perhaps; but more likely, both sides recognize that divided government requires the two parties to compromise on policy if any bills are to pass. Thus, to avoid gridlock, the majority collaborates with the minority on bill writing to produce legislation that appeals broadly to members from both parties. Either way, there is significant reason to expect that unified government encourages one-party deliberation and that divided government encourages two-party lawmaking. Hypothesis 3: The majority party will dominate bill deliberations to a greater extent in periods of unified control than in periods of divided control. 2.4.4 Republican Control Republican legislators are widely seen as having a "deep suspicion of deliberative processes" (Mann and Ornstein 2006, 98). While some scholars note that both Republicans and Democrats have set aside regular order (e.g., Sinclair 2008a), many observers argue that House Republicans, starting with Gingrich, have used an especially aggressive legislative style to shut Democrats out of bill development (Eilperin 2006; Fenno 1997; Mann and Ornstein 2012; Smith and Gramm 2009, 157). Indeed, Fenno describes Gingrich's institutional structuring of the House Republican Party as "the largest concentration of majority party power in a century" (1997, 34; see also Strahan 2007, 127). Speaker Dennis Hastert, Gingrich's successor, is said to have 19  maintained this approach, drafting bills himself and rewriting bills after markup (Baughman 2006, 170-171). Others, beyond congressional scholars, also see Republicans as anti-deliberative. The public (Hayes 2005), pundits (e.g., Dean 2006), and scholars from various disciplines (e.g., Altemeyer 1996; Lakoff 2002) view Republicans as strong leaders who, at times, display authoritarian tendencies. Democrats, by contrast, are often regarded as compassionate figures who, at times, act with alarming weakness. Chris Matthews provides the most succinct articulation of these partisan stereotypes, calling Republicans the "daddy party" and Democrats the "mommy party" (2001, 176). If there is truth to these stereotypes, the two parties should adopt different processes of bill development. Hypothesis 4: Republican control of the House will lead to greater levels of one-party deliberations than Democratic control. 2.5 Data and Variables To investigate one-party lawmaking, I compile a sample of bills drawn from all key votes selected by Congressional Quarterly between 1983 and 2008. Each year CQ identifies about fifteen key votes on congressional resolutions and major bills. I draw cases from only the legislative votes in which CQ mentions specific bills in its summary descriptions. There are approximately ten such legislative votes each year, and I select 156 bills from this pool. I examine major bills because party leaders and rank-and-file members tend to take an interest in such legislation; the same is not always true of less important bills (Clinton and Lapinski 2006).3  3  Key votes are a reliable measure of legislative importance because they "command the attention of the national press and of all political actors" (Mouw and Mackuen 1992, 88), and because "CQ has an excellent reputation for making sound and accurate judgments about congressional politics" (Bond and Fleisher 1990, 71).  20  The findings in this chapter may not generalize to secondary bills; a broader examination of legislation is for future investigation. Because I am interested in change over time, I stratify by year and randomly select six bills from a pool of approximately ten bills for every session of Congress. In the twenty-six-year period that my study covers, both Republicans and Democrats have enjoyed unified control and both Republican and Democratic presidents have faced a House controlled by the opposing party. The sample of bills covers diverse issues, including health care, education, taxes, and civil rights, among others. For appropriations bills, I look at deliberations on only the issue subject to the key vote. I examine, for example, only hearings on the MX intercontinental ballistic missile for a 1983 defense bill, because the key vote was over a funding provision on the missile, not over the whole bill. 2.5.1 Dependent Variables I construct four dependent variables that measure the level of majority-party control over legislative deliberations. To develop these measures, I identify the opportunities that the minority party is given to participate in lawmaking for each bill. I assume that as opportunities for twoparty participation increase, the level of legislative dominance by the majority party decreases.4 Hearings Held is a dummy dependent variable that denotes whether at least one legislative hearing has been held (1) or not (0) for a bill. To identify relevant hearings, I use legislative histories compiled by LexisNexis Congressional. For large legislation, a LexisNexis search often identifies hearings held by several different committees over multiple years. I count only hearings held before the committee or committees that actually report the bill, because these hearings will have involved the members who played a direct role in drafting the legislation. I  4  Whether the minority party takes full advantage of these opportunities is beyond the scope of my research here.  21  also include any relevant hearings from the previous Congress because often members deliberate on policy issues for several years. Some bills reach the floor without being reported by a committee. In such cases, committee chairs usually sponsor the bills. When an unreported bill is sponsored by a committee chair, I count all relevant hearings in the current and previous Congress held by that chair's committee. Witness Balance is the percentage of committee witnesses presumed to represent the ideological views of the minority party. This measure determines whether, and the degree to which, committee chairs stack hearings in the majority's favor. When Democrats are in the minority, I calculate the percentage of presumed liberal witnesses before committee and, when Republicans are in the minority, I calculate the percentage of presumed conservative witnesses. LexisNexis provides full witness lists in its legislative histories, and includes job titles and brief summaries of testimony for each witness. To determine ideological position, I use a method developed by O'Connor and Epstein (1983) and code witnesses as liberal if they represent the interests of minorities, criminal defendants, consumers, or workers, or if they are active on civil rights, civil liberties, or environmental-protection issues. I code groups as conservative if they represent the interests of employers or businesses, or if they "espouse socially conservative or law-and-order causes" (O'Connor and Epstein 1983, 480). I also record the party affiliations of lawmakers who testify at hearings, coding Democrats as liberal and Republicans as conservative. If a witness's ideological position is ambiguous, I read committee testimony and interestgroup websites to make a coding decision. When possible, I also use the ideological scores that Groseclose and Milyo (2005) estimate for major interest groups and think tanks to double check coding decisions. Finally, there are some witnesses who do not fit easily into ideological categories—such as private citizens, foreign officials, university-affiliated researchers, and  22  members of nonpartisan groups.5 These witnesses are coded as "other." I do not include them or government employees when computing the witness percentages.6 Markups Held is a dummy dependent variable that indicates whether a bill underwent a markup (1) or not (0). I use the Thomas website and CQ Committee Reports to compile the data. Floor Rule is a dummy dependent variable that indicates whether a bill was debated under an open rule, including a modified open rule (1), or under a restricted rule (0). For most bills, coding is straightforward. But in some cases, where bills come to the floor under complex rules, I use legislative descriptions by CQ Committee Reports to make coding decisions. CQ's articles explain the various restrictions placed on the floor debate for each bill and often determine whether the rule, in effect, is modified open or modified closed. I also code any bill brought to the floor as a privileged matter, without a special rule, as 1. Such bills are typically debated in a manner similar to an open amending process.7 2.5.2 Independent Variables I construct four independent variables to test my hypotheses on one-party lawmaking. To test Hypothesis 1, I create the variable Partisan Polarization by using the difference between the Republican and Democratic Party means on the first DW-NOMINATE dimension for each Congress (McCarty, Poole, and Rosenthal 2006). This measure enables me to determine whether increasing ideological conflict between the parties leads to increases in one-party bill drafting.8  5  There are three types of organizations that fit into this category: 1) groups that have members from both government and the business community (e.g., the Independent Sector and the American Association of Museums); 2) charitable foundations that rely on both public and private support (e.g., the American Cancer Society and the American Lung Association); and 3) think tanks that receive both government and corporate funding (e.g., Woodrow Wilson International Center for Scholars and Resources for the Future). 6 I examine witness lists for over 1,400 individual hearings. 7 I am grateful to Valerie Heitshusen for this explanation of House rules. 8 An endogenous relationship exists between the legislative process a bill undergoes and the floor votes that it receives. However, a score for party ideology based on nearly all floor votes cast in a Congress should not be affected by the legislative process undergone by a specific bill.  23  To test Hypothesis 2, I construct the variable Electoral Risk by calculating the difference between the percentages of voters who identify with the minority party and with the majority party in an election year. If, for example, 40 percent of voters identify with the minority and 30 percent identify with the majority, the House majority party faces an electoral risk of 10 for a given election. I use the seven-point party identification data that the National Election Study compiles before and after elections to develop this measure. A broad measure of party support, such as the NES data, is appropriate here. Politicians are not perfectly responsive to public opinion (Jacobs and Shapiro 2000; Page 2002), and they likely develop legislative strategies based on long-term forecasting rather than on day-to-day or week-to-week shifts in the polls. After all, many bills are developed over multiple months or over both sessions of a Congress. Although derived in part from post-election responses, NES data better reflect the conditions that lawmakers expect to face in an upcoming election than other polling data. This is because data collected frequently on party ID are extremely noisy. For example, the CBS/New York Times data on party identification often show five- to eight-point swings in support for the majority party within a month or less.9 I create two dummy variables to test Hypotheses 3 and 4. I construct the variable Unified Government to denote whether party control is unified (1) or divided (0). And I create the variable Republican Control to indicate whether House Republicans (1) or House Democrats (0) are in the majority. I also include two control variables in my models. The first, Time Remaining, is the number of days from a bill's floor vote to the end of Congress. Time restrictions likely influence legislative strategy. At the start of a new Congress, the majority party has time to allow lengthy, inclusive deliberations. As the end of Congress nears, time limitations may encourage the 9  I thank Gary Jacobson for sharing these data. Poll responses are from January 1980 to June 2011.  24  majority to omit legislative steps. I assume legislators anticipate the length of time before a floor vote, so that date functions as an indicator of perceived time pressure even at earlier stages. The second control variable, News Coverage, is the number of press items a bill receives in the month before it reaches the floor of the House. News stories on Congress tend to focus on whether legislation will pass, fail, or stall (Morris and Clawson 2005). With increased media attention on a bill, and thus increased public awareness, the majority likely takes steps to limit minority-party obstructionism by truncating deliberations. Using LexisNexis, I collect all relevant news stories and editorials from the New York Times and the Washington Post. Because formal bill names are rarely included in press reports, it is not possible to use a consistent method to identify all relevant news items for each measure. The most appropriate search terms for one bill often differ significantly from another bill. For this reason, I conduct several systematic LexisNexis searches for each bill to identify different key terms in a consistent way. I search the name of the bill, the name of the bill's sponsor, and the name of the policy area during the onemonth period of bill development before the floor vote to find the two or three most common press descriptors for each legislative proposal. I then use these descriptors as my key search terms to collect all pertinent news items. 2.6 Results I use logit and OLS regression models to produce estimates relating to congressional deliberations. Because logit coefficients are difficult to interpret, I show the impact that various explanatory variables have on the dummy outcomes by computing predicted probabilities. In each case, I report the predicted change in the probability of two-party participation occurring at each legislative step as the independent variable shifts from its minimum value to its maximum value. All other variables, if continuous, are held constant at their means and, if dichotomous, are 25  held constant at their modes. For all models, standard errors are corrected for clustering because the data for Partisan Polarization and Electoral Risk have repeated values in every two-year grouping of bills.10 To begin, I conduct a logit regression for the Hearings Held dependent variable. Results are reported in the first column of Table 2.1. I do not include News Coverage in the model, because it measures press attention one month prior to a bill's vote and relevant hearings can be held several years before floor actions. As expected, both increases in polarization and electoral pressure are negatively correlated with the presence of hearings. However, there is no evidence that unified government or Republican House control have an effect on this legislative stage. Column 2 in Table 2.1 displays the predicted change in the probability of hearings being held. It shows the likelihood of bills undergoing hearings in the most polarized House (110th Congress) is nearly 27 percent lower than it is for bills developed in the least polarized House (98th Congress) in the study. As the parties' preferences diverge, the majority caucus shows less interest in the views of the minority party. Electoral pressures appear to have a similar effect. The likelihood of bills undergoing hearings when the majority faces its steepest electoral disadvantage (14 points behind) is 40 percent less than it is for bills developed when the majority enjoys its best electoral lead (14 points ahead). When the majority lags in the polls, its leaders likely rush bill deliberations in the hope of passing legislation quickly to bolster the party brand. The majority may also skip hearings to limit the minority's opportunity to attack its proposals.  10  I should note that the correlation between Hearings Held and Markups Held is quite strong (r = 0.45). If majorityparty leaders make a joint decision to hold hearings and markups (so that the errors from two models are highly correlated), then a bivariate probit regression would generate better estimates of committee sessions than separate regressions for each legislative stage. Therefore, I ran a biprobit regression to see if it produces estimates that differ significantly from the estimates produced by separate regressions. It does not. For clarity, I report the results from the separate regressions in order to discuss findings on both hearings being held and the ideological balance of committee witnesses before I discuss results on markups.  26  Table 2.1. Logit Results for the Effects of Political Conditions on the Holding of Hearings Independent Variables  Estimated Coefficient  Change in Probability  Partisan Polarization  -3.966** (1.445)  -26.6 %  Electoral Risk  -0.076* (0.029)  -40.1%  Unified Government  0.192 (0.129)  —  Republican Control  0.929 (0.587)  —  Time Remaining  -0.001 (0.001)  —  Constant  Log-pseudolikelihood Pseudo R2 N (no. of bills)  4.535*** (1.207) -72.952 0.078 156  Note: Column 1 shows logit estimates with clustered standard errors in parentheses. Column 2 shows the change in probability that hearings are held, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their mean or mode. Estimates generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.05, **p<0.01, *** p<0.001 (two-tailed tests).  Although the absence of hearings indicates one-party lawmaking, the presence of hearings does not ensure that both parties participate equally. Because committee chairs can stack hearings with witnesses who favor the majority's position, it is important to consider the 27  ideology of committee witnesses. For this reason, I conduct OLS regressions on the dependent variable Witness Balance. Table 2.2 presents the results. Model 1 includes bills for which at least one hearing was scheduled; Model 2 includes all bills in the sample, even bills for which no hearings were held (in such cases, the percentage of minority-party witnesses is 0). In this way, I examine separately the extent to which the majority party controls witness selection and the extent to which it controls the entire hearings process. As Table 2.2 shows, the effect of partisan polarization is negative and significant in both models. Model 1 shows that minority-party witnesses decline by 76 percent when the parties move apart 1 point on the liberal-conservative dimension (i.e., half the NOMINATE scale). However, Democrats and Republicans in the House do not polarize this sharply in the years of my study, and so the estimates in Models 1 and 2 overstate the drop in minority-party witnesses. To determine the real drop, I multiply the distance that the two parties polarize between 1983 and 2008 (0.36) by the Partisan Polarization coefficient. Based on the results in Model 1, I find that the proportion of minority-party witnesses drops by almost 28 percent from the 98th Congress to the 110th Congress due to polarization. There is also some evidence that increasing electoral pressure on the majority party reduces the opportunities of the minority to select committee witnesses. The estimate for Electoral Risk achieves statistical significance in Model 2, and shows that for every 1-point increase in electoral risk for the majority, there is a 1-percent decline in minority-party witnesses. Election concerns thus encourage the House majority caucus to both stack and skip hearings. The findings for Unified Government in Table 2.2 suggest that, contrary to expectations, unified control encourages rather than diminishes minority-party involvement in hearings. This finding is statistically significant in both models. Model 1, for example, shows that bills drafted  28  Table 2.2. OLS Results for the Effects of Political Conditions on the Ideological Balance of Committee Witnesses Independent Variables  Model 1  Model 2  -76.613*** (12.156)  -80.991*** (13.446)  -1.054 (0.50)  -1.288* (0.459)  Unified Government  12.510** (3.826)  10.622** (2.870)  Republican Control  21.830 (11.551)  22.614 (11.201)  Time Remaining  -0.015 (0.016)  -0.019 (0.015)  90.803*** (15.001)  88.482*** (15.391)  0.108 124  0.145 156  Partisan Polarization  Electoral Risk  Constant  R2 N (no. of bills)  Note: Clustered standard errors in parentheses. *p<0.05, **p<0.01, *** p<0.001 (two-tailed tests)  in unified periods see a 12.5 percent increase in the proportion of minority-party witnesses as opposed to bills drafted in divided periods. It appears that when the president is the "legislator in chief," his team in the House makes an effort to include the minority in early policy discussions. Next, I conduct a logit regression on the Markups Held dependent variable. Results are reported in Table 2.3, Column 1. Once again, I find support for Hypotheses 1 and 2, find evidence against Hypothesis 3, and find no evidence in support of Hypothesis 4. In other words, 29  polarization and electoral pressure appear to encourage one-party bill drafting; unified control appears to foster two-party deliberation; and Republican House control appears to have no significant effect on bill-writing processes. Column 2 in Table 2.3 displays the predicted change in the probability of committee markups being held. It shows the likelihood of bills undergoing markups in the most polarized period is 44 percent lower than it is for bills drafted in the least polarized period of the study. Moreover, Column 2 shows that bills developed when the majority faces its worst reelection prospects are almost 48 percent less likely to go to markup than bills developed when the majority faces its best electoral chances. Column 2 also shows that bills developed in unified periods are about 9 percent more likely to undergo markups than bills developed in divided periods. This result matches the findings on witness selection and suggests a rather strong trend: bipartisan lawmaking is more likely when one party controls government than when both parties share governing power. Finally, I conduct a logit regression on the Floor Rule dependent variable. Results are reported in the first column of Table 2.4. I find that partisan polarization and electoral pressure are negatively correlated with open rules. I also find, surprisingly, that Republican control has a significant and positive correlation with open floor debate. Column 2 displays the predicted change in the probability of an open rule being used. Bills developed in times of either high polarization or high electoral competition are more than 40 percent more likely to undergo a restricted floor debate than bills developed in periods of either low polarization or low electoral competition. More noteworthy, bills developed during Republican control have a 56 percent greater likelihood of undergoing open debate than bills developed during Democratic control, other things equal. That Republicans, with their reputation for disciplined lawmaking, are more  30  likely to allow minority-party amendments than Democrats suggests a reevaluation of the Gingrich-Hastert-Pelosi years is necessary.  Table 2.3. Logit Results for the Effects of Political Conditions on the Holding of Markups Independent Variables  Estimated Coefficient  Change in Probability  Partisan Polarization  -7.817*** (2.362)  -44.4%  Electoral Risk  -0.103* (0.050)  -47.8%  Unified Government  0.727* (0.367)  +8.6%  Republican Control  1.718 (1.175)  —  Time Remaining  -0.001 (0.001)  —  News Coverage  -0.003 (0.004)  —  Constant  Log-pseudolikelihood Pseudo R2 N (no. of bills)  7.978*** (1.927) -53.967 0.124 156  Note: Column 1 shows logit estimates with clustered standard errors in parentheses. Column 2 shows the change in probability that markups are held, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their mean or mode. Estimates generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.05, **p<0.01, *** p<0.001 (two-tailed tests).  31  Table 2.4. Logit Results for the Effects of Political Conditions on the Use of Open Floor Rules Independent Variables  Estimated Coefficient  Change in Probability  Partisan Polarization  -8.764*** (1.443)  -44.0%  Electoral Risk  -0.121* (0.052)  -43.1%  Unified Government  0.430 (0.386)  —  Republican Control  2.678** (0.856)  +55.9%  Time Remaining  -0.0004 (0.001)  —  News Coverage  -0.017 (0.009)  —  Constant  Log-pseudolikelihood Pseudo R2 N (no. of bills)  5.432*** (0.975) -89.198 0.170 156  Note: Column 1 shows logit estimates with clustered standard errors in parentheses. Column 2 shows the change in probability that open rules are used, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their mean or mode. Estimates generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.05, **p<0.01, *** p<0.001 (two-tailed tests).  32  2.7 Discussion A review of the results shows a preponderance of evidence in support of Hypotheses 1 and 2. At all stages of bill development, partisan polarization and electoral pressures encourage decreases in minority-party participation. The findings on polarization are not surprising. Studies demonstrate that parties have voted with increasing unity on the floor over the last three decades, rejecting each other's bills and amendments with greater frequency (Roberts and Smith 2003; Rohde 1991). It follows that, as the two parties polarize, they also prevent each other from shaping legislative content at prefloor stages.11 There are two possible reasons why majority-party leaders dominate deliberations when their party faces electoral risk. Either they seek total control of bill development in the hope that their party receives full credit for passing popular legislation, or they exclude the minority in order to design bills that appeal specifically to the majority's base. Studies show that lawmakers facing tough reelection bids often develop extreme voting records to attract partisans within their districts (Fiorina 1974; Gulati 2004). They compensate for the loss of support among median voters by trying to maximize support among one set of extreme voters. Majority-party leaders may adopt this strategy in tight races, drafting ideologically pure legislation to rally the base. The evidence for witness selection and markups disconfirms Hypothesis 3, demonstrating that two-party participation is more likely in unified periods than in divided periods. This finding is counterintuitive. A party that controls Congress as well as the White House should have little interest in collaborating with or listening to a diminished minority. It can speed legislation 11  I also substituted the Partisan Polarization variable for a variable that measures the majority's level of ideological unity (constructed using the standard deviation of the majority members' first-dimension DW-NOMINATE scores for each Congress). The results on the majority-unity variable are similar to the results on polarization. However, majority unity does not have as strong a relationship with truncated deliberations as partisan polarization does. In the Hearings Held and Witness Balance models, for example, majority unity does not reach statistical significance. This finding reinforces the theory of conditional party government, and shows that the increasing homogeneity of both parties is crucial to the rise of one-party lawmaking.  