Parties and Procedural Choice in the House Rules Committee

Abstract

Party-based theories argue that leaders of the majority party use restrictive rules to increase that majority's influence over policy. However, informational theorists suggest that restrictive rules reduce information costs and encourage committee specialization. Using data on every bill on which the House Rules Committee placed a rule from the 94^th through 109^th Congresses, I test these explanations. Considering the use of restrictive rules, the evidence verifies some of the predictions of partisan theories, but unveils mixed support for informational theories. In addition, we learn that Democratic and Republican majorities differ with respect to the use of restrictive rules. The results indicate that Republican majorities behaved as more of a procedural cartel than their Democratic predecessors.

Acknowledgments

This article is based on my dissertation, Partisan Differences and Managing the Legislative Process in the Post-Reform House. I presented earlier versions of this article at the 2004 annual meetings of the Southern Political Science Association, New Orleans, Louisiana, and the Midwest Political Science Association, Chicago, Illinois. I thank Andrew Bargen, Fred Boehmke, Keith Poole, Megan Shannon, Steve Smith, and Matt Whittaker for their assistance in directing me toward appropriate data sources; Neal Beck and Fred Boehmke for their advice on estimating my model; and Fred Boehmke, Andrew Civettini, Scott Cody, Brian Harward, Michael Lewis-Beck, Eric Manning, Susan Miller, John Owens, Glenn Parker, Tom Rice, Scot Schraufnagel, Tracy Slagter, John Wilkerson, Don Wolfensberger, and especially Doug Dion and Chuck Shipan for all of their extremely helpful comments and suggestions on earlier versions. Finally, I thank Meredith Burkhart and Izak Post for research assistance. Any remaining errors and omissions are my own responsibility.

Notes

1. See Mayhew (2005) for a more complete discussion of what constitutes a major law.

2. During the 108th and 109th Congress, the House Rules Committee placed restrictive rules on approximately 80% of all bills that became major laws.

3. I define a restrictive rule as a modified closed or closed rule. Alternatively, I classify open rules as ones which do not significantly restrict a House member's ability to amend a bill. See Marshall (2002, Appendix A) for a more complete discussion of what constitutes a restrictive rule.

4. For example, the Republican Party had 15 deputy whips and 44 assistant whips, during the 106th Congress (Sinclair 2000), headed by majority leader Tom DeLay (R-TX), a strong Republican leader who was widely acknowledged as having as much power as the Speaker in his heyday. Conversely, earlier Democratic majorities had a higher number of whips than the Republicans, but were headed by much less powerful whips that were subordinate to the Speaker.

5. If this command structure operates in the way that is argued here, then Republican majorities might not need to employ restrictive rules on bills. After all, such a party has a structure through which it can alternatively reward compliant members and punish dissidents and members of the minority party. However, the control that leaders of any majority party have over their members is imperfect, even in a parliamentary system in which party discipline is paramount (Cox 1987; Searing 1994). Also, Republican majorities contended with an opposition that had increasingly divergent policy preferences. Without restrictive rules on bills, minority Democrats could have proposed amendments that modified the contents of bills, and/or debated bills endlessly to prevent them from passing in the form that they were originally proposed on the floor. Futhermore, minority Democrats might have coalesced with dissident Republicans to defeat a policy proposal that Republican leaders, but not all majority party members would prefer. For these reasons, Republican majorities needed restrictive rules to prevent such incidents from happening, and affecting their ability to pass their preferred policies.

6. Some argue that such a test may stack the deck against other models of legislative organization. However, my article corrects for overemphasizing such models in other studies of Rules Committee behavior.

7. A killer amendment is one that, if adopted, will likely cause a bill to fail (Enelow and Koehler, 1980; Wilkerson, 1999).

8. Proponents of CPG (e.g., Binder 1996, 1997; Dion 1997; Rohde 1991) argue that the use of rules that limit the rights of numeric minorities depend on the size of the majority, relative to the previous Congress. To test this prediction, I follow Binder (1996), Owens and Wrighton (2008) and Schickler (2000) and use a three-step process to generate majority partisan capacity. First, majority partisan strength is the percentage of seats held by that party divided by the standard deviation of that party's DW-NOMINATE scores. Second, minority partisan strength is the percentage of seats held by that party divided by the standard deviation of that party's DW-NOMINATE scores. Finally, I subtract majority from minority partisan strength to compute majority partisan capacity. When I replace the party cohesiveness measures with partisan capacity, the constitutive term that corresponds to this variable is positive and statistically significant. However, the interaction with Republican control is not significant. In addition, my results do not change from those reported in this article, with two exceptions: (1) the constitutive term for legislative profile is now positive and statistically significant; and (2) committee distance loses statistical significance.