33  through the House and use this momentum, coupled with the president's bully pulpit, to pressure the Senate into fast action (Mann and Ornstein 2006, 124-125). Conversely, a House majority that faces robust opposition in the Senate or the White House should compromise on legislation, preferring to achieve partial legislative goals than see its bills filibustered or vetoed. Despite the reasoning of Hypothesis 3, there may be two related explanations of why the opposite occurs. First, because partisan supermajorities are rare in the Senate, a party that enjoys unified control likely recognizes that it must still negotiate with the minority party if legislation is to reach the president's desk. A disgruntled opposition in the Senate can otherwise easily obstruct bills. If, however, the House majority develops measures with significant minority-party involvement, it reduces the ability of the Senate minority to filibuster. Indeed, minority-party senators are probably less likely to block a House bill that has undergone extensive two-party deliberations than a House bill that has circumvented regular order. Second, a House majority party in a divided government recognizes that it has little chance of seeing its preferred policy goals enacted. It knows that any House bill eventually signed into law will undergo tremendous revisions. Anticipating such revisions, House majority-party leaders likely seize control of bill writing on important legislation to ensure these bills strongly reflect the majority's preferences. They have no incentive to compromise in the House when they know compromise is later required with the Senate and the president (Binder 2003, 99). According to former Democratic Caucus Chairman Vic Fazio, majority-party leaders try to "get the strongest bill out of the House" possible in periods of divided government, knowing that the conference committee will dilute the legislation (quoted in Epstein 2010). No evidence indicates that Republicans omit committee deliberations more readily than Democrats; it appears that both parties have barred each other equally from bill drafting.  34  Findings on floor debates, however, show that Republicans are more likely to allow open rules than Democrats. This evidence not only refutes Hypothesis 4, but it also challenges the view that Gingrich and Hastert were unusually domineering. Indeed, these findings raise two important questions. First, why is Republican control not associated with truncated committee deliberations when Gingrich and Hastert used party task forces in apparently unprecedented ways? Second, what explains the difference between how the two parties have structured floor debates? In response to the first question, Republicans used bill-control strategies that did not always involve skipping committee. For hearings, Gingrich ordered chairs to schedule Republican experts in the morning before press deadlines and to schedule Democratic witnesses later in the day. This scheduling helped Republican witnesses receive full coverage and Democratic witnesses little media attention (Herrnson 1998, 122). Markups offered majority Republicans additional ways to frustrate Democrats. Some chairs would use markups to aggravate the Democrats by ramming bills through committee on strict party-line votes (Eilperin 2006, ch. 3). If we think of committees, then, as places where the majority can bully the minority in subtle and not-so-subtle ways, it is hardly surprising that both parties have scheduled markups with approximately the same consistency. As for floor debate, Republicans used open rules more often than Democrats for three reasons. First, after four decades in the minority, Republicans took control of the House in 1995 with the stated intention of making themselves look fairer and more accommodating than the Democrats (Towell 1994). Thus, in Gingrich's first year as Speaker, 63 percent of measures were debated under an open or modified open rule (Owens 1997, 252 n19). While Republican leaders eventually placed greater limits on floor activity, they never restricted debate as severely as  35  Democrats.12 Second, House Republican leaders used open rules frequently to weaken the legislative powers of committees and in turn bolster their own powers. Open rules allowed them to make important changes to committee bills by offering amendments on the floor that were often supported by loyal junior members (Owens 1997, 252). Third, House Republicans fostered open debate because they tended to face a lower risk of being rolled than Democrats. Republicans organized a very strong whip system and enjoyed extraordinarily high levels of party unity on major floor votes during the Gingrich years (Sinclair 1999, 439; Strahan 2007, 154), and they maintained this strong cohesion during the Hastert period (Evans and Grady 2009). The lack of a comparably high level of party unity perhaps forced House Democratic leaders to rely heavily on closed rules. 2.8 Conclusion In this chapter, I used data on prefloor legislative actions to show that the House majority party controls bill drafting for three reasons. It does not want to collaborate with an ideologically incompatible minority; it faces intense electoral competition; and it wants to report the strongest possible bill to boost its bargaining position at a later point in the legislative process. Combined, these results indicate that the majority takes control of bill deliberations when its majority status is under threat or its policy goals face strong opposition. These findings also suggest that even if the parties returned to the ideological center, the majority would still often exclude the minority from bill drafting. After all, electoral pressures and divided government are common features of U.S. politics. I agree with Lee that some core motivations for partisan conflict and partisan strategies go "beyond ideology" (2009).  12  Under Pelosi, House Democrats used a record number of closed rules, restricting debate on 88 percent of all bills in the 110th Congress (Binder et al. 2009, 39).  36  Although this chapter identifies key conditions that encourage one-party lawmaking, many questions concerning deliberative partisanship remain unanswered. Are certain types of legislation, such as tax bills or social-welfare bills, particularly prone to one-party control? Do the ideological positions of committee chairs or the ideological makeup of the committees themselves affect the likelihood of hearings and markups being held? Do significant increases in committee sessions necessarily lead to substantive policy concessions by the majority to the minority? When drafting bills behind closed doors, do leaders routinely overlook vital considerations or can they conduct thorough policy evaluation without the minority party’s assistance? I consider these questions in the next two chapters.  37  3.Bill Variation and Committee Leadership: Additional Explanations of OneParty Control  3.1 Introduction As the previous chapter shows, political factors can encourage significant reductions in formal legislative deliberations. However, theories of congressional parties suggest that there are two additional reasons why the majority excludes the minority from bill development in the House. First, there may be predictable areas of dispute between the parties that encourage the use of tough partisan strategies by the majority caucus. In particular, certain types of bills may be especially likely to undergo party-dominated procedures. Leaders of the majority caucus may prevent or limit the minority party from participating in bill discussions when specific policy issues are on the agenda. Second, certain types of committees may be prone to policy conflict and often excluded from the legislative process. According to the theory of conditional party government, certain issues foster unusually high levels of conflict between the parties and encourage the majority caucus to control bill development in order to pass its preferred policies (Aldrich and Rohde 1997, 2000a, 2000b, 2009). In fact, "the theory of conditional party government is applicable to only a specific subset of issues" and assumes that numerous policy questions do not produce significant partisan disputes (Aldrich and Rohde 2000b, 3). Remarkably, there has been little effort to identify this "specific subset of issues" critical to understanding the sources of partisan disagreement in Congress. Researchers have largely overlooked whether Republicans and Democrats are less likely to collaborate on bill writing when certain policy issues arise. Indeed, not since the early  38  1970s have congressional scholars even viewed policy substance as an important explanatory variable (Katznelson and Lapinski 2006, 98). Similarly, the theory of conditional party government posits that as legislative parties see their power increase, standing committees see their power decrease (Aldrich and Rohde 2000b, 2009). In era of partisan polarization, committees do the bidding of the House majority party or else find themselves bypassed from the decision-making process. However, the theory does not consider, and research has not yet identified, whether the majority excludes some committees more readily than others from deliberations. Numerous studies have examined committee rosters to see whether the preferences of congressional panels diverge from the preferences of the parent chamber and, if so, how this divergence affects bill formulation (e.g., Adler 2000; Adler and Lapinski 1997; Hall and Grofman 1990; Krehbiel 1990). However, the research has not examined whether committee rosters or other committee features directly affect whether panels participate in bill drafting. Because both policy substance and committee configurations are potential sources of partisan conflict, this chapter seeks to identify both the types of bills and the types of House panels prone to one-party deliberations. In doing so, this chapter contributes to the theory of conditional party government in two ways. It details the central but so far undeveloped policy story of the theory, and it explains variation in committee control over bill deliberations that the theory fails to consider. To conduct this research, I collect data on hearings and markups for major bills developed between 1983 and 2008 to create multiple measures of one- and two-party procedures. My assumption here, as in Chapter 2, is that the more opportunities the minority party has to participate in bill deliberations, the more likely it is to shape legislative content. I  39  then determine whether certain bills undergo high levels of party-directed lawmaking and whether some committees tend to be omitted from bill drafting. In the analysis, I focus especially on the types of bills that both parties use to build and maintain their electoral brands. These bills cover policy issues in which both parties hold strong but polar opposite positions. For example, Republicans tend to favor tax cuts for and Democrats tend to favor tax hikes on high-income earners. My findings show that bills designed to promote the majority party's brand−such as tax bills−tend to undergo truncated deliberations. Because members of the majority caucus hold fixed, long-standing positions on these policy issues, they have little need to discuss the details of these bills and, thus, they allow leaders to reduce the committee's involvement in bill formulation. I also find that bills that divide the majority caucus at the prefloor stage undergo mixed levels of one- and two-party deliberation. The majority party allows extensive hearings on these bills, on one hand, but tends to circumvent markups, on the other. I suggest that majority-party leaders allow extensive deliberations initially in order to explore and identify proposals that possibly reduce intraparty divisions. Once they settle on a proposal, however, they omit formal bill-drafting sessions to prevent committee members, especially minority-party members, from offering amendments that could undermine these already politically fragile measures. I also determine that, as committee chairs and ranking minority members hold increasingly extreme preferences, their panels hold fewer sessions. I proceed as follows in this chapter. First, I review research on congressional voting behavior and committee rosters to identify major findings on legislative conflict and cooperation. After discussing the limitations of this literature, I present testable hypotheses on policy issues and committee features that likely foster one-party lawmaking. Next, I discuss my methods of data collection and data coding, and I present the major findings of the statistical analysis. I  40  conclude by discussing the implications of these results and, in particular, by reassessing the role that markups play in the legislative process. 3.2 Agreement and Disagreement between the Parties Several different literatures examine the level of policy cooperation in Congress. From this diverse body of research, two sets of empirical questions emerge. First, what are the key factors that divide legislators on bills? How prevalent is partisan disagreement in Congress? Second, how do committee rosters affect committee activity? Are some committees stacked with ideologically extreme members who strive to produce bills that diverge from the preferences of the floor median? Or are committees organized to produce bills that reflect the broad demands of the parent chamber? As noted earlier, congressional scholars have not viewed policy substance as an important independent variable. However, in trying to determine the sources of legislative disagreement, scholars have identified several types of bills that tend to undergo contentious votes. Examining these contentious bills, they consider whether various characteristics of legislators−such as race, religion, gender, and seniority−affect vote choice. In several studies, they find that ideology and party affiliation have the greatest impact on members' voting behavior. Scholars demonstrate that roll calls on civil-rights legislation (Nye 1991, 1993), abortion bills (Oldmixon 2002), and environmental measures (Shipan and Lowry 2001), among others, tend to split overwhelmingly along party lines in Congress. Taken together, these studies produce a list of policy issues that appear to divide the congressional parties over several decades. None of these studies uses policy substance as an explanatory variable but, collectively, they provide evidence that different issues affect the level of legislative collaboration between the parties. Missing from this research, however, is a compelling explanation of why issues that 41  otherwise appear to have little in common−such as pollution control and reproductive health−consistently divide the parties. A second camp of scholars argues that lawmakers try to build large, nearly "universal" coalitions in which most members of Congress share both the legislative rewards (Shepsle and Weingast 1981; Weingast 1979) and the legislative blame (Balla et al. 2002). Initially, this literature focused on distributive policies to show that consensus rather than conflict is the norm in Congress. Recent work, however, has demonstrated that a high level of agreement exists on the vast majority of legislative measures. For example, Carson, Finocchiaro, and Rohde (2010) examine House voting over the last three decades, and find that about two-thirds of all bills receive high levels of support from both parties. Similarly, Lee (2009, ch5) shows that when partisan clashes do occur, they are often sparked not by moral or ideological questions but, rather, by concerns over the majority’s governing performance. Measures that address waste, fraud, and corruption undergo much higher levels of party-line voting than bills on abortion, affirmative action, or taxes. Differences in research findings on legislative conflict and consensus stem mostly from an important difference in sample selection. On one side, scholars interested in what factors divide legislators intentionally select contentious issues, such as abortion and the environment (e.g., Oldmixon 2002; Shipan and Lowry 200). They ignore legislation that encourages a high degree of compromise and, in the process, likely exaggerate the level of partisan disagreement in Congress. On the other side, scholars interested in how Congress generally manages bills use enormous samples of legislative votes on a broad array of policy issues (e.g., Lee 2009). They often fail to distinguish between minor and major measures and, in the process, likely understate  42  the level of conflict in Congress. After all, most floor votes are on "trivial" matters and provoke little if any disagreement (Arnold 1990, 269). One limitation present in both sets of studies is the use of roll-call data as the primary source of evidence for either partisan conflict or bipartisan collaboration. Floor votes represent a rather narrow part of legislative deliberation; many of the most crucial decisions over the legislative agenda and bill content occur at the prefloor stage (Cox and McCubbins 2005; Hall 1996; Sinclair 2002). Granted, Carson, Finocchiaro, and Rohde (2010) do look beyond floor data to determine whether certain types of bills are more likely to foster conflict or consensus between the parties. They combine floor votes with committee votes to measure the level of twoparty support for each bill in their sample. Although their study involves a broader examination of legislative decision-making than most research on the subject, it still has the same limitations as studies that focus exclusively on floor action. Just as most floor votes are on trivial matters, so are most committee votes. In fact, many amendments offered in markup are on minor technical issues, not substantive ones. High agreement on these technical amendments dilutes any disagreement on substantive amendments. Thus, the aggregate measure developed by Carson and his colleagues is likely biased toward two-party support and does not indicate the extent to which Democrats and Republicans work together on legislation.13 The other limitation with roll-call data is that they offer little indication of the extent to which the two parties collaborate in deliberations. A floor vote does not show whether the minority has had significant opportunity to participate in hearings and markups—the central venues in Congress for information gathering and legislative discussion. Because scholars have  13  To compound problems, Carson, Finocchiaro, and Rohde (2010) code all voice votes as unanimous votes. This coding decision biases their measure greatly toward consensus because most technical amendments are handled by voice vote. In markup, it is not unusual to have a dozen or more voice votes over minor language changes in a bill and only one or two recorded votes on major substantive amendments.  43  not examined the quantity or scope of committee deliberations for various policy issues, they have not determined which types of legislation are drafted primarily by the majority party and which types are developed by both parties. This research gap has encouraged some scholars to call for renewed interest in examining how, and determining why, Congress handles various policy issues differently (see e.g., Katznelson and Lapinski 2006). Questions over agreement and disagreement in Congress have also encouraged scholars to study the makeup of committees. A large literature examines the composition of House panels to see whether they are somehow unrepresentative of the parent chamber (e.g., Groseclose 1994; Krehbiel 1990, 1991, 1993). Committees with unrepresentative preferences likely produce bills that many legislators in the parent chamber oppose, increasing the likelihood of contentious floor votes on amendments designed to change committee bills. Some of this research has sought to determine whether majority-party leaders stack panels with ideologically extreme members to ensure partisan victories in committee rooms. So far, however, scholars have failed to identify a significant number of stacked panels. As Groseclose (1994) argues, a lack of statistical evidence has made it difficult to reject the possibility that committee assignments are handed out randomly (see also Espino and Franz 2004). If committees are not stacked with extreme members, they are less likely to work as agents of the majority party and more likely to operate as providers of specialized information to the parent chamber (Krehbiel 1991, 1993). In this capacity, committees work in a generally bipartisan manner and produce bills that the floor majority accepts without having to make significant changes. A counterargument to these findings is that majority-party leaders do not need to load committees with likeminded ideologues as long as committee chairs maintain close, cooperative relations with the Speaker. Two reasons support this point. First, on any given measure, only a  44  narrow set of panel members led by the chair (or subcommittee chair) has direct interest and involvement in a bill's development (Evans 2011; Hall 1996). Because so few lawmakers pay attention to bill drafting in any given case, the range of ideal points for committee members likely explains little about a bill's content or a bill's path to the floor (Evans 2011, 415). For this reason, the ideological scores of committee leaders may tell us more about a committee's activity and output than the median score of the panel itself. Second, with control over a large staff and the committee's agenda, chairs have the resources to ensure that their committees often produce legislation that majority-party leaders support (Deering and Smith 1997, 131-132; Oleszek 1996, 106-108). And if they do not have the ability to produce an acceptable bill in committee, they can coordinate with party leaders to draft the bill behind closed doors. However, an almost-total lack of recent large-N research on committee chairs makes it difficult to go beyond conjecture here. As Evans notes, "the analysis of leadership in committee has all but atrophied" (2011, 419). Because committee chairs are central to policy formulation, variation in their ideological preferences may encourage variation in one- and two-party bill development. By ignoring committee leadership, scholars are likely ignoring a key factor that explains important fluctuations in committee deliberations. 3.3 Explaining Policy Control According to theories of congressional parties, a majority caucus with relatively homogenous preferences will grant its leaders significant control over lawmaking (Aldrich and Rohde 1997, 2000a, 2000b; Sinclair 1995, 1999). Party theories assume that leaders follow the demands of their caucus and rarely make legislative decisions that rank-and-file members oppose (Sinclair 1999). Applied to House deliberations, these theories suggest that bills unifying the majority will undergo high levels of one-party control and bills dividing the majority will 45  undergo low levels of one-party control. Yet, it is unlikely that majority-party leaders refrain from intervening in policy matters that provoke internal disputes within their caucus. To allow serious disputes to persist would likely tarnish the party and damage the credibility of majorityparty leaders. As a result, leaders may have strong incentives to manage deliberations not only on bills that unify their caucus but also on bills that divide their caucus. Indeed, on measures that unify the majority, leaders may seek to circumvent committees because the party does not need to engage in lengthy discussions to determine its core policy goals and its preferred policy responses. Leaders can write bills behind closed doors in order to develop legislation efficiently and to minimize the minority's opportunity to criticize or delay the bill. On measures that divide the majority, leaders may seek to circumvent committees as a way of limiting intraparty conflict. Because committees often struggle to reach policy agreement on issues that divide its majority-party members, leaders may seize control of bill writing in these cases to produce legislation within a reasonable period and to avoid having bills changed by a coalition with minority-party members (Sinclair 1997, 218). If majority-party leaders have reason to dominate decision-making on bills that unify as well as on bills that divide their caucus, they appear to have incentives to control legislative deliberations on all or nearly all important measures. Yet, as Sinclair (2006, 148) shows, they intervene in only about half of all priority bills. Even in the partisan Congress, regular-order proceedings are frequently maintained. We thus have what appear to be two competing expectations on party-directed lawmaking: either majority-party leaders control bills that unify their caucus or they control bills that divide their caucus. However, these two expectations can be compatible if there are two causal mechanisms by which intraparty agreement operates. When disagreement within the majority caucus is low,  46  demand by rank-and-file members to participate in bill drafting may be low in turn; and when disagreement within the majority caucus is high, the cost of bill participation by rank-and-file members may be high, as well. These two extreme conditions provide incentives for leaders to intervene. Bills that fall somewhere between these two extremes of agreement and disagreement likely undergo regular-order proceedings, because majority-party leaders have neither the support nor the need to bypass committee. Thus, to hypothesize that majority-party leaders have reason to direct lawmaking on both bills that unify and divide their caucus does not imply that they have reason to direct lawmaking on all important bills. 3.3.1 Party-Brand Hypothesis Congressional parties develop strong national brands to provide voters credible guarantees of what their candidates will support and oppose in office (Grynaviski 2010). The party brands are, in effect, ideological reputations based on each party’s most important policy goals. Party leaders are the ones primarily responsible for maintaining and defending the ideological reputation of their respective parties. In the House, majority-party leaders uphold the brand by controlling which bills reach the floor for a vote (Cox and McCubbins 1993, 2005) and by ensuring that the content of these bills reflects the party’s main preferences (Aldrich and Rohde 1997, 2000a, 2009; Sinclair 1997, 2006). Majority-party leaders may draft legislation themselves (or at least place significant limits on formal deliberations) when the bill involves a policy objective that reinforces their party’s reputation. The majority has little incentive to engage in lengthy information gathering and policy discussion, because it already has a wellestablished, clearly fixed preference. On core policy issues, the majority's main concern is to advance its brand, not to consider new options.  47  Because voters, especially low-information voters, rely on these policy reputations to help determine which party best reflects their values and preferences, members of Congress have a strong electoral incentive to help maintain the brand by delegating legislative duties to leaders (Grynaviski 2010, ch. 2). Confusion over what the party stands for will diminish its reputation and harm its electoral chances. Parties thus operate as "long-term policy coalitions" that work to enact and defend legislation over several Congresses and sometimes over several generations (Smith 2007, 38). Again, rank-and-file members understand that their electoral success is tied directly to important policy gains, and they therefore allow party leaders to coordinate legislative strategies when important policy goals are at stake. Republicans and Democrats have developed party brands primarily by taking opposing positions on the same major policy issues (Aldrich and Rohde 2000b). Tax policy, for example, is crucial to both Republican and Democratic reputations, because the parties define themselves by taking divergent stands in this policy area (Grynaviski 2010, 183-185). Although the parties have staked out predictable positions on key policy matters, Republicans and Democrats do not always act as perfectly unified groups in all instances. Issue evolution has occurred in some cases, and the parties have sometimes adopted positions that they had previously opposed (see e.g., Karol 2009). But many of the most dramatic shifts in policy positions occurred in the 1970s, when the two parties were in the process of sorting themselves along ideological lines (Han and Brady 2007). Since then, the parties have only strengthened their ideological reputations, staking out clear policy positions to appeal to voters. Scholars have not developed a comprehensive list of policy positions that define each party brand; however, some political scientists argue that legislative proposals on taxes, social-welfare benefits, and moral issues frequently polarize the two parties.  48  Indeed, from the time of Ronald Reagan’s presidency onward, Republicans and Democrats have staked out opposing positions on tax policy (Bartels 2008, ch. 2; Rudder 2009). While Republicans tend to advocate tax cuts for affluent citizens and businesses, Democrats tend to support tax cuts for middle-class and working-class voters. What is more, as lawmakers have hardened their partisan positions on tax policy over the last three decades, the public’s views on tax issues have increasingly polarized in turn. The most politically attentive voters now hold ideologically extreme views on fiscal issues (Abramowitz 2010; Abramowitz and Saunders 2008). Consequently, a growing number of lawmakers in both parties maintain consistent positions on tax rates in an effort to satisfy their core supporters. This "race for the base," as Hacker and Pierson call it, ensures that bipartisan compromise on tax bills is exceedingly rare (2005, ch 4). That floor votes on tax measures have polarized along party lines since the early 1960s confirms this view (Cox and Poole 2002). Given the multi-decade disagreement over tax rates between the two parties, majority-party leaders may control deliberations on tax bills to ensure that these measures reflect the party's core preferences and strengthen the party's brand. McCarty argues that partisan conflicts over legislation are "most pronounced in the area of social policy" (2007, 240). He examines the minimum wage and welfare programs, and shows that benefits for the poor have eroded significantly over the last four decades. Because the minimum wage has not been indexed to inflation, congressional Republicans have successfully decreased the rate by using legislative tools to block Democratic measures designed to raise wages. At the same time, congressional Republicans have advocated and passed cuts to welfare programs that most congressional Democrats have opposed. Bartels et al. concur, arguing that "the 'have-nots' have benefited when the Democrats have won" control of Congress, because the party has reliably pushed legislation that aids the poor and enhances the social-safety net (2005,  49  104-105). Republicans, by comparison, have sought to weaken the safety net by cutting and, at times, attempting to dismantle antipoverty programs. Because both parties have very different social-policy goals, they have strong reasons to exclude each other from bill development when they draft legislation that affects social-welfare benefits. In addition to taxes and social policies, the two parties consistently disagree over a broad collection of moral-values issues that comprise the "culture war" (Layman and Carmines 1997; Oldmixon 2002). Republicans commonly support reductions in abortion procedures, while Democrats typically support expansions in abortion access (Adams 1997; Brady and Schwartz 1995); Republicans strongly support bans on gay marriage and other restrictions on homosexuals, while Democrats strongly support gay-rights legislation (Lubin 2005); Republicans tend to support weakening gun-control statutes, while Democrats tend to support strengthening gun-control laws (Lindaman and Haider-Markel 2002, 101). In short, a central component of the Republican brand is linked to conservative moral values, whereas a central component of the Democratic brand is connected to liberal moral values. Furthermore, publicopinion data reveal that each party's base strongly supports their parties on values issues, giving both Republican and Democratic lawmakers strong electoral incentives to advocate divergent policies (Bafumi and Shapiro 2009; Lindaman and Haider-Markel 2002). Because of the predictable partisan split on culture-war issues, majority-party leaders have incentives to control deliberations on moral-values bills to ensure this legislation advances the party brand. In sum, the majority caucus will exercise greater control over deliberations on bills that promote its party brand than on bills that do not promote its party brand. Republicans and Democrats define their brands typically by taking opposing positions on the same set of issues.  50  Therefore, party-brand bills will involve the same set of policies for each party (although each party will advocate very different policy solutions). Hypothesis 1: Bills that affect tax rates, social benefits, and moral issues will undergo greater levels of one-party deliberations than other types of bills. 3.3.2 Intraparty-Conflict Hypothesis Majority-party leaders may also have incentives to impose control on bill deliberations when their caucus is divided over an important issue. Leaders can potentially minimize party disputes by circumventing committees and restricting floor debate to ensure no divisive amendments are offered during a measure's development. In all likelihood, majority-party leaders try to avoid taking up legislation on policy matters that they know will split their caucus. However, public pressure, national crises, or other exogenous factors can sometimes force leaders to include bills in the legislative agenda that will divide the majority party. In such cases, the majority caucus may feel pressured to develop and support a bill located far from the party's median. Depending on the direction of policy change, either moderate or extreme members in the majority party may oppose the policy shift. Party divisions are also likely to occur on complex bills that encompass multiple policy issues simultaneously. When a bill attempts to shift the status quo in multiple dimensions, and not just one dimension, it attracts a mix of liberal and conservative supporters as well as a mix of liberal and conservative opponents (Jeong 2008). These liberal-conservative coalitions are not stable, however, because members switch their support for a bill whenever a more attractive proposal is offered on one of the bill's multiple dimensions. In fact, minority-party lawmakers often encourage majority-party defections by offering amendments that appeal specifically to pivotal legislators. Immigration bills illustrate the way in which complex measures produce 51  unstable coalitions. These types of bills spark conflicts within both parties because they simultaneously involve racial or ethnic concerns, labor policy, business interests, and welfare programs. Thus, over the last three decades, immigration reforms have encouraged pro-union Democrats to collaborate with socially conservative Republicans, and socially liberal Democrats to ally with pro-business Republicans (Jeong et al. 2011). Whether bills that divide the majority caucus are one-dimensional or multidimensional, they are vulnerable to minority-party amendments designed to exploit majority disunity. Any bills that threaten to split the majority caucus may encourage majority-party leaders to circumvent and restrict formal deliberations. By doing so, they prevent the minority party from offering destructive amendments designed to gain the support of wavering majority-party members. Leaders may also control deliberations, especially in the case of complex bills, to reduce policy instability created by ever-shifting coalitions that change measures in unpredictable ways. Hypothesis 2: Bills that divide the majority caucus prior to formal legislative deliberations will undergo greater levels of one-party control during the prefloor stage than bills that do not initially divide the majority caucus. 3.3.3 Committee-Leadership Hypotheses Disagreements between the parties and within the parties over policy often reveal themselves in committee. Thus, if majority-party leaders have incentives to control the development of certain types of bills in order to promote and protect their party's reputations, they also have incentives to control committees where the bulk of bill-drafting activity occurs. The question is: Which committees are most likely to see their deliberative roles reduced by  52  majority-party leaders? Or put another way, what factors will encourage chairs to reduce the formal involvement of their committees in the development of priority legislation? The theory of conditional party government provides a succinct explanation for partydirected lawmaking: as each party takes on increasingly homogenous and extreme views, members of the majority caucus increasingly delegate legislative control to their leaders (Aldrich and Rohde 1997, 2000a, 2000b, 2009). While the theory aims to explain coordination, or the lack of coordination, between leaders and rank-and-file members, the basic logic of the theory can also explain when committee chairs may decrease their panel's involvement in bill development and allow greater bill control by majority-party leaders. Chairs and majority-party leaders tend to agree with each other on crucial policy goals and usually work collaboratively on important bills (Aldrich and Rohde 2009; Sinclair 2005; Stewart 2012). After all, rank-and-file members expect both of them to maintain the party's brand and demonstrate the party's capacity to legislate effectively. For this reason, significant policy disputes between chairs and majority-party leaders occur infrequently in the House, especially in recent years (Aldrich and Rohde 2009, 235). The need for extensive committee deliberations likely diminishes if strong policy agreement exists between party leaders and committee leaders. That said, not all committee chairs have policy preferences that match their leaders' positions. Party leaders tend to be ideologically extreme (Grofman, Koetzle, and McGann 2002; McGann, Grofman, and Koetzle 2002), and relatively moderate chairs may disagree in key instances with their leaders over the legislative agenda. Moderate chairs may even seek to dilute policy proposals by holding extensive committee deliberation and allowing significant opportunity for amendments to be considered and adopted. Moreover, once a bill has been referred to a committee, the leadership does not have the ability to stop the committee chair  53  from scheduling lengthy and extensive sessions on the bill. Thus, we should expect extreme chairs taking steps to reduce formal deliberations and moderate chairs taking steps to expand panel sessions. Hypothesis 3: Formal committee deliberations will decrease for bills as chairs on the committees of jurisdiction become increasingly extreme. The ideological positions of ranking minority members on committees can also possibly affect the scope and quantity of bill deliberations. Ranking members obviously cannot exercise control over the agenda the way that committee chairs can. However, they can use their seniority and visibility to draw attention to bills under consideration in formal deliberative settings. A ranking member can use committee sessions to criticize bills in an effort to sway public sentiment against the measure. They can also propose markup amendments designed to change the bill significantly. In fact, the minority party has very strong electoral incentives to oppose and hinder the majority party (Hibbing and Tiritilli 2000; Jones 2010), and extreme minorityparty leaders are much more likely to oppose and attack than to bargain and collaborate. Facing an extreme partisan rival, neither the committee chair nor the rest of the majority party has incentives to hold lengthy deliberations and leave key bills open to extended criticism and destructive amendments. Hypothesis 4: Formal committee deliberations will decrease for bills as ranking minority members on the committees of jurisdiction become increasingly extreme. 3.4 Data and Coding To conduct this investigation on party-dominated lawmaking, I use the same sample of 156 major bills that I used in Chapter 2. This sample includes major bills developed between 1983 and 2008 in the House. In the second chapter, I examined whether various political factors 54  lead to increases in one-party deliberations. The findings confirm that partisan polarization and electoral competition, among other factors, encourage the majority party to exclude its partisan opponents from bill development. However, the models explain only some of the variance in the dependent variables, suggesting that other important factors affect deliberative procedures. Indeed, the three logit models in the second chapter—which identify the determinants of hearings being held, markups being scheduled, and floor rules being open—all have pseudo Rsquare values below 0.20. Because the interpretation of pseudo R-square values differs from the interpretation of R-square values, we cannot say that each model explains less than 20 percent of the variation in the dependent variables.14 However, pseudo-R square values below 0.20 are relatively low and suggest that variables omitted from the models account for some variation in the dependent outcomes.15 Further examination of the data on major bills is therefore warranted. 3.4.1 Dependent Variables I use four dependent variables to determine the level of minority-party participation at the two main stages of prefloor deliberations. For the first legislative stage, I use the dummy variable Hearings Held to indicate whether a bill has undergone committee hearings (1) or not (0). I also use the variable Number of Hearings to examine the relationship between the type of bill and the quantity of deliberations. This variable is constructed using the number of scheduled hearings held for each measure. For the second legislative stage, I use the dummy variable Markups Held to denote whether a bill has undergone committee markups (1) or not (0). I also create the variable Number of Markups to 14  The pseudo-R square simply provides a point of comparison between the log-likelihood of the full model and the log-likelihood of the restricted model (Bowen and Wiersema 2004, 99). 15 Two OLS models in Chapter 2 identify the determinants of hearings having ideologically balanced witness lists. Both models have low R-square values (0.11 and 0.14), thus indicating that each model accounts for less than 15 percent of the variance in the committee-witness data. While I do not examine the balance of witnesses in this chapter, the relatively low R-square values reinforce that major political factors account for only a portion of the variance.  55  explore the relationship between the type of bill and the quantity of bill-writing sessions. This variable is developed using the number of scheduled markups for each bill. As in Chapter 2, all data on hearings are from LexisNexis Congressional and all data on markups are from CQ.com and the Thomas website. 3.4.2 Bill Variables To test Hypothesis 1, I construct three dummy variables that each denote whether a bill concerns one of the issues that scholars have identified as central to the party's brand. I indicate whether a bill changes taxes (1) or not (0); whether a bill affects social benefits (1) or not (0); and whether a bill involves moral issues or not. Coding tax bills is straightforward.16 For socialwelfare measures, I include any bill that affects minimum-wage rates, welfare benefits, lowincome housing subsidies, food stamps, and other food-assistance programs; for moral-values legislation, I include any bill that involves abortion, gun rights, gay rights, pornography, or school prayer. I use summary descriptions of the bills in CQ Almanac to make coding decisions. To test Hypothesis 2, I construct the dummy variable Majority-Conflict Bills to denote whether a legislative proposal divides the majority party (1) or not (0) before formal deliberations begin. To identify these bills, I examine the earliest CQ Weekly articles on each bill in my sample to see whether party divisions are reported at the prefloor stage. Although floor votes provide an accurate measure of party cohesion on legislation, I use prefloor reports of majority-party disagreement to avoid concerns over endogeneity. Any reported divisions prior to formal bill development are likely the result of the specific policy matters under consideration, and not the result of the deliberative process that the bills undergo. 16  Some tax bills affect only minor or narrow provisions; we would not expect to observe a brand effect in such cases. However, major tax bills virtually always affect general rates and distribution, and the tax bills in my sample are all major pieces of legislation designed to make important changes to tax policy.  56  CQ often discusses sensitive or difficult political considerations that majority-party leaders face before taking up a bill. For example, on a 1986 proposal to cut defense spending, CQ reported that the issue "is likely to illuminate differences among majority party Democrats that until now have been aired in private" (Calmes 1985). On a 1995 plan to reform lobbying, CQ noted, "A Republican rebellion over proposed limits on the gifts that House members can accept threatens to produce results unacceptable to the lawmakers who forced the issue onto the agenda in the first place" (Salant 1995). And on a 2005 proposal to tighten rules on driver's licenses, CQ noted that a proposal supported by some House Republican leaders "is likely to plunge Congress into a fierce debate over immigration policy and expose a rift in the GOP on the issue" (Dlouhy 2005). I also include two control variables in the models to account for other kinds of legislation that potentially affect deliberations. First, I construct the dummy variable Discretionary Bills to denote whether bills are optional (1) or mandatory (0). All bills that do not fall under the mustpass category are coded as discretionary; examples include campaign-finance reforms, expansions in prescription-drug benefits, and changes to federal criminal sentencing rules. Mandatory bills, by contrast, are appropriations bills that require annual congressional approval or are reauthorization bills (Adler and Wilkerson 2009).17 Optional bills are more likely to skip legislative steps than mandatory bills, because majority-party leaders have greater flexibility in intervening on bills that are not regularly part of the legislative agenda.18  17  The mandatory bills in the sample are, for the most part, appropriations bills. There are no cases like proposed extensions of the Bush tax cuts, which undo the sunset provisions in the original bills and which are significantly different from must-pass legislation. If extensions of the Bush tax cuts were in the sample, however, I would code them as mandatory. This is because, as with a reauthorization bill, the absence of legislative action would result in major policy change−in this case, significant tax increases. 18 Optional bills do not necessarily unify the majority, because they can involve policy issues that arise unexpectedly and that the party feels pressured to address.  57  Second, I create the dummy variable Presidential Proposal to denote whether a bill has been proposed by a president (1) or not (0). Because of a president's central role in the American political system, a White House proposal likely garners significant public attention and encourages Congress to hold lengthy deliberations. I code a bill as a presidential proposal if CQ identifies it as such either in its annual study of voting patterns and presidential success or in its legislative summaries. 3.4.3 Committee Variables To test Hypothesis 3, I create the variable Chair Extremity by using the first-dimension DW-NOMINATE score for the committee chair of jurisdiction for each bill (Poole and Rosenthal 2007). I convert all NOMINATE data to positive values, so that relatively extreme liberals and conservatives have similar scores.19 When multiple committees report the same bill, I collect the NOMINATE data for all relevant chairs and calculate their mean score. To identify relevant chairs, I use the lists of committee assignments compiled by Nelson (2005) and by Stewart and Woon (2009). To test Hypothesis 4, I construct the variable Ranking Member Extremity by using the first-dimension DW-NOMINATE score for the ranking minority-party member on the committee of jurisdiction for each bill. Again, I convert all NOMINATE scores to positive values and I use the mean score of all relevant ranking members when multiple committees report a bill.20  19  NOMINATE scores run from -1 (most liberal) to +1 (most conservative); 0 represents the ideological midpoint. Using NOMINATE data, I could also calculate the distance between the chair and the ranking member and use their distance from each other as a key explanatory variable. I have decided to examine their positions separately, however, because the relative extremity of one leader may have a greater effect on legislative procedures than the relative extremity of the other leader. For example, an extreme chair may omit hearings regardless of the ranking member's preferences. Furthermore, Chair Extremity and Ranking Member Extremity are not strongly correlated (r = 0.278). 20  58  When a bill does not undergo hearings or markups, it can be difficult to identify the relevant committee and leaders that should have conducted formal deliberations. In these cases, I perform a multistep, systematic search to determine the omitted committee. First, I examine the record of referral for each bill on the Thomas website. When a bill is referred to only one committee and the committee deliberations are skipped, I consider it the omitted panel. However, when a bill is referred to multiple committees, I check to see if the bill's sponsor is a member of any of these panels. If so, I code the sponsor's committee as the omitted committee. Sometimes, however, a bill undergoes multiple referrals, skips all prefloor deliberations, and is sponsored by the Speaker or a member of the Rules Committee. In these cases, I see if CQ identifies the omitted committee in its reporting. Occasionally, it provides this information. For example, on a community-development bill drafted in 2000, CQ reports, "To get the bill to the floor, House Republican leaders bypassed the Ways and Means Committee, where Chairman Bill Archer, RTexas, was holding out" (Nather 2000). And if CQ in these cases does not explicitly name the omitted committee, I then see which minority-party members publicly objected to the committees being skipped. I examine both CQ articles and floor debates for this information. I then code the committee that these minority members sit on as the excluded panel.21 I use separate models to estimate the effects of bill type and committee leadership on deliberations, because each model examines distinct sets of hypotheses derived from the theory of conditional party government. For this reason, I construct two separate control variables for the committee models. The first control variable, Majority Committee Margin, measures the difference in the number of majority- and minority-party members on the committees of  21  For example, Speaker Newt Gingrich sponsored the Patient Protection Act of 1998 and brought it directly to the floor after drafting it in closed-door sessions. CQ reported that Democrats on the House Ways and Means Committee objected to Gingrich's total control over the bill's content and legislative path (Carey and Kirchhoff 1998).  59  jurisdiction for each bill. For example, if the majority has 23 members on the relevant committee and the minority has 13 members, then the majority's margin is 10. Presumably, as the majority's margin on a committee decreases, the risk it faces of losing crucial votes in markup because of defections increases. Thus, the smaller the majority's margin, the more likely it is to exclude legislative steps. Data on party ratios are collected from Tong (2009). The second control variable, Number of Referrals, measures the number of committees that review and work on a bill. Speakers often refer a bill to multiple committees to encourage policy innovation by involving a large and diverse set of lawmakers in bill development (Davidson, Oleszek, and Kephart 1988). I include this variable in the models to account for increases in hearings and markups that usually occur whenever a Speaker refers a bill to more than one panel. Even if all committee chairs are ideologically extreme, a multiple referral will likely ensure that a bill undergoes considerable deliberations. I use the Thomas website to identify all committee referrals. 3.5 Results I use both binary and count data to measure deliberative processes in the House. For this reason, I use logit and negative binomial models to produce estimates of one- and two-party lawmaking. Because both models generate coefficients that are not directly interpretable, I determine the effect that the explanatory variables have on the dependent outcomes by computing either predicted probabilities or expected values. For the logit models, I calculate the predicted change in the probability of hearings or markups being held when a key independent variable shifts from its minimum value to its maximum value. Similarly, for the negative  60  Table 3.1. Logit Results for the Effects of Bill Type on the Holding of Hearings Independent Variables  Estimated Coefficient  Change in Probability  Tax Bills  0.011 (0.637)  —  Social-Welfare Bills  -1.078* (0.644)  -21.64%  Moral-Values Bills  -1.270** (0.522)  -25.67%  Majority-Conflict Bills  -0.325 (0.430)  —  Discretionary Bills  -0.752 (0.491)  —  Presidential Proposals  0.563 (0.484)  —  Constant  Log-likelihood Pseudo R2 N (no. of bills)  2.200*** (0.515) -72.629 0.0825 156  Note: Column 1 shows logit estimates with standard errors in parentheses. Column 2 shows the change in probability that hearings are held, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their modes. Estimates generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  61  binomial models, I report the expected number of committee sessions held when a key explanatory variable shifts from its minimum value to its maximum value.22 I begin by conducting a logit regression on the Hearings Held dependent variable to determine if certain types of bills discourage committee sessions from taking place. Table 3.1 reports the results and provides strong support for the expectation that party-brand bills undergo high levels of one-party deliberations. The first column shows a negative correlation between social-welfare bills and the presence of hearings, as well as a negative correlation between moral-values bills and the presence of hearings. The second column reports the predicted probabilities of at least one hearing being held when these two types of bills are under consideration. Social-welfare bills are 22 percent less likely and moral-values bills are 26 percent less likely to undergo hearings than other legislation. While Table 3.1 reports no significant relationship between tax bills and omitted hearings, the findings here certainly indicate that bills designed to promote the party brand undergo truncated deliberations. Because the quantity of deliberations varies from measure to measure, I conduct a negative binomial regression on the Number of Hearings variable to see whether bill type affects panel frequency. Table 3.2 displays the results. As expected, I find that the majority party schedules fewer hearings for tax bills and moral-values bills than for other kinds of measures. Looking at the expected values reported in column 2, tax measures undergo about two fewer hearings and moral-values bills undergo nearly four fewer hearings than other kinds of legislation. Between the findings in Tables 3.1 and 3.2, there is strong support for Hypothesis 1.  