9. Cox and McCubbins (2005) consider majority and minority roll rates in the context of setting the agenda in a broad way that goes beyond the use of restrictive rules. That said, their work has implications for the use of restrictive rules that I explore in this analysis.

10. While cohesiveness and ideological distance tend to be positively co-related to one another, many partisan theorists consider these concepts as distinct from one another (see e.g., Aldrich and Rohde 2000b; Roberts 2010; Rohde 1991; Smith 2007). Consistent with many partisan theorists, I refer to cohesiveness (or intraparty homogeneity), as the degree of similarity that exists within a party, and ideological distance as the distance between that party's median and some other member, depending on the prediction being tested. When viewed in this way, the contradiction that appears between the cartel model's predictions with respect to the decreased likelihood of restrictive rules on bills when the: (1) majority party is cohesive; and (2) ideological distance between the majority party's median member and the median member of a substantive committee evaporates as a consequence of this theoretical distinction.

11. Hereafter, I refer to such a situation as one in which the median members of a substantive committee and the Rules Committee are located on the same side as the floor median.

12. However, Krehbiel (1997) notes that minor specification changes make a huge difference in his results.

13. Schickler's (2000) theory was designed to explain the conditions under which the minority party is advantaged or undermined through the use of rules. Since restrictive rules are one method by which the majority party can undermine the minority party, his insights can prove useful to explain the conditions under which the Rules Committee places restrictive rules on bills.

14. I replaced this measure with one in which I used the committee whose median member has preferences that are the furthest from zero (using first dimension, DW-NOMINATE scores). When I do so, my results do not differ from those that are presented this paper, except that committee distance is only significant at the relatively loose .10-level of significance, using a one-tailed test.

15. Since this variable is a percentage, its values range from zero to 100. A party cohesiveness score of zero indicates a totally incohesive party and a cohesiveness score of 100 indicates a completely cohesive one. I acquired data for both measures from Joseph Cooper and Gerry Young (Cooper and Young 1997, 2002).

16. I should note that MEDIAN_FLOOR represents the floor median in each House, MEDIAN_MINPTY represents the median member of the minority party, MEDIAN_MAJPTY represents the median member of the majority party and MEDIAN_ADV represents the absolute value of the differences between the floor and the: (1) minority party, and (2) majority party in the current Congress. Finally, ΔMEDIAN_ADV denotes the difference between the MEDIAN_ADV of the current Congress from that of the previous Congress.

17. I substituted the number of bills on which the Rules Committee has placed a rule as a substitute for my workload measure. This variable was not statistically significant, and did not affect the significance patterns of the other variables. This substitution allowed me to test whether workload was not statistically significant because of how I measured it.

18. To compute the change in predicted probability of a bill receiving a restrictive rule, I held the values of all continuous variables at their mean, and all binary variables at zero. To compute this change, I changed the value of that variable from the mean to one standard deviation above it if it was a continuous variable, and from 0 to 1 for each binary variable, while holding all others at their baseline values. For each interactive term, I computed the change in predicted probability by changing the value of each term and its constitutive term from the mean to one standard deviation above it (or, for legislative profile, from 0 to 1).

19. I also operationalized majority and minority party cohesiveness as the standard deviation of the first dimension, DW-NOMINATE scores for each party in the House in each Congress. I used this operationalization as an attempt to lessen my dependence on vote-based measures. I found that majority and minority party cohesiveness are not statistically significant. While the results are mostly unchanged, committee distance is only significant at the relatively loose, .10-level of significance, using a one-tailed test.

20. To check whether the dummy for constituency service committees sapped the power of the legislative profile hypothesis, I excluded this dummy and found that the constitutive term for this hypothesis was not statistically significant. The results remain the same, except that the dummies for prestige and policy committees become positive and statistically significant.

21. I checked my results to verify that bills from any particular Congress did not drive the results that I obtained. To conduct my robustness checks, I ran my models, and excluded bills from one Congress at a time while including all other Congresses. I discover that the signs on my coefficients and the significance patterns mostly do not change. These models are available from the author on request.