22  Because the correlation between Hearings Held and Markups Held is rather strong (r = 0.45), I conducted a bivariate probit regression to see whether it produces significantly different estimates than the separate regressions for each committee stage. The results are similar. As in Chapter 2, I have opted here to report the findings from the separate regressions in order to discuss all the findings on hearings before I discuss the findings on markups.  62  Table 3.2. Negative Binomial Results for the Effects of Bill Type on the Number of Hearings Independent Variables  Estimated Coefficient  Expected Value  Tax Bills  -0.449* (0.255)  -1.803  Social-Welfare Bills  -0.378 (0.333)  −  Moral-Values Bills  -1.294*** (0.335)  -3.722  Majority-Conflict Bills  1.072*** (0.222)  +10.197  Discretionary Bills  -0.249 (0.233)  −  Presidential Proposals  0.365* (0.218)  +2.498  Constant  1.895*** (0.245)  Log-pseudolikelihood  -469.444  Alpha  N (no. of bills)  1.542 (0.188) 156  Note: Column 1 shows negative binomial estimates with standard errors in parentheses. Column 2 shows the expected value in the number of hearings, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their modes. Expected values generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  63  That is, on bills that cover issues central to the majority's reputation, the majority limits formal information gathering and proceeds quickly to bill drafting. Not all the findings in Table 3.2 support my predictions, however. Contrary to the hypothesis on intraparty conflict, I find that bills dividing the majority party prior to deliberations undergo more, not less, hearings than other types of measures. As column 2 shows, majority-conflict bills receive ten more legislative hearings than other types of measures. The bills in my sample receive on average about ten panel sessions, which means that majorityconflict bills undergo twice as many hearings as compared to other measures. It appears that when the majority caucus is initially divided over a legislative proposal, leaders of the majority avoid bill control and allow the committee of jurisdiction to engage in lengthy policy discussion. Next, I conduct a logit regression on the Markups Held dependent variable to see whether different types of bills discourage the scheduling of bill-drafting sessions. Table 3.3 reports the findings. Although there is no support for the hypothesis on party-brand bills in this model, there is support for the hypothesis on intraparty conflict. Indeed, column 1 shows that majorityconflict bills have a significant and negative correlation with committee markups. Column 2 displays the predicted probability of markups being held, and shows that majority-conflict bills are about 16 percent less likely to undergo markups than other types of bills. While these findings confirm expectations, they are nonetheless perplexing because they contradict the results on majority-conflict bills in the previous model. On the one hand, bills that divide the majority tend to undergo extensive legislative hearings but, on the other hand, they tend to skip the subsequent bill-drafting sessions. These apparently contradictory results merit further examination. I explore why majority-conflict bills are prone to both one- and two-party deliberations in the discussion section of this chapter.  64  Table 3.3. Logit Results for the Effects of Bill Type on the Holding of Markups Independent Variables  Estimated Coefficient  Change in Probability  Tax Bills  0.719 (0.830)  −  Social-Welfare Bills  -0.707 (0.747)  −  Moral-Values Bills  -0.454 (0.655)  −  Majority-Conflict Bills  -0.980** (0.497)  -15.73%  Discretionary Bills  -1.851** (0.781)  -10.72%  0.231 (0.547)  −  Presidential Proposals  Constant  Log-pseudolikelihood Pseudo R2 N (no. of bills)  3.737*** (0.830) -55.078 0.1063 156  Note: Column 1 shows logit estimates with standard errors in parentheses. Column 2 shows the change in probability that hearings are held, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their modes. Estimates generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  65  Table 3.4. Negative Binomial Results for the Effects of Bill Type on the Number of Markups Independent Variables  Estimated Coefficient  Expected Value  -0.703** (0.300)  -1.227  Social-Welfare Bills  -0.252 (0.306)  −  Moral-Values Bills  -0.572*** (0.211)  -1.093  Majority-Conflict Bills  0.310 (0.216)  −  Discretionary Bills  0.058 (0.226)  −  Presidential Proposals  -0.221 (0.212)  −  Tax Bills  Constant  0.831*** (0.228)  Log-pseudolikelihood  -299.927  Alpha  N (no. of bills)  0.718 (0.178) 156  Note: Column 1 shows negative binomial estimates with standard errors in parentheses. Column 2 shows the expected value in the number of hearings, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their modes. Expected values generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  66  To see whether particular types of measures affect the quantity of bill-drafting sessions, I conduct a negative binomial regression on the Number of Markups dependent variable. The findings reported in Table 3.4 provide confirming evidence for the hypothesis on party-brand legislation. Tax bills and moral-values bills are both negatively correlated with the presence of markups, supporting the expectation that legislation designed to promote the majority's brand undergoes limited deliberations. Column 2 shows that both types of bills receive approximately one less markup than other types of bills that go to committee. To put these numbers in context, the bills in my sample undergo on average two scheduled markups. Thus, majority-party leaders permit tax bills and moral-values bill to undergo about half the number of committee-writing sessions as other types of legislation. Because these two types of bills also undergo truncated deliberations at hearings, the results here strongly suggest that bills on issues that promote the majority's brand do not receive significant consideration at the prefloor stage. I turn now to the question of how committee factors−in particular, the ideological positions of committee leaders−affect prefloor deliberations. I conduct logit regressions on the Hearings Held and Markups Held dependent variables, this time including committee variables in the models instead of bill-type variables. In both cases, I find no statistically significant relationship between the ideological extremity of committee leaders and the presence or absence of panel sessions (see Tables A1 and A2 in Appendix A for full results). Neither extreme chairs nor extreme ranking members are linked to omitted prefloor deliberations. When I examine the frequency of committee sessions, however, I do find that leadership extremity is strongly correlated with declines in committee sessions. I conduct negative binomial regressions on the Number of Hearings and the Number of Markups dependent variables. Both models provide confirming evidence for my hypotheses on committee leadership. Table 3.5  67  Table 3.5. Negative Binomial Results for the Effects of Committee Factors on the Number of Hearings Independent Variables  Estimated Coefficient  Expected Value  Chair Extremity  -1.690** (0.860)  -13.142  Ranking Member Extremity  -1.823** (0.790)  -12.321  Majority Committee Margin  -0.008 (0.050)  −  Number of Referrals  0.138* (0.079)  +44.005  Constant  3.313*** (0.706)  Log-pseudolikelihood  -482.463  Alpha  N (no. of bills)  1.861 (0.267) 156  Note: Column 1 shows negative binomial estimates with standard errors in parentheses. Columns 2 shows the expected value in the number of hearings, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their means. Expected values generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  shows that as committee chairs and ranking members each become increasingly extreme, the number of legislative hearings on a bill decreases in turn. As the second column in Table 3.5 shows, a bill sent to a committee headed by the most extreme chair undergoes 13 fewer hearings  68  Table 3.6. Negative Binomial Results for the Effects of Committee Factors on the Number of Markups Independent Variables  Estimated Coefficient  Expected Value  Chair Extremity  -1.454** (0.732)  -2.756  Ranking Member Extremity  -1.120* (0.573)  -1.871  Majority Committee Margin  0.042 (0.032)  −  0.141*** (0.047)  + 9.531  Number of Referrals  Constant  Log-pseudolikelihood Alpha  N (no. of bills)  1.184* (0.646) -301.241 0.540 (0.142) 156  Note: Column 1 show negative binomial estimates with standard errors in parentheses. Columns 2 shows the expected value in the number of markups, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their means. Expected values generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  than a bill referred to a panel led by the least extreme chair. The shift from least to most extreme ranking member has nearly the same impact on the number of hearings. Similarly, Table 3.6  69  shows that as committee leaders from both parties become increasingly extreme, the number of bill-writing sessions declines. For example, a measure that goes to a committee run by the most extreme chair receives nearly three fewer markups than a committee directed by the least extreme chair. Clearly, then, when committee leaders have extreme preferences, formal prefloor deliberations are significantly reduced. 3.6 Discussion Many of the findings in this chapter confirm expectations. As predicted, bills that affect taxes, social benefits, and moral values all see reduced levels of prefloor deliberations. Indeed, tax measures undergo truncated hearings and markups; social-welfare bills tend to skip hearings altogether; and moral-values legislation circumvents hearings and receives limited consideration at the bill-writing stage. Taken together, these findings provide strong evidence in support of Hypothesis 1 and show that the majority controls deliberations on bills designed to promote the party's national brand. To include the minority in deliberations on these bills would be counterproductive from the majority's standpoint because bipartisan compromise only dilutes these bills and ultimately displeases its base. The implication is that parties are least likely to conduct thorough legislative deliberations on the very policies that they consider most important. Paradoxically, parties may even run the risk of overlooking important substantive concerns and producing defective bills on the measures that they expect will boost their appeal to voters. I focus on this very issue in the next chapter. One concern with the findings on partisan bills is that, by using three different types of legislation to test one prediction, I have given myself several chances to confirm the party-brand hypothesis. For example, I find that the majority party tends to omit hearings for social-welfare measures and moral-values bills, but not for tax bills. Despite the mixed findings, I interpret the 70  results overall as substantiating the hypothesis on party-brand legislation. Have I designed the study in such a way that it provides too many possible sources for confirming evidence? Perhaps; but I have examined each partisan bill separately primarily to identify the policy issues central to the theory of conditional party government. I am interested in telling a policy story, and an examination of several different types of bills is necessary in telling a full story. This approach, moreover, enables me to determine not only whether these bills undergo truncated procedures, but also which of these bills undergoes the highest or most consistent levels of one-party lawmaking. The analysis here shows that moral-values bills undergo the most consistently partisan procedures in prefloor development, followed by tax measures and then social-welfare bills.23 The findings on committee leadership confirm expectations to some extent. Although leadership extremity does not lead to the frequent exclusion of committees, increasing extremity of both the committee chair and the ranking member reduces the number of deliberative sessions a panel holds for a bill. In short, Hypotheses 3 and 4 bear out. Keeping in mind that previous research has found no evidence of committees being stacked with ideologues, the results here suggest that the preferences of committee leaders, rather than the ideological composition of the committees as a whole, may be the crucial factor in explaining panel activity. These findings also suggest that when majority-party leaders rush bills through the House, extreme chairs are likely acting as willing partners or, more accurately, brand managers.  23  Another research strategy is to create a party-brand dummy variable, in which all tax measures, social-welfare bills, and moral-values legislation comprise one category and all other bills comprise the other. This approach may provide a better test for the party-brand hypothesis than separate examinations of tax, welfare, and moral-values bills. I reran the four models on bill types using a party-brand dummy variable, instead of three separate bill variables, to see if the results differ. They do not. The majority omits hearings, reduces hearings, and limits markups on party-brand bills. See Tables A3 to A6 in Appendix A for full results.  71  The most puzzling findings in this chapter center on bills that divide the majority caucus prior to formal deliberations. On the one hand, majority-conflict bills tend to undergo a higher number of committee hearings than other bills, suggesting that the majority allows significant two-party discussion on policy matters that spit its caucus. This finding clearly challenges Hypothesis 2. On the other hand, majority-conflict bills are prone to skipping markup sessions, suggesting that majority-party leaders control bill writing in instances where their caucus is divided. This finding clearly supports Hypothesis 2 How can the results on majority-conflict bills both challenge and confirm predictions? What explains these contradictory findings? The majority likely has two incentives to engage in lengthy deliberations on policy matters that divide the party. First, the majority party may be interested in identifying policy solutions that will satisfy the greatest number of members in its caucus. Lengthy hearings allow the majority to hear from a large number of experts and to assess a broad array of proposals. The majority thus increases its ability to design a policy response that avoids, or at least minimizes, intraparty rifts. Second, the majority party may also be interested in allowing many of its members to raise concerns, offer solutions, and possibly testify formally before the committee, thereby ensuring that no grievances within the majority caucus are ignored. Obviously, lengthy discussions do not guarantee that everyone's suggestions will be adopted. But inclusive deliberations held in the early stages of bill development may help to reduce majority-party dissension in the later stages of the legislative process. However, once hearings end, majority-party leaders have incentives to draft the final versions of the majority-conflict bills behind closed doors. First, it gives them the opportunity to draft legislation that will have the broadest possible appeal to their caucus. Indeed, private billdrafting sessions give leaders the flexibility to negotiate delicate compromises with hesitant  72  caucus members who may be pivotal to the bill's passage. Second, bill drafting conducted in private obviously protects majority-conflict bills from destructive markup amendments. In sum, this chapter helps fill in the policy story that the theory of conditional party government has suggested but has never detailed. To maintain its brand, the majority party uses partisan strategies to push its core bills through the House, beyond minority-party members. Tax measures, welfare legislation, and moral-values bill undergo consistently limited deliberations at the prefloor stage. Extreme chairs appear to act as close partners with majority-party leaders in maintaining the electoral brand by reducing their committees' deliberative roles and limiting the minority's influence on bill content. Finally, in cases where the majority takes up legislation that divides its caucus, majority-party leaders protect the bill from potentially disruptive committee sessions in an attempt, presumably, to produce policy outcomes that they consider to be in the party's best interests. 3.7 Are Markups a "Formality"? A final point to consider: while there is a strong relationship between majority-conflict bills and the absence of markups, there is no such relationship between party-brand legislation and the absence of bill-writing sessions. Tax measures and moral-values bills tend to undergo significantly fewer markup sessions than other bills, but they do not appear to skip them frequently. The findings in this chapter, then, call into question the role that bill-drafting sessions play in the legislative process. If highly partisan bills often go to markup, even if it is only one markup, the presence of bill-writing sessions may not indicate the performance of genuine, twoparty deliberations. Sinclair makes this very observation, noting, "The public markup with minority party members present is simply a formality [today]" (2006, 148). This point raises  73  some obvious questions. Are all markups formalities? If a bill undergoes two, three, or even ten markup sessions, are they all formalities, too? Sinclair does not say. Possibly, the quantity of markup meetings provides a good indication of the degree to which a bill has undergone a two-party process of development. There are a number of notorious cases in which committee chairs rushed a bill through a single markup, giving the minority little or no time to draft germane amendments. Rep. Bill Thomas (R-CA), the Chairman of the House Ways and Means Committee, once prevented Democrats from even reading a bill in markup and called the Capitol Police to oust Democrats when they attempted to locate a copy of the bill in a library (Hetherington 2009, 413). Clearly, a single markup session can be a mere formality, if not an opportunity for partisan bullying. Multiple sub- and full-committee sessions, by contrast, may signal the majority's willingness to consider minority-party proposals. Unfortunately, the most thorough research on markups is more than two decades old and tells us little about how markups operate in the partisan Congress (see Hall 1987; Hall and Evans 1990). More research on markups is therefore required. Although I focus primarily on the consequences of one-party lawmaking in the next chapter, I also reexamine the role of bill-writing sessions and consider in greater depth what their absence indicates.  74  4. Deliberative Deficit: The Consequences of One-Party Lawmaking  4.1 Introduction Cases of congressional majority parties rushing to pass problematic legislation abound. Exhibit A: In 2003, House Republicans attached a $20 billion measure for reconstruction projects in Iraq to a large troop-funding bill. Republicans inserted the provision into the war supplemental fund to curtail deliberations and to make it politically difficult for Democrats to raise objections. Many observers predicted that the reconstruction money would be wasted, because the bill included no guidelines for awarding contracts or monitoring projects ("Iraq Reconstruction Bonanza" 2003; Krugman 2003). Two years later, an audit found that nearly half the funds had disappeared because of corruption ("Audit" 2005). Exhibit B: In 2007, after President George W. Bush had ordered 30,000 additional soldiers to Baghdad, House Democrats passed a measure to initiate a military exit from Iraq. The press noted that Democrats were ignoring reports of declining violence due to the presence of extra U.S. forces, and editorials argued that a pullout would jeopardize the recent security gains ("House's Rush to Judgment" 2007; "Rush to Judgment" 2007). The House bill authorizing troop withdrawal failed to pass the Senate; but instead of contemplating ways to bolster Iraq's growing stability, antiwar Democrats omitted committee hearings and rushed another troop-withdrawal bill to the floor (Rogin 2008). Does total control of bill deliberations by the majority party produce "bad" legislation? Conversely, does the inclusion of the minority party in the decision-making process reduce policy mistakes and produce "good" legislation? In a recent edited volume on partisan polarization and its effects, leading congressional scholars, including Krehbiel (2008) and Sinclair (2008b), raise these very questions. However, they do not provide definitive answers.  75  Beyond a few scholars—notably Hacker and Pierson (2005, 2007, 2010) and Mann and Ornstein (2006)—most political scientists hesitate to evaluate bills, pointing out that there are no generally accepted criteria to make judgments about legislative quality. Furthermore, scholars recognize that they risk injecting their own values into the analysis and producing highly biased research that other scholars, with different ideological views, will likely reject. Sinclair goes so far as to conclude that assessing legislative quality is "a fool's errand" (2008b, 82). The decision to avoid judging bills, while understandable, has encouraged scholars to avoid a pressing question in American politics today: Does increasing partisanship lead to the development of increasingly problematic legislation? In this chapter, I take on the fool's errand and try to answer this question. Specifically, I attempt to identify the policy consequences of the House majority party dominating bill development and excluding the minority party from formal legislative deliberations. To conduct this study, I develop two new methods for assessing bill quality in a large sample of bills. First, I measure the level of negative criticism a bill receives in unsigned editorials from a large selection of partisan and nonpartisan daily newspapers. Legislation that receives negative evaluations in newspapers representing a variety of regions and viewpoints likely contains severe policy defects. Second, I develop a new measure to estimate the level of ideological extremity in legislation based on DW-NOMINATE and roll-call data. Extreme bills likely respond to the demands of a narrow constituency at the expense of the broad public and arguably foster a deterioration of democratic responsiveness. After developing these measures, I collect data on hearings, markups, and floor debates for major bills drafted in the last three decades, and conduct regression analyses to determine whether the exclusion of the minority party from bill deliberations affects the quality of legislation.  76  For the most part, the results of this chapter reinforce concerns that one-party lawmaking produces defective and extreme bills. I find that the level of editorial criticism is significantly greater for legislation that has skipped hearings and has undergone restricted floor debate than it is for legislation that has been considered under the regular order. These results suggest that when the minority party is barred from the legislative process, the final legislative product contains obvious shortcomings. Likewise, I find that bills tend to be more ideologically extreme if hearings have been omitted instead of held. However, contrary to expectations, I also find that bills tend to be more ideologically moderate if markups have been circumvented instead of maintained. Truncated deliberation, then, appears to produce both one-sided and balanced legislation. To explain this apparent contradiction, I conduct a series of case studies that examine how hearings and markups separately affect bill content. My findings indicate that hearings can lead to significant legislative changes that reflect minority-party preferences, but that markups often function as symbolic exercises, not as genuine bill-drafting sessions. Thus, not only does this chapter demonstrate that one-party lawmaking tends to produce problematic legislation, but it also shows that hearings often play a much more important deliberative role than markups do. I proceed as follows in this chapter. First, I discuss previous studies on the policy consequences of limited deliberations, and I explain why we should expect one-party decisionmaking to produce worse legislative outcomes than two-party decision-making. Next, I describe my measures of bill quality, and I discuss the major results of the data analysis. Because findings on hearings and markups run in different directions, I then conduct case studies to explore how committee operations affect bill content. Finally, I conclude by suggesting several new research questions.  77  4.2 Concerns over One-Party Deliberations Few studies examine the consequences of truncated deliberations in Congress (for a review, see Quirk and Bendix 2011). Although limited, existing research indicates that one-party lawmaking produces three negative, or arguably negative, outcomes: legislative gridlock, policy mistakes, and ideologically extreme bills. A number of scholars argue that partisan polarization has produced unprecedented levels of legislative stalemate, preventing Congress from implementing solutions to rising entitlement costs, increasing poverty rates, and continuing flows of illegal immigrants, among other issues (Brady, Ferejohn, and Harbridge 2008; Galston and Nivola 2006; Jones 2001; McCarty 2007). Sinclair (2006), sharing these concerns over polarization and gridlock, considers whether bills that have undergone one-party processes face greater legislative obstacles than bills that have undergone two-party processes. Examining major bills from the last forty years, she finds that measures developed by the majority party have a lower rate of enactment than bills drafted by both parties (2006, 358-359). Deliberative partisanship, in other words, contributes to stalemate. While deadlock over one-party bills allows serious problems to persist, passage of oneparty bills can create new problems. Mann and Ornstein (2006), in a series of case studies, show that the majority party inserts poorly designed provisions into bills drafted in closed-door sessions. They note that majority-party leaders often fail to conduct a "reasonable assessment of costs and benefits" (128) and sometimes "slap together" legislation that contains measures even the leaders themselves come to regret (173). To illustrate, Mann and Ornstein discuss a 3,000page appropriations bill that Republican House leaders assembled, in haste, in 2004. Working through the night, lawmakers unintentionally added a provision that empowered congressional staff to search and circulate private tax records without criminal punishment.  78  According to Hacker and Pierson (2005, 2007, 2010), one-party deliberation leads to the development of extreme legislation. They argue that extreme bills are troubling because they provide disproportionate benefits to the majority's base at the expense of most other voters. They point to the Bush tax cuts as evidence, noting that these bills allocated generous cuts to affluent voters and provided only modest cuts to middle-class voters. These bills satisfied conservative interest groups and wealthy Republican donors but also helped produce long-term structural deficits and greater income disparities (see also Bartels 2008; Jones and Williams 2008; Shaviro 2007). Hacker and Pierson conclude that party-dominated lawmaking encourages Congress members to make crass deals that benefit special interests at a cost to most citizens. Democratic values, such as representativeness and accountability, are undermined. 4.2.1 Limitations of the Literature Overall, the literature provides compelling examples of one-party deliberation producing policy mistakes and ideologically extreme bills. But because this literature is based on only a few case studies, all of which focus on Republican-drafted legislation, some scholars question the generalizability of the findings. Sinclair (2008b, 79) wonders whether examples of problematic one-party bills are truly more prevalent than examples of problematic two-party bills. Selection bias may account for the dearth of cases that illustrate shortcomings in bipartisan measures. Likewise, Brady, Ferejohn, and Harbridge concede that extreme bills can produce problematic results, but they argue that extreme bills are not prevalent even in today's partisan Congress. For example, they (2008, 194-195) point out that Hacker and Pierson's singular focus on the Bush tax cuts ignores the many bipartisan measures enacted during the Bush presidency. The absence of large-N studies on bill quality has ultimately encouraged an inconclusive scholarly debate. Yet, the difficulty in developing widely accepted measures of problematic 79  legislation has discouraged scholars from examining large samples of bills to determine the policy consequences of one-party deliberation. As Sinclair (2008b, 79-82) explains, political scientists have neither identified objective standards by which to judge bills nor have they found systematic data necessary to conduct evaluations of legislative quality. Thus, the two greatest concerns over one-party lawmaking—that it produces policy defects and ideologically extreme legislation—remain largely untested. Furthermore, the congressional literature has not developed clear explanations of why we should expect one-party deliberation to produce significantly worse policy outcomes than two-party deliberation.24 Indeed, bipartisan collaboration may often involve little more than logrolling, where both parties develop and support "win-win" agreements that promote excessive spending (Guttmann and Thompson 2012, 14). For this reason, the next section of this chapter considers why bipartisan committees are generally able to develop better legislative solutions than party-run policy groups. 4.3 The Effectiveness of Group Deliberations The case studies on partisan lawmaking suggest that we can conceptualize problematic bills in two ways. First, we can think of them as being defective bills that produce unnecessary costs or foreseeable inefficiencies. Such bills contain mistakes that lawmakers would avoid if they recognized the defective features in the bills before supporting them. Second, we can think of problematic bills as ideologically extreme measures that satisfy special interests and overlook important concerns. American constitutional traditions favor policy deliberations and policy  24  Case studies on policy development in the British parliament suggest that domination of deliberations by government ministers, i.e., majority-party leaders, results in the passage of widely recognized legislative "blunders" (Crewe and King 2010). But the literature on parliamentary deliberations includes the same gap as the congressional literature, namely, the absence of large-N analyses on the consequences of one-party lawmaking.  80  outcomes that serve broad interests (Bessette 1994). Extreme bills undermine these traditions and neglect the values of an electorate that tends to hold centrist preferences. Why should we expect these types of problematic bills to be produced in partisan, rather than bipartisan, contexts? When the parties work together, they study and discuss policies in committees to identify options that will help them achieve favored goals and avoid unsatisfactory outcomes (Krehbiel 1991; Quirk 2005). By including the minority in discussions, the majority exposes itself to diverse policy suggestions and often cedes some control over bill content to its opponents. Hearings allow the minority to call witnesses and to highlight a bill's shortcomings; markups and open floor debate give the minority opportunity to alter a bill by offering amendments. The comprehensive information gathering and thorough policy debate that comprises two-party deliberation likely helps lawmakers reduce, or avoid, policy mistakes. Conversely, the majority gains total control over legislative content when it sets aside regular order and excludes the minority party from deliberations, but it risks overlooking ideas and concerns that minority-party legislators would raise in committee. Of course, the majority party can use ad hoc task forces to conduct comprehensive deliberations on its own, in an effort to fill the gap that it creates by skipping hearings and markups (Sinclair 1997, 2006). Yet, party deliberations are unlikely to be as effective or informative as committee deliberations. Research shows that heterogeneous groups tend to perform better as problem solvers than homogenous groups (Hong and Page 2001, 2004; Page 2007). As groups become more diverse, they become better at identifying and solving problems because they are able to analyze issues from multiple perspectives. That is, heterogeneous groups see concerns that homogenous groups fail to see, and  81  they develop superior remedies as a result.25 Because committees have members from both parties, they are likely better at detecting and correcting flaws in bills than party policy groups. Moreover, because committees include both Republicans and Democrats, they are less vulnerable to bias and policy extremism than party task forces. Research on group decisionmaking shows that a diversity of opinions encourages open-mindedness in groups, because it makes members realize that their views are not unanimously shared (Sunstein 2005, 993). By contrast, when likeminded people work together, they tend to adopt positions more extreme than they held before deliberations began (Sunstein 2000, 2005, 2006, 2007; see also Janis 1982). This "group polarization," as it is called, involves three steps. First, similar points are heard more frequently than contrary arguments among the likeminded members. This high frequency of similar arguments encourages the group to consider primarily one type of solution. Second, social pressure within the group encourages all members to adopt the same position; no one wants to be the odd man out in a roomful of likeminded peers. Third, the most extreme group members have the highest confidence in their positions and tend to adopt even more extreme positions if they see their views even partly corroborated. Their increasing extremism pushes the rest of the group to follow suit (Sunstein 2006, 204). So, when majority-party lawmakers draft legislation in private, without minority-party input, they encounter "limited argument pools" and become susceptible to group polarization (Sunstein 2000, 104). Given the blind spots and biases of homogenous groups, we should expect that one-party deliberations are more likely to produce defective bills and extreme bills than two-party deliberations. I determine whether these two expectations are correct in the remaining sections.  25  In fact, a heterogeneous group can outperform a homogenous group in problem solving, even when the heterogeneous group lacks the qualifications of the homogenous group. A diversity of views and perspectives can be a greater asset than expertise (Hong and Page 2004; Page 2007).  82  4.4 Data and Methods To investigate the consequences of party-directed lawmaking, I use the sample of 156 major bills that I compiled in the second chapter. Developed between 1983 and 2008, these bills are drawn from the key votes identified by Congressional Quarterly. Each year CQ identifies about ten key votes on important bills debated in the House. To gather my sample, I stratify by year and randomly draw six bills from a pool of about ten CQ key votes from every session of Congress. The sample covers diverse policy issues—such as national security, gay rights, and energy regulations, among others—and it includes bills developed under both Republican and Democratic majorities. It is especially necessary to focus on important bills for this study, because editorials (which I use to develop several dependent variables) are unlikely to be written on secondary bills or on bills that fail to reach the House floor. 4.4.1 Measuring Legislative Defects I use unsigned editorials from forty-five daily newspapers, representing both large and small cities from every major region in the United States, to measure how problematic each bill is in the sample (see Table B1 in Appendix B for a list of newspapers). Presumably, the more criticism a bill receives in the press, the greater the likelihood that it contains serious mistakes.26 Unsigned editorials in newspapers are written by the editorial-board writers and represent the official opinion of the publication. Major newspapers have over a dozen editorial writers who convene daily to produce "collective judgments" on major national issues (Gottlieb 1999). Newspapers frequently publish unsigned editorials that provide relatively detailed evaluations of important bills that are under debate or have been recently passed. 26  A possible counterargument to this assumption is that journalists often frame news stories around, or "index" them to, the conflicts played out between politicians; that is, the arguments they present reflect those that are made by elites (Bennett 1991). If this indexing effect is at work, then increased bill criticism in the press could simply reflect increased criticism raised by opponents of the majority. Indexing is unlikely to account for the judgments of editorial boards, however. Unlike news article, editorials are intended to reflect the board's judgment.  83  To develop measures of problematic bills based on editorial criticism requires multiple steps. First, using LexisNexis, I collect all relevant unsigned editorials that appear from one week before to three years after the first House vote on each bill. The collected editorials focus on only House bills, and do not include Senate or conference committee versions.27 I look for editorials over a long time span because Congress often debates major policy issues for years, inviting editorial boards to offer multiple appraisals.28 To identify relevant editorials, I limit LexisNexis searches of bill descriptors to headlines and leads. Descriptors include the name of the bill used in the press (e.g., the Brady Bill, "Don't Ask, Don't Tell," and the Patriot Act), as well as the main policy issues that the bill covers. Press names of bills are found in descriptions of major legislation in CQ Almanac. I do not collect editorials on bills developed before 1987, even though my sample includes bills developed as early as 1983, because LexisNexis searches do not yield results from a sufficient number of newspapers before 1987. All models that include measures based on editorial criticism thus have only 132 cases, not the full 156 cases. Next, I code editorials as "positive" if they clearly endorse the bill or a major provision in the bill, or if they criticize counter proposals. I code editorials as "negative" if they clearly oppose the bill or a major provision in the bill, or if they support a counter proposal. I code an editorial as "neutral" if it provides a description of the bill without offering a clear evaluation, or if the evaluation is both positive and negative. About 95 percent of the collected editorials 27  However, I do include editorials that compare and evaluate various versions of the same bill. Early in the 109th Congress, for example, House Republicans introduced the Energy Policy Act of 2005, which authorized drilling in the Arctic National Wildlife Refuge (ANWR). Editorial boards made appraisals of this bill, focusing in particular on the two main policy arguments that Republicans made in support of the bill: first, that oil operations in the Arctic would reduce fuel prices by encouraging increased crude supplies and, second, that these new drilling operations would follow safe environmental procedures. The House passed the bill in April 2005 and the Senate passed an alternative energy bill, one that did not authorize Arctic drilling, two months later. Both House and Senate votes prompted editorial boards to evaluate the issue of drilling in ANWR. Two years later, Democrats took control of the House and promptly offered their own energy bill. This bill, like the 2005 Senate bill, did not include a drilling provision. Again, editorial boards considered not only the merits of this latest measure but also the merits of drilling in ANWR. Policy reexaminations are not unique to the issue of arctic drilling. In the years of this study, Congress and the press frequently revisited campaign-finance reform on soft-money contributions, the banning of assault weapons, and caps on entitlement spending, among other policy issues. 28  84  (almost 2,500 in all) provide either clear-cut endorsements of or strong objections against bills. They often call legislation "good" or "bad," judge a legislative vote as either "right" or "wrong," or use other directly evaluative language to describe a measure. Many editorial boards have stable partisan biases. For this reason, I code each newspaper as an advocate, opponent, or neutral observer of the majority party depending on which presidential candidates the newspaper supports and which party controls the House at the time of each bill's development.29 For example, if a newspaper consistently supports Republican candidates for president, I code the newspaper as an advocate during years of Republican House control and as an opponent during years of Democratic House control. Newspapers that have a record of either endorsing both Republican and Democratic presidential candidates or of avoiding endorsements altogether are coded as neutral. In total, I collect unsigned editorials from nineteen Democratic-leaning newspapers, fifteen Republican-leaning newspapers, and eleven nonpartisan newspapers. I use Editor & Publisher's Quadrennial Presidential Poll to collect endorsements by editorial boards from 1992 to 2008. I also use data compiled by Hayes (2008) to help complete endorsement lists. After collecting and coding data, I measure the level of negative criticism against bills in three ways. First, I create the dependent variable All-Sources Editorial Criticism by compiling the total number of negative editorials published in both partisan and nonpartisan newspapers. This variable provides a measure of the press's overall disapproval of a bill. Second, I compile the number of negative editorials published in only neutral newspapers for each bill to create the dependent variable Nonpartisan Editorial Criticism. This measure enables me to determine whether one- and two-party bills receive similar levels of criticism from impartial sources.  29  I coded the newspapers after I coded the editorials, so that I was unaware of the partisan biases of the newspapers as I examined their assessments of bills.  85  Third, I create the dependent variable Partisan Editorial Divergence by subtracting the percentage of editorials that are negative in opponent newspapers from the percentage of editorials that are negative in advocate newspapers for each bill. By doing so, I measure the difference in the level of negative editorial criticism a bill receives from papers with opposing partisan loyalties. For example, if during a period of Democratic majority control the level of criticism for a bill is 75 percent in Republican-leaning papers and 25 percent in Democraticleaning papers, then the level of editorial disagreement for the legislation is 50. This measure allows me to determine the extent to which one-party bills have narrow partisan appeal. Presumably, the narrower a bill's appeal, the more likely it contains serious defects. Although some partisan-leaning papers offer predictable criticism (the New York Times rarely supports Republican legislation and the Washington Times rarely supports Democratic bills), most partisan-leaning papers support measures from both parties. When these newspapers criticize legislation, they tend to offer detailed analysis based on substantive policy concerns, not ideological objections. For example, in 2007, the Republican-leaning Omaha World-Herald opposed the Democrats' plan to expand government-funded health insurance to all children in families of four earning less than $62,000 a year. The newspaper noted that, while a family of four would likely find this income inadequate in a major urban center, the same family would not struggle financially in a rural region. The World-Herald thus argued that an across-the-board increase in coverage would allow comfortable middle-class families to qualify for benefits that they did not need ("Budgetary Failure" 2007). Democratic-leaning newspapers, meanwhile, overlooked these points. As this example illustrates, a large divergence of opinion between partisan newspapers is not simply an indicator of competing party loyalties over goals. Rather, it is indicator of significant bill defects.  86  4.4.2 Measuring Legislative Extremity Concerns over legislative content have encouraged scholars to develop multiple methods of estimating the spatial location of bills (Clinton and Meirowitz 2001; Jeong 2008; Woon 2008). However, because I am interested in estimating a bill's level of ideological extremity, not its precise ideological location, these estimation methods are not appropriate for my purposes. I do not need to determine whether a bill is extremely conservative or extremely liberal, only the degree to which it is extreme. To create a measure of bill extremity, I multiply the percentage of lawmakers opposed to the bill by the tendency for the vote to be ideologically divided. That is:  [Nay Votes / (Yea Votes + Nay Votes)] x PRE (Proportional Reduction in Error)  I use the PRE from DW-NOMINATE data as a measure of ideological division for each bill (Poole and Rosenthal 2007).30 The PRE is 1 if all lawmakers vote as expected in ideological order on a proposal and, in turn, the NOMINATE model makes no classification errors. The PRE is 0 if many lawmakers vote out of ideological order on a proposal and, in turn, the NOMINATE model makes as many classification errors as the benchmark model that predicts 100 percent support for the winning policy alternative. Lawmakers closest to the cutting line, and thus indifferent to the vote's outcome, account for most classification errors (Poole and Rosenthal 2007, 36-41). A high rate of error in a specific case indicates, then, that many lawmakers beyond the indifferent ones have voted out of order. By itself, the PRE does not specify whether a bill is extreme, only the extent to which lawmakers vote in or out of sequence. When the chamber splits into two perfectly homogenous ideological camps, the PRE does not distinguish whether the yea and nay votes are similar in number or whether the nay votes are relatively few. For this reason, I 30  PRE = (Majority Errors - Model Errors) / Model Errors.  87  estimate a bill's extremity by multiplying the percentage of nay votes by the PRE, so that I take into account both the size and the ideological consistency of the opposition. The most ideologically extreme bills have a score close to 50, where nearly half the chamber has voted against the measure and where the PRE is 1.31 The least ideologically extreme bills have a score of 0, in which case no lawmaker has voted against the measure. I then multiply all bill scores by 2, so that they run approximately from 0 to 100, to create the dependent variable Bill Extremity.32 4.4.3 Independent Variables I use two sets of variables to measure the degree to which the majority party dominates bill deliberations. The first set measures whether the majority has included or excluded the minority in legislative decision-making. An omitted legislative step represents the exclusion of the minority party and indicates a bill has undergone some level of one-party deliberation. Hearings Held and Markups Held are dummy variables that indicate whether a formal stage of committee deliberation has been held (1) or not (0). When hearings are skipped, the majority denies the minority the opportunity to invite and question relevant policy experts on a legislative proposal. Similarly, when markups are skipped, the majority denies the minority the opportunity to participate in bill drafting at the prefloor stage. Floor Rule is a dummy variable denoting whether a bill undergoes open (1) or restricted floor debate (0). When the majority implements a restricted rule, it reduces or eliminates the minority's ability to offer amendments during floor debate. Data on hearings are collected from LexisNexis Congressional; data on  31  It is possible for an extreme bill to receive very little support on the floor and to have the proportion of nay votes exceed 50 percent. In practice, however, House majority-party leaders rarely bring bills to the floor where an overwhelming number of lawmakers vote nay. 32 When the majority party proposes a small change on a purely ideological bill—an incremental tax hike, for example—the vote is likely to divide the chamber into two nearly equal factions. This tax hike will appear to be extreme even though it is relatively minor because the vote is divisive. However, because I use only important bills in my sample, these bills likely propose only large shifts in the status quo. Thus, highly divisive votes in this study likely indicate a bill's extremeness.  88  markups and floor debate are collected from CQ.com and the Thomas website. (See Chapter 2 for complete details on data collection for these variables.) The second set of independent variables measures the quantity and scope of legislative deliberations for each bill. Presumably, as legislative deliberations increase, especially in formal settings, the minority party gains greater opportunity to shape a bill's language. Number of Hearings is the number of scheduled hearings held for each bill. The more hearings that are scheduled, the more opportunity the minority party has to question witnesses, raise policy concerns, and call its own experts to testify. Number of Markups is the number of markups held for each bill. As more markups are held, the minority party has greater opportunity to offer and discuss germane amendments. Length of Floor Debate is the number of hours granted for general debate on each bill. Debate time is always split equally between the two parties (Davidson and Oleszek 1994, 338339). Thus, longer debate always gives the minority more opportunity to discuss measures. Duration of Deliberations is the number of days from a bill's introduction in the House to its vote on the floor. The opportunity for minority-party members to engage in both formal and informal deliberations decreases when a bill is rushed through the House. The Thomas website provides complete timelines for each bill from introduction to final action. Cosponsor Balance is the percentage of cosponsors from the minority party. This measure indicates the degree to which a bill has circulated throughout the House before it reaches the floor for a vote. Cosponsorship is often understood as a signal that a bill matches a lawmaker's preferences (e.g., Koger 2003; Kressler and Krehbiel 1996; Woon 2008). But it also denotes whether legislation has undergone a relatively open or closed deliberative process. Bill sponsors have an incentive to circulate their proposals outside committee to attract broad support  89  for their measures before they reach the floor (Davidson and Oleszek 1994, 322; Krutz 2005). Sponsors expect that as the number of cosponsors increases, their bill's chances of surviving the legislative process increase in turn. Meanwhile, minority-party lawmakers have an incentive to cross party lines and cosponsor majority-party bills to show constituents that they are effective legislators willing to work in a bipartisan manner (Campbell 1982; Harbridge 2011). In fact, minority-party members are more likely to cosponsor bills than majority-party members. The percentage of minority-party cosponsors thus indicates not only the level of minority-party support for a bill, but also the level of minority-party inclusion in prefloor deliberations.33 4.5 Results I use count data to construct the dependent variable All-Sources Editorial Criticism. These data contain a large number of zeroes, suggesting that overdispersion is an issue. As a result, I need to use a model that does not assume the conditional variance and the conditional mean of the dependent variable are equal. A count model is suitable in these cases. I conduct a Vuong test and determine that a zero-inflated negative binomial (ZINB) model is favored over its non-inflated counterpart. ZINB models are appropriate when the data contain two types of zeroes: "true zeroes" and "excess zeroes." The data for All-Sources Editorial Criticism contain zeroes when papers publish only positive editorials and when they publish no editorials on a bill. Table 4.1 reports the first-stage results from the ZINB regressions on All-Sources Editorial Criticism. Model 1 shows that the number of negative editorials from all newspapers declines when hearings are held. Model 2 shows that the number of negative editorials decreases as the number of markups and the percentage of minority-party cosponosors increases. In short, both models show that editorial boards are less critical of two-party bills than one-party bills.  33  Using the number of minority-party cosponsors in the models produces similar results.  90  Table 4.1. ZINB Results for the Effects of Legislative Deliberations on All-Sources Editorial Criticism Independent Variables  Model 1  Model 2  Hearings Held  -0.564* (0.333)  —  Markups Held  0.080 (0.357)  —  Floor Rule  0.045 (0.259)  —  Number of Hearings  —  -0.003 (0.013)  Number of Markups  —  -0.080* (0.048)  Length of Floor Debate  —  -0.142 (0.115)  Duration of Deliberations  —  0.0003 (0.001)  Cosponsor Balance  —  -0.009* (0.007)  Constant  2.672*** (0.213)  2.840*** (0.239)  Log pseudolikelihood Wald chi-square (3) Alpha  -374.605 4.27 1.185 (0.196) 132  -368.002 21.19 1.104 (0.163) 130  N (no. of bills)  Note: Robust standard errors in parentheses. *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests)  91  Because ZINB models generate coefficients that are difficult to interpret, I compute the marginal effects of all significant explanatory variables on the dependent outcome. The effect of hearings being held on the overall level of editorial criticism is -4.83. A bill that undergoes hearings, in other words, receives about five fewer negative editorials than a bill that skips hearings. To put this finding into context, the bills in my sample receive on average 8.5 negative opinion pieces (see Table B2 in Appendix B for the frequency distribution of All-Sources Editorial Criticism). A decrease of almost five editorials due to the presence of hearings represents a 56 percent decrease in the level of disapproval for a bill. The effect that the number of markups has on editorial criticism is -0.69, indicating that the increase of three markup sessions for a bill produces a decrease of two negative editorials. And the effect of Cosponsor Balance is -0.08, suggesting that a 12 percent increase in the proportion of minority-party cosponsors produces a decrease of one negative editorial. I also use count data for the dependent variable Nonpartisan Editorial Criticism, and find that a ZINB model is appropriate here as well. Table 4.2 reports the first-stage results from the ZINB regressions. Model 3 shows that the number of negative editorials from nonpartisan papers declines when hearings are held. Model 4 shows that the number of negative editorials decreases as the number of markups and the percentage of minority-party cosponsors increases. Notably, the results in Table 4.2 are almost identical to the results in Table 4.1. Again, I calculate the marginal effects to identify the impact that deliberation has on bill evaluations. The effect of hearings on nonpartisan criticism is -0.73. On average, bills receive one negative editorial from nonpartisan newspapers (see Table B3 in Appendix B for the frequency distribution of Nonpartisan Editorial Criticism). Thus, a 0.73 decline in negative editorials because of hearings represents a 73 percent decrease in nonpartisan disapproval. The  92  Table 4.2. ZINB Results for the Effects of Legislative Deliberations on Nonpartisan Editorial Criticism Independent Variables  Model 3  Model 4  Hearings Held  -0.727** (0.360)  —  Markups Held  -0.153 (0.421)  —  Floor Rule  -0.260 (0.377)  —  Number of Hearings  —  0.011 (0.020)  Number of Markups  —  -0.191* (0.114)  Length of Floor Debate  —  -0.241 (0.165)  Duration of Deliberations  —  0.0002 (0.002)  Cosponsor Balance  —  -0.016** (0.007)  Constant  0.991*** (0.279)  1.227*** (0.335)  Log pseudolikelihood Wald chi-square (3) Alpha  -152.386 5.85 1.308 (0.412) 132  -150.048 17.54 1.224 (0.405) 130  N (no. of bills)  Note: Robust standard errors in parentheses. *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests)  93  effect that the number of markups has on nonpartisan criticism is -0.20, suggesting that an increase of five bill-writing sessions produces a decrease of one negative editorial from neutral papers. And the effect of cosponsorship is -0.02, which indicates that an increase of 50 percent in the proportion of minority-party cosponsors produces a reduction of one negative editorial from nonpartisan newspapers. Next, I conduct OLS regressions on the Partisan Editorial Divergence variable. Table 4.3 reports the results. Model 5 shows that the gap in bill assessments by opponent and advocate papers decreases when formal legislative steps are maintained. Specifically, the difference in the level of criticism between the two types of papers declines by over 22 percent when hearings are held and by almost 19 percent when floor debate is open. The majority appears to produce less controversial legislation when it increases rather than limits the minority's opportunities to participate in bill development. Model 6 shows how the quantity of legislative deliberations affects bill evaluations by advocate and opponent papers. As the length of floor debate, the duration of House deliberations, and the percentage of minority-party cosponsors all increase, bill assessments by advocate and opponent papers increasingly converge. For example, as general floor debate on a bill increases by one hour, the difference in the level of criticism between opponent and advocate papers decreases by 14 percent. As in Model 5, the findings in Model 6 suggest that two-party deliberations tend to reduce editorial disagreement. To determine the effects of one-party deliberations on the ideological position of bills, I conduct OLS regressions on the Bill Extremity dependent variable. Results are displayed in Table 4.4. Because of possible endogeneity between cosponsorship and a bill's ideological character, I do not include Cosponsor Balance in these regression equations. Model 7 shows that bills for  94  Table 4.3. OLS Results for the Effects of Legislative Deliberations on Partisan Editorial Divergence Independent Variables  Model 5  Model 6  Hearings Held  -22.414* (13.021)  —  Markups Held  19.269 (15.465)  —  Floor Rule  -18.741* (10.257)  —  Number of Hearings  —  -0.078 (0.262)  Number of Markups  —  1.806 (2.077)  Length of Floor Debate  —  -14.350*** (4.554)  Duration of Deliberations  —  -0.143** (0.061)  Cosponsor Balance  —  -0.386** (0.162)  34.119*** (12.592)  63.917*** (8.858)  0.040 132  0.178 130  Constant  R2 N (no. of bills)  Note: Robust standard errors in parentheses. *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests)  95  Table 4.4. OLS Results for the Effects of Legislative Deliberations on Bill Extremity Independent Variables  Model 7  Model 8  Hearings Held  -23.753*** (7.064)  —  Markups Held  18.918** (8.974)  —  -4.095 (5.339)  —  Number of Hearings  —  -0.390** (0.153)  Number of Markups  —  0.005 (0.860)  Length of Floor Debate  —  3.033 2.463  Duration of Deliberations  —  -0.043 (0.031)  49.275*** (7.311)  46.994*** (4.949)  0.072 155  0.080 149  Floor Rule  Constant  R2 N (no. of bills)  Note: Robust standard errors in parentheses. *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests)  which the majority party holds hearings are almost 24 points lower on the 100-point extremity scale than legislation for which the majority party skips hearings. Model 8 shows that each  96  additional hearing lowers a bill's placement by 0.39 points on the extremity scale. Moderate legislation, therefore, has a strong correlation with the presence of and the increase in hearings. The findings on markups in Table 4.4 are much more surprising and challenge the expectation that truncated deliberation produces ideologically extreme legislation. As Model 7 reports, bills for which the majority party holds markups are nearly 19 points higher on the extremity scale than bills for which the majority party omits markups. These results suggest that when the majority allows the minority to participate in bill-drafting sessions, it produces bills that are more extreme than when it develops bills outside committee. Put another way, moderate legislation appears to have a strong association with the absence of markups. How can hearings and markups produce opposite findings? I raised a similar question in the previous chapter, when I found that bills that divide the majority caucus undergo a large number of hearings but, at the same time, skip markups. Before I pursue this question further, and attempt to link the results from Chapter 3 to the contradictory findings here, I need to determine first if the estimates in Table 4.4 are the result of multicollinearity. When independent variables are highly correlated, regressions can produce significant coefficients with the "wrong" sign (Kennedy 2006). A strong relationship between hearings and markups may explain why the results in Table 4 show two-party deliberation apparently fostering both ideological restraint and ideological extremity. To check for multicollinearity, I calculate the variance inflation factor for each independent variable and find that they are all below the standard 2.5 threshold (Allison 1999). I also conduct separate regressions in which I exclude Hearings Held in one model and Markups Held in another model to see if the findings otherwise remain consistent with the findings in Table 4.4. They do (see Table B4 in Appendix B), indicating again that there is no collinearity problem.  97  Because there are no obvious explanations for these contradictory findings, my next step is to examine hearings and markups separately to see how each committee stage affects bill formulation on its own. I conduct this examination in the proceeding sections. 4.6 Tracing the Effects of Hearings on Bills To determine what role hearings play in shaping legislative outcomes, I need to determine the direction of causality between hearings and bill extremity. Indeed, the results in Table 4.4 do not tell us whether hearings produce moderate bills or whether moderate bills are simply more likely to undergo hearings than extreme bills. A common solution to endogeneity problems is to conduct an instrumental-variable regression, in which a proxy variable correlated with the explanatory variables but unrelated to the outcome variable is used to produce consistent coefficients (Sovey and Green 2011). Unfortunately, I cannot use this solution to determine the causal direction between hearings and bill extremity. The majority party exerts tremendous strategic control over the legislative process (Cox and McCubbins 2005; Sinclair 1997, 2006), and no exogenous factor significantly affects deliberative procedures without also affecting legislative outcomes. Fire alarms in the Rayburn Building are not pulled often enough to provide a measurable disruption of committee sessions. A second possible solution is to estimate a bill's spatial location at introduction and then determine how its spatial location changes, if at all, as it moves through the legislative process. So far, no method has been developed to produce comparable estimates of prefloor and floor locations of legislation. To estimate bill location before committee deliberation requires a very different method of measurement than to estimate bill location at the floor stage.34 The logic of  34  For example, I could estimate a bill’s prefloor location using the mean NOMINATE score of all cosponsors and then estimate a bill’s floor location using the mean NOMINATE score of all yea voters. But these two measures are not comparable, even though both use NOMINATE data to produce estimates. Because bills have few (if any)  98  this kind of analysis, however, can be used in qualitative case studies to trace the direction of change (or lack of change) that a bill undergoes as it moves through the legislative process. By examining the content of a bill before and after hearings, it is possible to determine whether lawmakers have moderated legislation as it travels through the House. If the majority revises a bill to meet (or partially meet) opponents' demands, such activity would demonstrate that a bill has moderated to some extent because of committee hearings. I employ this method here to trace the effects of hearings on legislation.35 Specifically, I examine two Democratic bills and two Republic bills to see what revisions are made in response to arguments raised in hearings. I select bills that appear to be extreme at introduction, based on descriptions by CQ Weekly. Selection of extreme cases is necessary because it maximizes the potential for variance and provides the best opportunity to identify changes made to bills during deliberations. Bills that begin as moderate measures are unlikely to undergo significant revisions to moderate them further. If the estimates in Table 4.4 mainly reflect endogeneity, we should observe few changes to the four bills because of points raised in hearings. Any bill revisions should be minor and unrelated to witness testimony. But if the estimates are free of endogeneity bias, we should see all four bills undergo significant content changes based specifically on suggestions made in committee. For evidence, I use transcripts from the relevant hearings as well as press reports on bill content and bill development. 4.6.1Democratic Bills Democrats were disappointed that they had failed in 1991 to pass legislation banning the sale of all assault rifles (Idelson 1991). Determined to build support for their gun-control cosponsors, estimates based on cosponsorship will be derived from significantly less data than estimates based on floor votes. Furthermore, the decision not to cosponsor is much different from the decision to vote nay on the floor. 35 I recognize the irony in conducting case studies for this project when I noted earlier in this chapter the limitations of case-study research. Of course, I use case studies here to supplement, not to stand in place of, large-N findings.  99  measure, Democrats held over a dozen hearings on firearms regulations in the subsequent Congress. Gun advocates argued in these hearings that the failed bill had included a broad definition of assault weapons, which would have empowered authorities to arrest legitimate gun collectors. Critics also explained that any new ban needed to be very specific about which guns would be prohibited because federal authorities already restricted firearms in ways that breached existing laws ("Federal Firearms Licensing" 1993, 72-73). In 1994, Democrats introduced a revised assault-weapons bill that excluded 661 firearms from the proposed ban. These lists of rifles were included specifically to assuage gun-rights advocates and Republicans (Idelson 1994). In 2007, House Democrats proposed a $14.4 billion bill that would fund construction of a publicly owned, high-speed Amtrak line from New York to Washington. Democrats supported the scheme because rail service would reduce air traffic and thus air pollution. But Republicans opposed the bill and wanted instead to privatize Amtrak. In hearings, experts involved in the construction of rail systems around the world noted that governments typically collaborated with private companies to develop and operate high-speed train systems. As one witness noted, even the Chinese government encouraged "private and international investment" to reduce costs ("International High-Speed Rail" 2007, 15). In response to these hearings, Democrats moderated their bill by including a provision that would privatize some of the services for the high-speed line. Republicans in turn supported the measure (Itkowitz 2008). 4.6.2 Republican Bills Conservative groups objected to Oregon's 1994 legalization of doctor-assisted suicide on religious grounds. When court challenges failed to overturn this law, House Republicans drafted a bill in 1999 to ban euthanasia in all states. But healthcare groups argued in a hearing that the measure would encourage law enforcement to target doctors who treated terminally ill patients. 100  Fear of overzealous prosecution, critics claimed, would discourage physicians from treating pain adequately ("Pain Relief" 1999, 99-100). To address this concern, Republicans added a provision stating that doctors could use controlled substances to ease the pain of patients even if the drugs "increase[d] the risk of death" (quoted in Foerstel 1999). This change satisfied the medical groups that had initially opposed the bill. In 2003, House Republicans sought to revamp the Head Start program, which provided health and educational support to low-income children. The House Republican bill would reorganize this federal program by giving states control over Head Start and allowing them to replace antipoverty services with literacy initiatives (Swindell 2003b). The House bill appealed to conservatives who supported greater local control over education. In a hearing on the proposal, witnesses and minority-party members argued that transferring control of Head Start to the states would turn it into a block-grant program devoid of national standards. States running deficits would likely be tempted to use Head Start funds to balance their budgets ("H.R. 2210" 2003). In response, Republicans moderated the reform by establishing a five-year pilot project that would give eight states—rather than all states—greater control over childhood assistance (Swindell 2003a). Clearly, all four cases show that the majority modifies bills to address specific policy concerns raised by opponents in hearings. Witness testimony highlights provisions that push policy to untenable extremes, and it offers effective policy solutions that the majority has not considered. The majority then responds to the objections by making important content changes that moderate the bills. These four examples suggest that the results on hearings in Table 4.4 are not a product of endogeneity. Extreme bills do go to hearings, and the arguments raised in hearings appear to push the majority into making pragmatic legislative adjustments.  101  4.7 Leadership Strategies in Markups Although hearings tend to foster bill moderation, the majority party can decide to disregard witness testimony and simply use hearings to publicize its bills (DeGregorio 1992). Markups, on the other hand, provide the minority a formal opportunity to offer amendments and alter a bill's language. Extreme bills may be vulnerable to change in markups because the minority can suggest amendments that appeal to majority-party moderates on the committee. The surest way for majority-party leaders to shield extreme bills from amendments is to omit markup sessions. Yet, as Table 4.4 shows, extreme bills are more likely to undergo markups than moderate bills. Why do House majority-party leaders expose extreme bills to committee votes when these bills presumably face a significant risk of being amended? To answer this question, it is necessary to consider not only the location of bills, but also the location of status quos. We can think of the majority proposing bills to shift the status quo in two directions. To illustrate, consider a left-leaning majority party whose members take positions on a onedimensional policy space. First, if the status quo is right of the floor median, the majority will propose a bill that moves the status quo to the left. In fact, majority-party leaders can propose to replace a conservative policy with a very liberal bill and expect significant party support for the shift. After all, the location of the new legislation will be much closer to the majority-party median than the status quo. Second, if a status quo is so extremely liberal that even the majorityparty median opposes it, the majority will offer a bill to shift policy to the right. While a majority of the majority will support this shift, the most extreme members in the majority caucus will reject the proposal because it moves policy away from their ideal points (Cox 2001, 205). In other words, extreme bills likely unify the majority because they shift the status quo from the minority's side of the policy space to the majority's side. Conversely, moderate bills  102  likely divide the majority because they shift the status quo from the majority's extreme to the majority's median. Because estimating status quo locations is exceedingly difficult and limited to bills with very specific data (Jeong 2008; Richman 2011), a direct test of these two hypotheses is not possible. However, if both predictions are correct, we should observe two things. First, we should see the majority voting with a high degree of unity in committee to prevent the minority from altering bills that go to markup. Majority unity indicates that both median and extreme members support the proposed policy change. Second, we should see majority-party leaders omitting or cutting short markups when the most extreme members in their caucus oppose a bill. Objections from extreme members likely indicate that a relatively moderate bill is being considered. To confirm the first expectation, I examine all recorded votes on markup amendments for the bills in my sample developed between 1993 and 2008.36 I find that 95 percent of minorityparty amendments are voted down in markups, and 94 percent of majority-party amendments are adopted without the need for minority-party support. I also find that both Democrats and Republicans, when in the majority, demonstrate similar levels of party discipline in markups. For example, in the 103rd Congress, majority Democrats defeated all but one Republican amendment in committee; in the 104th Congress, majority Republicans defeated every Democratic amendment in committee. Typically, these votes are not close. On 70 percent of the minority's amendments, the majority does not cast a single yea vote; and on 88 percent of the minority's amendments, the majority defeats the measure with more than 90 percent of its committee members voting nay (usually only one member crosses party lines). As these numbers show, the majority party maintains nearly perfect unity in committee to prevent the minority party from  36  I limit my examination to these years because CQ does not report all recorded committee votes before 1993.  103  making bill modifications. All measures, including extreme ones, face little or no threat of undergoing change once majority-party leaders send them to markup. I also examine all CQ stories on the bills in my sample to seek anecdotal evidence of the majority party being rolled on committee votes. Instances of majority defeat in markups would indicate that majority-party leaders do not always know in advance their members' loyalties and preferences. On only one bill in my sample—a 1998 education bill—do I find reports of the minority party winning committee votes over the objections of senior majority-party members. These victories only occurred, however, because "[s]parse Republican attendance" in markup gave minority-party Democrats a brief advantage in numbers ("Higher Education Amendments" 1998). Rare disorganization, then, accounts for the only case of leadership loss in committee. To confirm the second expectation, I search CQ stories on the bills in my sample for cases of majority-party leaders omitting markups when the majority's most extreme members oppose a measure. In a sense, I am looking for the proverbial dog that does not bark, because I am looking for the absence of markups that produces an absence of intraparty conflict. Even so, I find several reports where majority-party leaders circumvent committee deliberations to avoid exposing divisions within their caucus. Two examples demonstrate this leadership strategy. First, in 1990, the National Endowment for the Arts (NEA) faced harsh criticism for awarding grants to artists who depicted homoerotic and sacrilegious images. Moderates and conservatives in the House wanted to restrict the NEA from awarding funds to artists who produced "obscene" work. But liberal Democrats opposed this proposal, arguing that Congress had no role in defining the boundaries of artistic expression and thus had no role in placing restrictions on the NEA. To prevent a caucus split, House Democratic leaders skipped markup sessions and drafted a bill with Republicans outside committee (Lawrence 1990).  104  Second, in 2003, Speaker Dennis Hastert referred the Medicare Prescription Drug bill to two committees, hoping that they would draft legislation that appealed broadly to Republicans. Both committees held acrimonious markups, however, and failed to report bills. Hastert thus rewrote the measure himself behind closed doors with other senior Republicans (Sinclair 2006, 143-144). The problem for Hastert was that his most conservative caucus members viewed the bill as a massive expansion of government spending and threatened to vote against it. Before the floor vote, President Bush tried to appease conservative Republicans by promising them a chance to revise the legislation in conference (Adams and Carey 2003). Enough Republicans finally relented to pass the bill. Taken together, the roll calls and the CQ stories provide confirming evidence for the two hypotheses on committee responses to policy change. Because extreme bills likely represent policy shifts toward the majority's side of the policy space, majority-party leaders anticipate strong caucus support for these measures and see little risk in allowing markups. By contrast, because moderate bills likely shift policy away from the majority's extreme members, leaders try to minimize caucus rifts by negotiating bills in backrooms. Returning to Table 4.4, the correlation between markups and bill extremity appears to be the result of these two leadership strategies. Furthermore, this explanation also clarifies the findings in Chapter 3 on bills that divide the majority caucus. While these divisive bills are not necessarily all moderate bills, majority-party leaders have the same motive to circumvent the markup sessions for these two types of measures: to avoid public forums that will draw attention to the weaknesses in and the disorganization of the majority caucus.  105  4.8 Discussion and Conclusion In this chapter, I have presented two sets of findings. First, using three different measures constructed from newspaper editorials, I provide large-N evidence that bills undergoing regularorder proceedings are less likely to receive negative criticism than bills undergoing truncated deliberations. These results confirm the concerns over one-party lawmaking raised in earlier case studies (e.g., Mann and Ornstein 2006; Hacker and Pierson 2005) and reinforce views that homogenous groups often overlook large problems. In short, the findings here suggest that oneparty deliberations lead to the development of defective legislation. Second, using a new measure of legislative extremity, I show that bills tend to be more extreme if hearings are skipped instead of held. This finding reinforces expectations drawn from previous case studies and it supports the group-polarization research. Surprisingly, I also show that bills tend to be more moderate if markups are omitted instead of maintained. This finding complicates the story of congressional deliberations and suggests that the absence of some formal legislative steps is not always detrimental to the formulation of workable policy. The case studies that I conduct to explain these contradictory findings on hearings and markups confirm that not all prefloor steps are equally important to the performance of bill deliberations. In particular, I show that hearings encourage moderation because they function as venues for genuine bipartisan compromise. Ideas and objections raised in hearings are often (partially) used to revise bills. By contrast, I show that markups are almost always defined by party-line votes. Although bill revisions are supposed to occur in markups, they rarely do. Moreover, in cases where the majority caucus splits over bill content, leaders omit markups to ensure that the committee does not make important legislative changes that the leaders oppose. Unlike the findings on hearings, the findings on markups are very much a result of endogeneity.  106  Why do we see this difference in deliberations between the two stages? Perhaps the majority is more open to compromise at early points of bill development when its ideas on a policy issue are less fixed, and it becomes less flexible the longer it works on a bill. After the committee has discussed the pros and cons of various legislative options in hearings, the majority caucus may have a rather clear sense of which option provides the best solution for its political and policy needs. By the time markups occur, the majority may already have a firm commitment to a bill's design and may be unwilling to make concessions. Additionally, the majority may be more receptive to expert suggestions offered in hearings than to minority-party amendments offered in markup. Expert advice does not come in the form of an either-or proposition. The majority can select how much expert advice it heeds in hearings, and can fully control the magnitude of bill changes that it makes in response to expert testimony. For example, after experts argued in 2003 that Head Start should be left unaltered, Republican House leaders decided to implement a small pilot project rather than a national reform. They did not fully follow the experts and abandon plans to reorganize Head Start; instead, they simply reduced the scope of the reorganization. Because markup amendments must be adopted or rejected, the majority does not have the same flexibility making bill changes in the markup stage as it does in the hearings stage. Whenever the minority suggests bill modification through amendment, the majority faces an allor-nothing choice. It may be less risky for the majority to vote down nearly all minority-party amendments than to allow a dynamic amending process that could potentially lead to significant and, from the perspective of majority-party leaders, detrimental bill changes. Even so, it is puzzling that majority-party moderates on the committee rarely cross party lines to support minority-party amendments that shift bills closer to their ideal points. Perhaps  107  majority-party moderates calculate that it is not worth crossing party lines to make bill changes that can be undone by leaders after the committee reports the bill or by amendments offered on the floor. Alternatively, majority-party leaders may use a combination of threats and promises to discourage their caucus members from voting sincerely in committee. Whatever the explanation, future research that examines policy formulation in the House should focus greater attention on hearings than on markups. Finally, although the results of this chapter demonstrate that markups are often devoid of serious deliberations, more research on markups is required. Why do committees schedule only one markup for some bills and multiple bill-drafting sessions for others? For bills that undergo extensive markups, does the majority party demonstrate consistent party unity through the entire process and vote down all minority-party amendments? The research in this chapter suggests the answer is yes. Yet, if the majority maintains near-perfect unity in committee to prevent bill revisions, why does it ever bother to hold more than one markup session? If the majority makes all key policy decisions prior to markup, why does it not simply hold one formal session for the purpose of voting the bill out of committee and get on with the floor debate? What are the strategic and deliberative purposes of lengthy markups? The literature does not provide answers to these questions, suggesting that there is still much work to be done on committee operations.  108  5. Conclusion This project set out to answer two basic questions about deliberations in the House: When does the majority seize control of bill development, and what are the consequences of this control? I argued that answers to these questions are crucial to understanding how Congress operates and how effective (or ineffective) it is in developing policy solutions in the partisan era. But as I also noted, despite deliberation being a central activity in Congress, scholars have generally overlooked the study of congressional deliberation (Quirk and Bendix 2011). Much of what we know about deliberative partisanship in the House comes from case-study work by Sinclair (1983, 1997, 2006) and Mann and Ornstein (2006). So in this project, I was interested not only in answering the two questions above, but also in developing large-N strategies for assessing congressional performance and legislative outcomes. I showed that several political factors encourage the House majority to take increasing control of bill deliberations. As the two parties polarize and as electoral pressures increase, the majority caucus allows its leaders with greater frequency to draft bills behind closed doors, rush measures to the floor, and use restrictive rules to limit amendments and debate. I also showed that one-party lawmaking is more likely in periods of divided government than in periods of unified government. The majority likely calculates that there is little reason to compromise with a minority party that has no ability to obstruct legislation. From the House majority's standpoint, it only makes sense to compromise with actors who have power to dispose of bills: the Senate and the president. As well, I showed that bills involving policy areas used by the parties to define their brands are more likely to undergo limited deliberations than other types of bills. Party-brand bills include tax measures, social-welfare legislation, and culture-war items. Importantly, these types  109  of bills have polarized the parties for several decades, not just for the last few years. Therefore, we should not expect House Democrats and Republicans to collaborate on reforming the tax code, alleviating poverty, or regulating firearms, to name a few examples. Although the parties share important common interests on these issues, the parties have incentives to avoid policy compromise because each one defines its brand, and thus maintains its electoral base, by taking the opposite position. I also showed that bills dividing the majority party tend to undergo hearings but not markups. Because hearings involve only information gathering, not bill revisions, majority-party leaders likely see limited risk in holding these panel meetings. However, they avoid markups on bills that split the majority caucus, because bill-drafting sessions can lead to significant changes to provisions that the leaders oppose. These findings are notable because they suggest that majority-party leaders do not always circumvent committees to exclude the minority party; sometimes they omit panels to prevent their own caucus members from drafting or revising bills. Finally, I showed that one-party bills are more likely to contain defects than two-party bills. This result supports concerns raised by many congressional observers who lament the deterioration of regular-order proceedings (Ahuja 2008; Eilperin 2006; Hamilton 2004; Mann and Ornstein 2006). It also challenges arguments made by some commentators that partisan lawmaking is preferable to bipartisan collaboration because it produces coherent policy, not messy compromise (see e.g., Begala 2009; Cost 2009). I also showed that bills that skip hearings are more ideologically extreme than bills that undergo hearings. It appears that policy arguments made at this stage of deliberations often persuade the majority to make significant changes to their legislative proposals. In short, hearings matter.  110  However, contrary to expectations, I found that the majority tends to omit markups on ideologically moderate legislation. Extreme members of the majority caucus often oppose moderate bills, and they make public objections against these measures or they try to block their progress in committee. The disruptive actions taken by extreme caucus members push majorityparty leaders to omit markups and to take control of bill writing themselves. These findings, like the previous findings on bills that split the majority caucus, demonstrate that leaders do not always circumvent committees for partisan reasons. I also showed that when markups are held on priority legislation, the minority party has little opportunity to shape or add provisions. The majority typically rejects all minority-party amendments in committee, indicating that billdrafting sessions serve little or no genuine deliberative function. This project contributes to the congressional literature in several ways. First, I have developed new measures for and have presented new data on prefloor deliberative procedures in the House. Congressional scholars have long used floor votes to examine legislative decisionmaking, but they have focused much less attention on prefloor activity. This project has shown that important data can be collected on committee operations (and the lack of committee operations). Second, I have provided explanations of why parties exclude each other from bill development. Many previous studies on partisanship have focused much attention on the growing ideological gap between the parties to explain increased leadership control over bill content and the legislative agenda. This project has demonstrated that ideological polarization is a crucial factor in explaining partisan lawmaking, but hardly the only factor. Electoral pressures, policy issues, and the preferences of committee leaders all affect the level of one-party control that bills undergo. Third, and perhaps most noteworthy, I have developed multiple measures of legislative quality and have shown that scholars can evaluate a large number of bills, on a wide  111  range of policy issues, without injecting their own biases into the analysis. The measures that I have developed here will hopefully encourage other scholars to seek creative ways to assess policy outcomes. After all, one of the motivations for assessing the quality of legislative deliberation is to determine whether it affects the quality of the legislation itself (Quirk and Bendix 2011, 568-569). 5.1 Future Directions While this project builds our understanding of deliberative partisanship in Congress, there is certainly more work to be done on the subject. First, this study presented evidence indicating that markups on major bills are largely symbolic exercises, where little serious deliberations occur. This finding raises serious questions for which the congressional literature does not have answers. Indeed, the presence of multiple markup sessions on major legislation is perplexing because, as I showed, committee votes on priority bills tend to be cast overwhelmingly along party lines. If the two parties enter these panel meetings with little interest in compromise, they should anticipate little opportunity for bill revisions and thus see little need for lengthy committee discussions. Why are multiple markups scheduled and what, beyond party-lines votes, occur in these sessions? One possible answer is that majority-party leaders encourage extensive committee meetings on bills that deal with very popular policies. Multiple markups in such cases may provide majority-party members extended opportunity to claim credit and to show constituents that they are active on important policy matters. Furthermore, multiple markups give their committee members an opportunity to demonstrate their expertise and policy influence by voting for or against key amendments. This kind of activity may be crucial for satisfying important interest groups. 112  Another possibility is that the committees are conducting serious deliberations in these panel sessions, but that markup votes do not capture the kind of policymaking activity that occurs. On particularly complex legislation, the majority may want to put a bill through extensive committee meetings to see whether the minority party uncovers any serious weaknesses before the measure reaches the floor. The majority may not adopt minority-party amendments in the course of these sessions, but it may make revisions between or after markup meetings to eliminate defects. Furthermore, the majority may want to learn at the prefloor stage what the minority's main objections are to the bill, so that it can produce persuasive counterarguments in time for the floor debate. In any case, future research needs to explain not only the variation in the number of markup meetings but also the purpose (deliberative or otherwise) of holding extensive bill-writing sessions. Markups are just one area of congressional deliberation that requires further study. There are many others. For example, the congressional literature has not fully considered how parties deliberate when they take control of bill development (Quirk and Bendix 2011, 569). Case studies have suggested that parties overlook major concerns and focus only on policy information that reinforces their own policy views. My findings here certainly support the casestudy research. Even so, there is no reason to assume that all party policy groups, under all conditions, operate the same way, or that they all produce substantively problematic bills. They may perform narrow or comprehensive deliberations depending on a number of factors−including the size of the party policy group, the amount of time allotted for party deliberation, and the types of outside groups that contribute to closed-door discussions. Furthermore, Republican and Democratic majority leaders may conduct party deliberations differently. Comparisons between the two parties' ad hoc groups could reveal differences in the  113  internal workings of the parties themselves and thus the process by which they arrive at important policy decisions. To explore these issues would require researchers to interview ad hoc group members and their staff aides, examine briefing papers and, ideally, observe party policy groups in action as they develop legislation. Admittedly, such research would pose tremendous challenges to political scientists. It is obviously much easier to collect data on legislative activity than to gain access to party leaders and monitor their backroom discussions. However, such research would reveal the kinds of data, evidence, and arguments that these party groups consider as they draft legislation. Do they rely mostly on partisan sources of information, or do they make serious efforts to collect data from nonpartisan sources (such as the Congressional Research Service)? The question of how the majority decides policy is becoming increasingly important. Because both Republicans and Democrats have incentives to adopt highly combative strategies when in the minority, the majority caucus may have little choice but to deliberate on its own. Another issue to examine is whether differences in prefloor deliberations affect the tone and content of floor debates. Of particular concern is whether floor debates on party-drafted bills are more likely to overlook major considerations than debates on committee-drafted bills. The quality of floor debate is important because it reveals what lawmakers know or do not know about a bill before the final legislative vote (Mucciaroni and Quirk 2006, 5-6). In turn, the quality of floor debate may indicate the quality of the legislation under consideration. A large number of false and unsubstantiated claims on the floor−especially if made by bill sponsors, cosponsors, and others involved in prefloor deliberations−may be a clear sign that a bill contains serious defects. Severe objections and dire policy predictions may also be strong indicators of  114  problematic legislation. Ultimately, measures of floor quality and floor criticism would help researchers determine whether important differences exist between one- and two-party bills. The above suggestions do not represent a comprehensive list of research topics on deliberative partisanship that need exploring. Far from it. Although I have focused on the House in this project, the same or similar questions that I have posed here could be asked about Senate deliberations in the partisan era. Moreover, legislative deliberations are not the only kinds of deliberations that Congress conducts. Partisanship no doubt affects oversight and confirmation hearings; thus, research that examines and assesses Congress's performance of these deliberative duties is needed, as well.  115  References Abramowitz, Alan I. 2010. The Disappearing Center. New Haven, CT: Yale University Press. Abramowitz, Alan I., and Kyle L. Saunders. 2008. "Is Polarization A Myth?" Journal of Politics 70 (2): 542-555. Adams, Greg D. 1997. "Abortion: Evidence of an Issue Evolution." American Journal of Political Science 41 (3): 718-737. Adams, Rebecca, and Mary Agnes Carey. 2003. "Compromise Will Come Hard In MedicareOverhaul Conference." CQ Weekly, June 28, 1611. Adler, E. Scott. 2000. "Constituency Characteristics and the 'Guardian' Model of Appropriations Subcommittees, 1959-1998." American Journal of Political Science 44 (1): 104-114. Adler, E. Scott. 2002. Why Congressional Reforms Fail: Reelection and the House Committee System. Chicago: University of Chicago. Adler, E. Scott, and John S. Lapinski. 1997. "Demand-Side Theory and Congressional Committee Composition: A Constituency Characteristic Approach." American Journal of Political Science 41 (3): 895-918. Adler, E. Scott, and John Wilkerson. 2009. "Wait Until Next Year: The Timing and Degree of Legislative Change." Paper presented at the Annual Meeting of the Midwest Political Science Association, Chicago, IL, April 2-5. Ahuja, Sunil. 2008. Congress Behaving Badly: The Rise of Partisanship and Incivility and the Death of Public Trust. Westport, CT: Praeger. Aldrich, John H. 1995. Why Parties? The Origin and Transformation of Political Parties in America. Chicago: University of Chicago Press. Aldrich, John H., and David W. Rohde. 1997. "The Transition to Republican Rule in the House: Implications for Theories of Congressional Politics." Political Science Quarterly 112 (4): 541-567. Aldrich, John H., and David W. Rohde. 2000a. "The Consequences of Party Organization in the House: The Role of the Majority and Minority Parties in Conditional Party Government." In Polarized Politics: Congress and the Partisan Era, eds. Jon R. Bond and Richard Fleisher. Washington, DC: CQ Press, 31-72. Aldrich, John H., and David W. Rohde. 2000b. "The Republican Revolution and the House Appropriations Committee." Journal of Politics 62 (1): 1-33. Aldrich, John H., and David W. Rohde. 2009. "Congressional Committees in a Continuing Partisan Era." In Congress Reconsidered, 9th ed., eds. Lawrence B. Dodd and Bruce I. Oppenheimer. Washington, DC: CQ Press, 217-240. Allison, Paul D. 1999. Logistic Regression Using the SAS System: Theory and Application. Cary, NC: SAS Institute. Altemeyer, Bob. 1996. The Authoritarian Specter. Cambridge, MA: Harvard University Press. Arnold, R. Douglas. 1990. The Logic of Congressional Action. New Haven, CT: Yale University Press. "Audit: U.S. lost track of $9 billion in Iraq funds." 2005. CNN.com, January 30, <http://edition.cnn.com/2005/WORLD/meast/01/30/iraq.audit/>. Bafumi, Joseph., and Robert Y. Shapiro. 2009. "A New Partisan Voter." Journal of Politics 71 (1): 1-24.  116  Balla, Steven, J., Eric D. Lawrence, Forrest Maltzman, and Lee Sigelman. 2002. "Partisanship, Blame Avoidance, and the Distribution of Legislative Pork." American Journal of Political Science 46 (3): 515-525. Bartels, Larry M. 2008. Unequal Democracy: The Political Economy of the New Gilded Age. Princeton, NJ: Princeton University Press. Bartels, Larry M., Hugh Heclo, Rodney E. Hero, and Lawrence R. Jacobs. 2005. "Inequality and American Governance." In Inequality and American Democracy: What We Know and What We Need to Learn, eds. Lawrence R. Jacobs and Theda Skocpol. New York: Russell Sage, 88-155. Baughman, John. 2006. Common Ground: Committee Politics in the U.S. House of Representatives. Stanford, CA: Stanford University Press. Begala, Paul. 2009. "Forget Bipartisanship." The Daily Beast, November 8, http://www.thedailybeast.com/blogs-and-stories/2009-11-08/forget-bipartisanship.html. Bennett, W. Lance. 1991. "Toward a Theory of Press-State Relations in the United States." Journal of Communication 40 (2): 103-125. Bessette, Joseph M. 1994. The Mild Voice of Reason: Deliberative Democracy and American National Government. Chicago: University of Chicago Press. Binder, Sarah A. 2003. Stalemate: Causes and Consequences of Legislative Gridlock. Washington, DC: Brookings Institution. Binder, Sarah A., Thomas E. Mann, Norman J. Ornstein, and Molly Reynolds. 2009. "Mending the Broken Branch: Assessing the 110th Congress, Anticipating the 111th." Brookings Institution Report. Bond, Jon R., and Richard Fleisher. 1990. The President in the Legislative Arena. Chicago: University of Chicago Press. Bowen, Harry P., and Margarethe F. Wiersema. 2004. "Modeling Limited Dependent Variables: Methods and Guidelines for Researchers in Strategic Management." In Research Methodology in Strategy and Management, Vol. 1, eds. David J. Ketchen and Donald D. Bergh. Oxford, UK: Elsevier, 87-134. Brady, David W., John Ferejohn, and Laurel Harbridge. 2008. "Polarization and Public Policy: A General Assessment." In Red and Blue Nation? Consequences and Correction of America's Polarized Politics, Vol. 2, eds. Pietro S. Nivola and David Brady. Washington, DC: Brookings Institution, 185-216. Brady, David, and Edward P. Schwartz. 1995. "Ideology and Interests in Congressional Voting: The Politics of Abortion in the U.S. Senate." Public Choice 84 (1): 25-48. Brown, William Holmes, and Charles W. Johnson, III. 2003. House Practice: A Guide to the Rules, Precedents, and Procedures of the House, 108th Congress, 1st Session. Washington, DC: U.S. GPO. "Budgetary Failure; Bush rightly wants fiscal discipline, but remember saying 'All hat, no cattle'?" 2007. Omaha World-Herald, October 22, B06. Calmes, Jacqueline. 1985. "Trade-off for Pentagon Cut? Social Security Hikes Divide House Democrats." CQ Weekly, May 11, 875. Campbell, James E. 1982. "Cosponsoring Legislation in the U.S. Congress." Legislative Studies Quarterly 7 (3): 415-422. Carey, Mary Agnes. 2002. "Complex Drug Bill May Prove More Difficult for GOP to Sell." CQ Weekly, June 22, 1667.  117  Carey, Mary Agnes. 2003. "House Republicans Struggle to Create Workable Drug 'Reimportation' Bill." CQ Weekly, July 19, 1828. Carey, Mary Agnes, and Sue Kirchhoff. 1998. "GOP's Managed Care Bill Rushes Through House." CQ Weekly, July 25, 2007. Carson, Jamie L., Charles J. Finocchiaro, and David W. Rohde. 2002. "Partisanship, Consensus, and the Committee-Floor Divergence: A Comparison of Member Behavior in the 96th and 104th Congresses." American Politics Research 30: 3-33. Carson, Jamie L., Charles J. Finocchiaro, and David W. Rohde. 2010. "Consensus, Conflict, and Partisanship in the House Decision Making: A Bill-Level Examination of Committee and Floor Behavior." Congress & the Presidency: A Journal of Capital Studies 37 (3): 231253. Clinton, Joshua D., and John Lapinski. 2006. “Measuring Legislative Accomplishment, 18771994.” American Journal of Political Science 50 (1): 232-249. Clinton, Joshua D., and Adam Meirowitz. 2001. "Agenda Constrained Legislator Ideal Points and the Spatial Voting Model." Political Analysis 9: 1-18. Cohen, Richard. 1999. "Crackup of the Committees." National Journal, July 31, 2210-2217. Cost, Jay. 2009. "Three Cheers for Partisanship!" Real Clear Politics, February 16, http://www.realclearpolitics.com/horseraceblog/2009/02/three_cheers_for_partisanship_1 .html. Cox, Gary W. 2001. "Agenda Setting in the U.S. House: A Majority-Party Monopoly?" Legislative Studies Quarterly 26 (2): 185-210. Cox, Gary W., and Matthew D. McCubbins. 1993. Legislative Leviathan: Party Government in the House. Berkley, CA: University of California Press. Cox, Gary W., and Matthew D. McCubbins. 2005. Setting the Agenda: Responsible Party Government in the U.S. House of Representatives. New York: Cambridge University Press. Cox, Gary W., and Keith T. Poole. 2002. "On Measuring Partisanship in Roll-Call Voting: The U.S. House of Representatives, 1877-1999." American Journal of Political Science 46 (3): 477-489. Crewe, Ivor, and Anthony King. 2010. "On the Causes of Domestic Policy Failure in U.K. Government." Presented at the annual meeting of the American Political Science Association, Washington, DC, September 2-5. Davidson, Roger H., and Walter Oleszek. 1994. Congress and Its Members, 4th ed. Washington, DC: CQ Press. Davidson, Roger H., Walter J. Oleszek, and Thomas Kephart. 1988. "One Bill, Many Committees: Multiple Referrals in the U.S. House of Representatives." Legislative Studies Quarterly 13 (1): 3-28. Dean, John. 2006. Conservatives without a Conscience. New York: Viking Adult. Deering, Christopher J., and Steven S. Smith. 1997. Committees in Congress, 3rd Edition. Washington, DC: CQ Press. DeGregorio, Christine. 1992. "Leadership Approaches in Congressional Committee Hearings." Western Political Quarterly 45 (4): 971-983. Dion, Douglas, and John Huber. 1996. "Procedural Choice and the House Committee on Rules." Journal of Politics 58 (1): 25-53. Dlouhy, Jennifer A. 2005. "Immigration Fight Looms." CQ Weekly, January 31, 246.  118  Eilperin, Juliet. 2006. Fight Club Politics: How Partisanship Is Poisoning the House of Representatives. Lanham, MD: Rowan & Littlefield. Epstein, Edward. 2010. "Dusting off Deliberation." CQ Weekly, June 14, 1436. Espino, Rodolfo, and Michael M. Franz. 2004. "Retesting Committee Composition Hypotheses for the U.S. Congress." Political Analysis 12 (2): 196-198. Esterling, Kevin M. 2004. The Political Economy of Expertise: Information and Efficiency in American National Politics. Ann Arbor, MI: University of Michigan. Esterling, Kevin M. 2011. "'Deliberative Disagreement' in U.S. Health Policy Committee Hearings." Legislative Studies Quarterly 36 (2): 169-198. Evans, C. Lawrence. 2011. "Congressional Committees." In The Oxford Handbook of the American Congress, eds. Eric Schickler and Francis E. Lee. New York: Oxford University Press, 396-425. Evans, C. Lawrence, and Claire E. Grandy. 2009. "The Whip Systems of Congress." In Congress Reconsidered, 9th ed., eds. Lawrence B. Dodd and Bruce I. Oppenheimer. Washington, DC: CQ Press, 189-215. Fenno, Richard J., Jr. 1973. Congressmen in Committees. Boston: Little, Brown. Fenno, Richard F., Jr. 1997. Learning to Govern: An Institutional View of the 104th Congress. Washington, D.C.: Brookings Institution Press. "Federal Firearms Licensing." 1993. Hearing before the Subcommittee on Crime and Criminal Justice of the Committee on the Judiciary, House of Representatives, 103rd Congress, 1st Session, June 17. Fiorina, Morris P. 1974. Representatives, Roll Calls, and Constituencies. Lexington, MA: D.C. Heath. Foerstel, Karen. 1999. "Assisted-Suicide Ban Approved by House Judiciary Panel; Senate Filibuster Threatened." CQ Weekly, September 18, 2173. Galston, William A., and Pietro S. Nivola. 2006. "Delineating the Problem." Red and Blue Nation?: Characteristics and Causes of America’s Polarized Politics, Vol. 1, eds. Pietro S. Nivola and David Brady. Washington, DC: Brookings Institution, 1-47. Garand, James C. 1988. "The Socialization of Partisan Legislative Behavior: An Extension of Sinclair’s Task Force Socialization Thesis." Western Political Quarterly 41 (2): 391-400. Garand, James C., and Kathleen M. Clayton. 1986. "Socialization to Partisanship in the U.S. House: The Speaker’s Task Force." Legislative Studies Quarterly 11 (3): 409-428. Gilligan, Thomas W., and Keith Krehbiel. 1987. "Collective Decisionmaking and Standing Committees: An Informational Rationale for Restrictive Amendment Procedures." Journal of Law, Economics, and Organization 3 (2): 287-335. Gottlieb, Bruce. 1999. "Who Writes Unsigned Editorials?" The Slate, March 8, <http://www.slate.com/id/1002028/>. Grofman, Bernard, William Koetzle, and Anthony J. McGann. 2002. "Congressional Leadership 1965-96: A New Look at the Extremism versus Centrality Debate." Legislative Studies Quarterly 27 (1): 87-105. Groseclose, Tim. 1994. "Testing Committee Composition Hypotheses for the U.S. Congress." Journal of Politics 56 (2): 440-458. Groseclose, Tim, and Jeffery Milyo. 2005. "A Measure of Media Bias." Quarterly Journal of Economics 120 (4): 1191-1237. Grynaviski , Jeffrey D. 2010. Partisan Bonds: Political Reputations and Legislative Accountability. New York: Cambridge University Press. 119  Gulati, Girish J. 2004. "Revising the Link between Electoral Competition and Policy Extremism in the U.S. Congress." American Political Research 32 (5): 495-520. Guttmann, Amy, and Dennis Thompson. 2012. The Spirit of Compromise: Why Governing Demands It and Campaigning Undermines It. Princeton, NJ: Princeton University Press. Hacker, Jacob S., and Paul Pierson. 2005. Off Center: The Republican Revolution and the Erosion of American Democracy. New Haven, CT: Yale University Press. Hacker, Jacob S., and Paul Pierson. 2007."Tax Politics and the Struggle over Activist Government.” In The Transformation of American Politics: Activist Government and the Rise of Conservatism, eds. Paul Pierson and Theda Skocpol. Princeton, NJ: Princeton University Press, 256-280. Hacker, Jacob S., and Paul Pierson. 2010. Winner-Take-All Politics: How Washington Made the Rich Richer—And Turned Its Back on the Middle Class. New York: Simon & Schuster. Hall, Richard L. 1987. "Participation and Purpose in Committee Decision Making." American Political Science Review 81 (1): 105-128. Hall, Richard L. 1996. Participation in Congress. New Haven, CT: Yale University Press. Hall, Richard L., and C. Lawrence Evans. 1990. "The Power of Subcommittees." Journal of Politics 52 (2): 335-355. Hall, Richard L., and Bernard Grofman, 1990. "The Committee Assignment Process and the Conditional Nature of Committee Bias." American Political Science Review 84 (4): 11491166. Hamilton, Lee. 2004. How Congress Works and Why We Should Care. Bloomington, ID: Indiana University Press. Harbridge, Laurel. 2011. "Congressional Agenda Control and the Decline of Bipartisan Cooperation." Paper presented at the 69th Annual Meeting of the Midwest Political Science Association, Chicago, IL, The Palmer House, March 31-April 3. Han, Hahrie, and David W. Brady. 2007. "A Delayed Return to Historical Norms: Congressional Party Polarization after the Second World War." British Journal of Political Science 37: 505-531. Hayes, Danny. 2005. "Candidate Qualities through a Partisan Lens: A Theory of Trait Ownership." American Journal of Political Science 49 (4): 908-923. Hayes, Danny. 2008. "Party Reputations, Journalistic Expectations: How Issue Ownership Influences Election News." Political Communication 25: 377-400. Hetherington, Marc. 2009. "Putting Polarization in Perspective." British Journal of Political Science 39: 413-448. Hibbing, John R., and Eric Tiritilli. 2000. "Public Disapproval of Congress Can Be Dangerous to Majority Party Candidates: The Case of 1994." In Continuity and Change in House Elections, eds. David W. Brady, John Cogan, and Morris P. Fiorina. Stanford, CA: Stanford University Press, 110-129. Herrnson, Paul S. 1998. "Directing 535 Leading Men and Leading Ladies: Party Leadership in the Modern Congress." In Great Leadership: The American Congress in the 1990s, eds. Herbert F. Weisberg and Samuel C. Patterson. New York: Cambridge University Press, 109-134. "Higher Education Amendments." 1998. CQ Committee Coverage, March 18, CQ.com. Hong, Lu, and Scott E. Page. 2001. "Problem Solving by Heterogeneous Agents." Journal of Economic Theory 97 (1): 123-163.  120  Hong, Lu, and Scott E. Page. 2004. "Groups of Diverse Problem Solvers Can Outperform Groups of High-Ability Problem Solvers." Proceedings of the National Academy of Sciences of the United States of America 101 (46): 16385-16389. "House's Rush to Judgment." 2007. The Post and Courier, July 15, A10. "H.R. 2210: The School Readiness Act of 2003." 2003. Hearing before the Subcommittee on Education Reform of the Committee on Education and the Workforce, House of Representatives, 108th Congress, 1st Session, Washington, DC, June 3. Idelson, Holly. 1991. "House Members Duel on Crime: Assault-Gun Ban Rejected." CQ Weekly, October 19, 3038. Idelson, Holly. 1994. "House Nearing Showdown on Assault Weapons Ban." CQ Weekly, April 30, 1069. "International High-Speed Rail Systems." 2007. Hearing before the Subcommittee on Railroads, Pipelines, and Hazardous Materials of the Committee on Transportation and Infrastructure, House of Representatives, 110th Congress, 1st Session, April 19. "Iraq Reconstruction Bonanza." 2003. New York Times, October 1, A22. Itkowitz, Colby. 2008. "House Reaches Compromise on Amtrak." CQ Weekly, June 16, 1622. Jacobs, Lawrence M., and Robert Y. Shapiro. 2000. Politicians Don't Pander: Political Manipulation and the Loss of Democratic Responsiveness. Chicago: University of Chicago Press. Janis, Irving L. 1982. Groupthink: Psychological Studies of Policy Decisions and Fiascos, 2nd edition. Boston: Houghton Mifflin. Jeong, Gyung-Ho. 2008. "Testing the Predictions of the Multidimensional Spatial Voting Model with Roll Call Data." Political Analysis 16: 179-196. Jeong, Gyung-Ho, Gary J. Miller, Camilla Schofield, and Itali Sened. 2011. "Cracks in the Opposition: Immigration as a Wedge Issue for the Reagan Coalition." American Journal of Political Science 55 (3): 511-525. Jones, Bryan D., and Walter Williams. 2008. The Politics of Bad Ideas: The Great Tax Cut Delusion and the Decline of Good Government in American. New York: Pearson Longman. Jones, David R. 2001. "Party Polarization and Legislative Gridlock." Political Research Quarterly 54 (1): 125-141. Jones, David R. 2010. "Partisan Polarization and Congressional Accountability in House Elections." American Journal of Political Science 54 (2): 323-337. Karol, David. 2009. Party Position Change in American Politics: Coalition Management. New York: Cambridge University Press. Katznelson, Ira, and John S. Lapinski. 2006. "The Substance of Representation: Studying Policy Content and Legislative Behavior." In The Macropolitics of Congress, eds. E. Scott Adler and John S. Lapinski. Princeton, NJ: Princeton University Press, 96-128. Keefe, William J., and Marc J. Hetherington. 2003. Parties, Politics, and Public Policy in America, 9th ed. Washington, DC: Congressional Quarterly Press. Kennedy, Peter E. 2006. "Oh No! I Got the Wrong Sign! What Should I Do?" Journal of Economic Education 36 (1): 77-92. Kessler, Daniel, and Keith Krehbiel. 1996. "Dynamics of Cosponsorship." American Political Science Review 90 (3): 555-566. Koger, Gregory. 2003. "Position Taking and Cosponsorship in the U.S. House." Legislative Studies Quarterly 28 (2): 225-246. 121  Krehbiel, Keith. 1990. "Are Congressional Committees Composed of Preference Outliers?" American Political Science Review 84 (1): 149-163. Krehbiel, Keith. 1991. Information and Legislative Organization. Ann Arbor, MI: University of Michigan Press. Krehbiel, Keith. 1993. "Where's the Party?" British Journal of Political Science 23 (2): 235-266. Krehbiel, Keith. 2008. "Comment on Chapter Two." In Red and Blue Nation? Consequences and Correction of America's Polarized Politics, Vol. 2, eds. Pietro S. Nivola and David Brady. Washington, DC: Brookings Institution, 93-103. Krugman, Paul. 2003. "Who's Sordid Now?" New York Times, September 30, A19. Krutz, Glen S. 2001. Hitching a Ride: Omnibus Legislation in the U.S. Congress. Columbus, OH: Ohio State University Press. Krutz, Glen S. 2005. "Issues and Institutions: 'Winnowing' in the U.S. Congress." American Journal of Political Science 49 (2): 313-326. Lakoff, George. 2002. Moral Politics: How Liberals and Conservatives Think. Chicago: University of Chicago Press. Lawrence, Christine C. 1990. "Obscenity Debate: Off-Stage Negotiations Begin Over Curbs on Arts Grants." CQ Weekly, June 16, 1894. Layman, Geoffrey C., and Edward G. Carmines. 1997. "Cultural Conflict in American Politics: Religious Traditionalism, Postmaterialism, and U.S. Political Behavior." Journal of Politics 59 (3): 751-777. Lee, Frances E. 2009. Beyond Ideology: Politics, Principles, and Partisanship in the U.S. Senate. Chicago: University of Chicago Press. Lindaman, Kara, and Donald P. Haider-Markel. 2002. "Issue Evolution, Political Parties, and the Culture Wars." Political Research Quarterly 55 (1): 91-110. Lipinski, Daniel. 2009. "Navigating Congressional Policy Processes: The Inside Perspective on How Laws Are Made." In Congress Reconsidered, 9th ed., eds. Lawrence C. Dodd and Bruce I. Oppenheimer. Washington, DC: CQ Press, 337-360. Lublin, David. 2005. "The Strengthening of Party and Decline of Religion in Explaining Congressional Voting Behavior on Gay and Lesbian Issues." PS: Political Science and Politics 38 (2): 241-245. Mann, Thomas E., and Norman J. Ornstein. 2006. The Broken Branch: How Congress Is Failing America and How to Get It Back on Track. New York: Oxford University Press. Mann, Thomas E., and Norman J. Ornstein. 2012. "Let's Just Say It: The Republicans Are the Problem." Washington Post, April 27. Matthews, Chris. 2001. Now, Let Me Tell You What I Really Think. New York: Free Press. McCarty, Nolan. 2007. "The Policy Effects of Political Polarization." In The Transformation of American Politics: Activist Government and the Rise of Conservatism, eds. Paul Pierson and Theda Skocpol. Princeton, NJ: Princeton University Press, 223-255. McCarty, Nolan, Keith Poole, and Howard Rosenthal. 2006. Polarized America: The Dance of Ideology and Unequal Riches. Cambridge, MA: MIT Press. McGann, Anthony J., Bernard Grofman, and William Koetzle. 2002. "Why Party Leaders Are More Extreme Than Their Members: Modeling Sequential Elimination Elections in the U.S. House of Representatives." Public Choice 113 (3): 337-356. Morris, Jonathan S., and Clawson, Rosalee A. 2005. "Media Coverage of Congress in the 1990s: Scandals, Personalities, and the Prevalence of Policy and Process." Political Communication 22 (3): 297-313. 122  Mouw, Calvin, and Michael MacKuen. 1992. "The Strategic Agenda in Legislative Politics." American Political Science Review 86 (1): 87-105. Mucciaroni, Gary, and Paul J. Quirk. 2006. Deliberative Choices: Debating Public Policy in Congress. Chicago: University of Chicago Press. Nather, David. 2000. "House Passes Bill to Help Revive Inner City, Rural Economies." CQ Reports, July 29, 1882. Nelson, Garrison. 2005. Committees in the U.S. Congress, 1947-1992: House, 97th to 102th Congresses, September 15. Nye, Mary Alice. 1991. "The U.S. Senate and Civil-Rights Roll Calls." Western Political Science Quarterly 44 (4): 971-986. Nye, Mary Alice. 1993. "Changing Support for Civil Rights: House and Senate Voting, 19631988." Political Research Quarterly 46 (4): 799-822. O’Connor, Karen, and Lee Epstein. 1983. "The Rise of Conservative Interest Group Litigation." Journal of Politics 45 (2): 479-489. Oldmixon, Elizabeth A. 2002. “Culture Wars in the Congressional Theater: How the U.S. House of Representatives Legislates Morality, 1993-1998.” Social Science Quarterly 83 (3): 775-788. Oleszek, Walter J. 1996. Congressional Procedures and the Policy Process, 4th ed. Washington, DC: CQ Press. Owens, John E. 1997. "The Return of Party Government in the U.S. House of Representatives: Central Leadership-Committee Relations in the 104th Congress." British Journal of Political Science 27: 247-272. Page, Benjamin I. 2002. "The Semi-Sovereign Public." In Navigating Public Opinion: Polls, Policy, and the Future of American Democracy, eds. Jeff Manza, Fay Lomax Cook, and Benjamin I. Page. New York: Oxford University Press, 325-344. Page, Scott E. 2007. The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies. Princeton, NJ: Princeton University Press. "Pain Relief Promotion Act of 1999." 1999. Hearing before the Subcommittee on the Constitution of the Committee on the Judiciary, House of Representatives, One Hundred Sixth Congress, First Session, June 24. Pearson, Kathryn, and Eric Schickler. 2009. "The Transition to a Democratic Leadership in a Polarized House." In Congress Reconsidered, 9th ed., eds. Lawrence B. Dodd and Bruce I. Oppenheimer. Washington, DC: CQ Press, 165-188. Poole, Keith T., and Howard Rosenthal. 2007. Ideology and Congress. New Brunswick, NJ: Transaction Publishers. Quirk, Paul J. 2005. "Deliberation and Decision Making." In The Legislative Branch, eds. Paul J. Quirk and Sarah A. Binder. New York: Oxford University Press, 314-348. Quirk, Paul J., and William Bendix. 2011. "Deliberation in Congress." In The Oxford Handbook of the American Congress, eds. Eric Schickler and Francis E. Lee. New York: Oxford University Press, 550-574. Richman, Jesse. 2011. "Parties, Pivots, and Policy: The Status Quo Test." American Political Science Review 105 (1): 151-165. Roberts, Jason M. 2010. "The Development of Special Orders and Special Rules in the U.S. House, 1881-1937." Legislative Studies Quarterly 35 (3): 307-336.  123  Roberts, Jason M., and Steven S. Smith. 2003. "Procedural Contexts, Party Strategy, and Conditional Party Voting in the U.S. House of Representatives, 1971-2000." American Journal of Political Science 47 (2): 305-319. Rogin, Josh. 2008. "2008 Legislative Summary: War Supplemental." CQ Weekly, December 8, 3255. Rohde, David W. 1991. Parties and Leaders in the Postreform House. Chicago: University of Chicago Press. Rohde, David W. 2005. "Committees and Policy Formulation." In The Legislative Branch, eds. Paul J. Quirk and Sarah A. Binder. New York: Oxford University Press, 201-223. Rudder, Catherine E. 2009. "Transforming American Politics through Tax Policy." In Congress Reconsidered, 9th ed., eds. Lawrence B. Dodd and Bruce I. Oppenheimer. Washington, DC: CQ Press, 263-283. "Rush to Judgment." 2007. Washington Times, July 16, A16. Shaviro, Daniel N. 2007. Taxes, Spending, and the U.S. Government's March Toward Bankruptcy. New York: Cambridge University Press. Shepsle, Kenneth A., and Barry R. Weingast. 1981. “Political Preferences for the Pork Barrel: A Generalization.” American Journal of Political Science 25 (1): 96-111. Shipan, Charles, and William Lowry. 2001. "Environmental Policy and Party Divergence in Congress." Political Research Quarterly 54 (2): 245–264. Sinclair, Barbara. 1981. "The Speaker's Task Force in the Post-Reform House of Representatives." American Political Science Review 75 (2): 397-410. Sinclair, Barbara. 1983. Majority Leadership in the U.S. House. Baltimore: Johns Hopkins University Press. Sinclair, Barbara. 1994. "House Special Rules and the Institutional Design Controversy." Legislative Studies Quarterly 19 (4): 477-494. Sinclair, Barbara. 1995. Legislators, Leaders, and Lawmaking: The U.S. House of Representatives in the Postreform Era. Baltimore: Johns Hopkins University Press. Sinclair, Barbara. 1997. Unorthodox Lawmaking: New Legislative Processes in the U.S. Congress. Washington, D.C.: CQ Press. Sinclair, Barbara. 1999. "Transformational Leader or Faithful Agent? Principal-Agent Theory and House Majority Leadership." Legislative Studies Quarterly 24 (3): 421-449. Sinclair, Barbara. 2002. "Do Parties Matter?" In Party, Process, and Political Change in Congress: New Perspectives on the History of Congress, eds. David W. Brady and Matthew D. McCubbins. Stanford, CA: Stanford University Press, 36-63. Sinclair, Barbara. 2005. "Parties and Leadership in the House." In The Legislative Branch, eds. Paul J. Quirk and Sarah A. Binder. New York: Oxford University Press, 224-254. Sinclair, Barbara. 2006. Party Wars: Polarization and the Politics of National Policy Making. Norman, OK: University of Oklahoma Press. Sinclair, Barbara. 2008a. "Orchestrators of Unorthodox Lawmaking: Pelosi and McConnell in the 110th Congress." The Forum 6 (3). Sinclair, Barbara. 2008b. "Spoiling the Sausages? How a Polarized Congress Deliberates and Legislates." In Red and Blue Nation? Consequences and Correction of America's Polarized Politics, Vol. 2, eds. Pietro S. Nivola and David Brady. Washington, DC: Brookings Institution, 55-87.  124  Smith, Steven S. 1985. "Decisionmaking in Congress." In The New Direction in American Politics, eds. John E. Chubb and Paul E. Peterson. Washington, DC: The Brookings Institution, 203-233. Smith, Steven S. 2007. Party Influence in Congress. New York: Cambridge University Press. Smith, Steven S., and Gerald Gramm. 2009. "The Dynamics of Party Government in Congress." In Congress Reconsidered, 9th ed., eds. Lawrence C. Dodd and Bruce I. Oppenheimer. Washington, DC: CQ Press, 141-164. Sovey, Allison J., and Donald P. Green. 2011. "Instrumental Variables Estimation in Political Science: A Reader's Guide." American Journal of Political Science 55 (1): 188-200. Steiner, Jürg, André Bächtiger, Markus Spörndli, and Marco R. Steenbergen. 2004. Deliberative Politics in Action: Analyzing Parliamentary Discourse. New York: Cambridge University Press. Stewart, Chares, III. 2012. "Congressional Committees in a Partisan Era: The End of Institutionalization as We Know It?" In New Directions in Congressional Politics, ed. Jamie L. Carson. New York: Routledge. Stewart, Chares, III, and Jonathan Woon. 2009. Congressional Committee Assignments, 103rd to 111th Congresses, 1993−2009: House, November 13. Strahan, Randall W. 2007. Leading Representatives: The Agency of Leaders in the U.S. House. Baltimore: Johns Hopkins University Press. Sunstein, Cass R. 2000. "Deliberative Trouble? Why Groups Go to Extremes." The Yale Law Journal 110: 71-119. Sunstein, Cass R. 2005. "Group Judgments: Statistical Means, Deliberation, and Information Markets." New York University Law Review 80: 962-1049. Sunstein, Cass R. 2006. "Deliberating Groups versus Prediction Markets (or Hayek’s Challenge to Habermas)." Episteme: A Journal of Social Epistemology 3 (3): 192-213. Sunstein, Cass R. 2007. Republic.com 2.0. Princeton, NJ: Princeton University Press. Swindell, Bill. 2003a. "Bill Expands Role of States in Head Start." CQ Weekly, June 14, 1463. Swindell, Bill. 2003b. "GOP Rewrite of Head Start Program Guaranteed Quick, Stiff Resistance." CQ Weekly, May 31, 1334. Theriault, Sean M. 2008. Party Polarization in Congress. New York: Cambridge University Press. Tomz, Michael, Jason Wittenberg, and Gary King. 2003. "CLARIFY: Software for Interpreting and Presenting Statistical Results." Journal of Statistical Software 8 (1): 1-30. Tong, Lorraine H. 2009. "House Committee Party Ratios: 98th-111th Congresses." Congressional Research Service, March 30. Towell, Pat. 1994. "GOP Drive for a More Open House Reflects Pragmatism and Resentment." CQ Weekly, May 19, 3320. Trubowitz, Peter, and Nicole Mellow. 2005. "'Going Bipartisan': Politics by Other Means." Political Science Quarterly 120 (3): 433-453. Weingast, Barry R. 1979. "A Rational Choice Perspective on Congressional Norms." American Journal of Political Science 23 (2): 245-262. Woon, Jonathan. 2008. "Bill Cosponsorship in Congress: The Moderating Effect of the Agenda Positions on Legislative Proposals." Journal of Politics 70 (1): 201-216.  125  Appendices  Appendix A: Estimates of Committee Factors and Bill Types on Hearings and Markups  Table A1. Logit Results for the Effects of Committee Factors on Hearings Being Held Independent Variables  Estimated Coefficient  Chair Extremity  -0.993 (1.294)  Ranking Member Extremity  -1.435 (1.326)  Majority Committee Margin  0.084 (0.087)  Number of Referrals  0.022 (0.103)  Constant  1.701 (1.219)  Log-likelihood Pseudo R2 N (no. of bills)  -76.099 0.040 156  Note: Standard errors are in parentheses. *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  126  Table A2. Logit Results for the Effects of Committee Factors on Markups Being Held Independent Variables  Estimated Coefficient  Chair Extremity  -1.400 (1.427)  Ranking Member Extremity  -2.047 (1.627)  Majority Committee Margin  0.137 (0.113)  Number of Referrals  -0.050 (0.109)  Constant  2.488* (1.439)  Log-likelihood Pseudo R2 N (no. of bills)  -56.705 0.080 156  Note: Standard errors are in parentheses. *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  127  Table A3. Logit Results for the Effects of Bill Type on the Holding of Hearings Independent Variables  Estimated Coefficient  Change in Probability  -0.792** (0.368)  -15.16%  Majority-Conflict Bills  -0.257 (0.424)  —  Discretionary Bills  -0.655 (0.480)  —  Presidential Proposals  0.669 (0.475)  —  Party-Brand Bills  Constant  Log-likelihood Pseudo R2 N (no. of bills)  2.069*** (0.501) -74.278 0.0617 156  Note: Column 1 shows logit estimates with standard errors in parentheses. Column 2 shows the change in probability that hearings are held, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their modes. Estimates generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  128  Table A4. Negative Binomial Results for the Effects of Bill Type on the Number of Hearings Independent Variables  Estimated Coefficient  Expected Value  Party-Brand Bills  -0.642*** (0.182)  -2.404  Majority-Conflict Bills  1.15*** (0.229)  +11.269  Discretionary Bills  -0.167 (0.229)  −  Presidential Proposals  0.440** (0.212)  +2.855  Constant  1.774*** (0.238)  Log-pseudolikelihood  -471.868  Alpha  N (no. of bills)  1.599 (0.192) 156  Note: Column 1 shows negative binomial estimates with standard errors in parentheses. Column 2 shows the expected value in the number of hearings, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their modes. Expected values generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  129  Table A5. Logit Results for the Effects of Bill Type on the Holding of Markups Independent Variables  Estimated Coefficient  Change in Probability  Party-Brand Bills  -0.144 (0.436)  —  Majority-Conflict Bills  -0.927* (0.490)  -14.96%  Discretionary Bills  -1.823** (0.776)  -10.92%  0.282 (0.537)  —  Presidential Proposals  Constant  Log-likelihood Pseudo R2 N (no. of bills)  3.672*** (0.815) -56.104 0.0897 156  Note: Column 1 shows logit estimates with standard errors in parentheses. Column 2 shows the change in probability that markups are held, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their modes. Estimates generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  130  Table A6. Negative Binomial Results for the Effects of Bill Type on the Number of Markups Independent Variables  Estimated Coefficient  Expected Value  Party-Brand Bills  -0.531*** (0.173)  -1.027  Majority-Conflict Bills  0.287 (0.214)  −  Discretionary Bills  0.061 (0.220)  −  Presidential Proposals  -0.229 (0.211)  −  Constant  0.838*** (0.224)  Log-pseudolikelihood  -300.449  Alpha  N (no. of bills)  0.724 (0.179) 156  Note: Column 1 shows negative binomial estimates with standard errors in parentheses. Column 2 shows the expected value in the number of markups, when the independent variable shifts from its minimum value to its maximum value, with all other variables held constant at their modes. Expected values generated with Clarify 2.1 in Stata 11 (Tomz, Wittenberg, and King 2003). *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests).  131  Appendix B: Supplementary Data and Analyses on Bill Quality  Table B1. Editorial Endorsements of Presidential Candidates by Party, 1992-2008 Papers Arkansas Democrat-Gazette Atlanta Journal and Constitution Augusta Chronicle (GA) Austin American-Statesman Birmingham News Boston Herald Buffalo News Capital Times (WI) Chapel Hill Herald (NC) Charleston Daily Mail (WV) Charleston Gazette (WV) Chicago Sun-Times Christian Science Monitor Dallas Morning News Fresno Bee Grand Rapids Press (MI) Herald-Sun (NC) Houston Chronicle New York Times The News and Observer (NC) The Oklahoman Omaha World-Herald The Oregonian The Philadelphia Daily News The Philadelphia Inquirer Pittsburgh Post-Gazette The Plain Dealer (OH) Portland Post Herald (ME) The Post and Courier (SC) Providence-Journal (RI) Richmond Times Dispatch Roanoke Times (VI) Sacramento Bee  Democratic  Republican  Nonpartisan  x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x  132  Table B1 (cont’d). Editorial Endorsements of Presidential Candidates by Party, 1992-2008 Papers Salt Lake Tribune San Antonio Express-News San Francisco Chronicle San Jose Mercury News St. Louis Post-Dispatch St. Paul Pioneer Press (MN) St. Petersburg Times (FL) The Star-Ledger (NJ) Star-News (NC) Times-Picayune (LA) Tulsa World (OK) Washington Post Washington Times Total  Democratic  Republican  Nonpartisan x  x x x x x x x x x x x x 19  15  11  133  Table B2. Frequency Distribution of All-Sources Editorial Criticism No. of Editorials  Freq.  Percent  Cum.  0 1 2 3 4 5 6 7 8 9 10 11 12 16 17 19 20 21 23 24 25 26 29 30 34 36 40 53 67 69  34 12 10 5 10 2 7 4 6 4 7 2 4 3 2 2 1 1 1 2 1 1 1 3 1 2 1 1 1 1  25.76 9.09 7.58 3.79 7.58 1.52 5.30 3.03 4.55 3.03 5.30 1.52 3.03 2.27 1.52 1.52 0.76 0.76 0.76 1.52 0.76 0.76 0.76 2.27 0.76 1.52 0.76 0.76 0.76 0.76  25.76 34.85 42.42 46.21 53.79 55.30 60.61 63.64 68.18 71.21 76.52 78.03 81.06 83.33 84.85 86.36 87.12 87.88 88.64 90.15 90.91 91.67 92.42 94.70 95.45 96.97 97.73 98.48 99.24 100.00  Total  132  100.00  134  Table B3. Frequency Distribution of Nonpartisan Editorial Criticism No. of Editorials  Freq.  Percent  Cum.  0 1 2 3 4 5 6 8 9 10  85 21 9 5 2 1 5 1 2 1  64.39 15.91 6.82 3.79 1.52 0.76 3.79 0.76 1.52 0.76  64.39 80.30 87.12 90.91 92.42 93.18 96.97 97.73 99.24 100.00  Total  132  100.00  135  Table B4. OLS Results for the Effects of Legislative Deliberations on Bill Extremity Independent Variables  Model 1  Model 2  Hearings Held  -16.641*** (6.276)  —  Markups Held  —  4.506 (8.147)  -0.3887 (5.098)  -2.061 (5.482)  58.412*** (5.955)  42.048*** (7.221)  0.044 155  0.002 155  Floor Rule  Constant  R2 N (no. of bills)  Note: Robust standard errors in parentheses. *p<0.10, **p<0.05, *** p<0.01 (two-tailed tests)  136  Appendix C: Descriptive Statistics  Table C1. Descriptive Statistics for Dependent and Independent Variables Variables Hearings Held Number of Hearings Witness Balance Markups Held Number of Markups Open Floor Rule Length of Floor Debate Duration of Deliberations Cosponsor Balance Partisan Polarization Electoral Risk Unified Government Republican Control Time Remaining News Coverage Tax Bills Social-Welfare Bills Moral-Values Bills Majority-Conflict Bills Discretionary Bills Presidential Proposals Chair Extremity Ranking Member Extremity Majority Committee Margin Number of Referrals All-Sources Editorial Criticism Nonpartisan Editorial Criticism Partisan Editorial Divergence Bill Extremity  Mean  Std. Dev.  Min  Max  9.500 0.794 40.402 0.865 2.153 0.455 1.732 76.826 19.134 0.781 -1.153 0.237 0.461 384.538 28.903 0.141 0.083 0.141 0.365 0.673 0.326 0.445 0.370 7.827 2.064 8.545 1.007 25.715 45.028  18.905 0.405 28.573 0.342 3.274 0.499 1.136 86.479 29.177 0.127 10.542 0.426 0.500 183.335 42.724 0.349 0.277 0.349 0.483 0.470 0.470 0.153 0.180 3.005 1.969 12.411 2.0133 54.428 32.099  0 0 0 0 0 0 0.5 1 0 0.603 -14 0 0 21 0 0 0 0 0 0 0 0.071 0.011 2 1 0 0 -100 0  154 1 100 1 28 1 7.5 489 100 0.963 14 1 1 722 282 1 1 1 1 1 1 0.973 0.77 15 13 69 10 100 95.529  137  Table C2. Percentage of Bills that Undergo Truncated Deliberations at Each Legislative Stage for Each Congress Congress  98th 99th 100th 101st 102nd 103rd 104th 105th 106th 107th 108th 109th 110th  Omitted Hearings  Omitted Markups  Restricted Debate  16.67% 8.33% 0.00% 0.00% 16.67% 16.67% 33.33% 41.67% 25.00% 25.00% 33.33% 25.00% 25.00%  0.00% 0.00% 16.67% 0.00% 8.33% 0.00% 8.33% 33.33% 25.00% 16.67% 25.00% 16.67% 25.00%  25.00% 8.33% 41.67% 25.00% 50.00% 58.33% 50.00% 50.00% 58.33% 58.33% 75.00% 66.67% 83.33%  138  Table C3. Data for Partisan Polarization and Electoral Risk Variables by Congress Congress  98th 99th 100th 101st 102nd 103rd 104th 105th 106th 107th 108th 109th 110th  Partisan Polarization  Electoral Risk  0.603 0.632 0.641 0.650 0.673 0.726 0.805 0.837 0.856 0.892 0.922 0.955 0.963  -9 -13 -7 -14 -12 -4 12 14 12 6 10 3 -13  139  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0073053/manifest

Comment

Related Items