Members and Leaders in Senate
Obstruction
Nicholas O. Howard
Ph.D. Candidate
University of North Carolina
[email protected]
August 26, 2014
Abstract
Obstruction is one of the most widely discussed, but least understood, aspects of
the Senate. As most obstruction occurs in off-floor negotiation than through public filibusters, my paper focuses on what is privately obstructed by senators. Utilizing a newly
gathered dataset of Republican holds under the leadership of Howard Baker (R-TN) and
Robert Dole (R-KS), I explore the usage of Senate holds and their effects. I establish
theory driven hypotheses for when private obstruction occurs, and subsequently test if
these holds change outcomes. I find that institutional circumstances, bill sponsor identities, committee consideration, and chamber support all affect which bills are targeted
by holds. I then utilize data on bill passage and on-floor amendments to explore if these
holds change outcomes, and find that holds do lower the probability of bill passage, but
that leaders routinely do not accommodate amendment requests. This paper sheds light
on the world of lawmaking through illustrating the targets and effects of obstruction in
its most widely-used forms.
1
Introduction
Students of legislatures are, at some level, primarily concerned with how bills are
made into laws. This often centers around questions of institutional design and legislator
interests.1 Elected officials can design their institutions in order to further their interests
through controlling how laws are made, providing themselves with benefits from passing
laws only available to incumbents. Thus elected officials have an interest in passing laws
for the purpose of receiving benefits. The problem with this assertion is that individual
elected officials need not, and often do not, agree on what should be done. No laws are
made if elected officials cannot agree to pass legislation.
This tension between individual elected officials and legislative efficiency appears
most clearly in Senate procedure. In the Senate, leaders tenuously balance member
needs against the ability to pass legislation in an expeditious manner. Individual senators have a great deal of power in deciding how the chamber as a whole functions.
The lack of a readily available centralizing figure and ability for extended debate limit
chamber efficiency.2 Given these institutional rules, threats to slow, if not outright stop,
proceedings must be taken seriously. The question is how do party leaders in the Senate
structure and control the legislative agenda when individual senators are given numerous opportunities to halt the legislative process. Individual senators have the capacity
1
Reelection naturally plays a pivotal role in why some bills become laws while others do not. All
other goals, such as policy or power, require this goal in order to be realized. For a discussion of how
reelection goals alter institutional design, see Mayhew (1974).
2
I return to these rules below in order to explain Senate debate and how holds fit into the process.
1
and procedural tools, such as holds and filibusters, to stop leaders’ efforts to control
the agenda. Both an individual senator’s capacity to obstruct and the capability of
leaders to successfully structure and limit these senators’ obstructive capabilities vary
across types of legislation. The two following examples will illustrate that the results of
this conflict between individual obstructionist capacity and leader control shapes both
deliberation and outcomes.
1.1
Obstruction and Management: Two Examples
Pursuing health care insurance reform for private employers and extending cover-
age for the unemployed became a central focus of Bill Clinton’s administration after his
promises in the 1992 presidential campaign.3 Clinton’s strategy included the secret, and
controversial, Health Care Task force and working with congressional Democrats to craft
a nearly universal coverage health care policy.4 These efforts failed in Congress, a victory
for congressional Republicans which they utilized to raise questions about Democrats’
effectiveness in the 1994 midterm elections.
After their victory in the 1994 midterms, the new Republican majority felt pressure to act on healthcare.5 Republicans introduced the Health Insurance Portability and
Accountability Act of 1996 (hereafter HIPAA) on March 18, 1996, in the House. Designed to limit exclusions based on preexisting conditions, HIPAA was the Republican
3
Ifill, Gwen. “The 1992 Campaign: The Democrats; Clinton Proposes Making Employers Cover
Health Care.” The New York Times. September 25, 1992.
4
Pear, Robert. “Ending Its Secrecy, White House Lists Heath-Care Panel.” The New York Times.
March 27, 1993.
5
Haveman, Judith. “House GOP Releases Plan on Health Care.” Washington Post. March 9, 1996.
2
counter to Clinton’s pursuit of health care reform and shaped related rhetoric in the
1996 presidential campaign. Although somewhat controversial, HIPAA moved rapidly
through the House.6
7
The Senate fight over HIPAA was much different. Introduced on April 15, 1996,
the more individualistic Senate sought changes to the bill.8 A bipartisan filibuster by
Edward Kennedy (D-MA) and Nancy Kassebaum (R-KS) sought to either fundamentally
change the bill’s content or stop passage of the House’s legislation.9 These filibusters
presented the public with the possibility that this bill would be talked to death like
others before it. Despite the obstructionist efforts of senators Kennedy and Kassebaum,
party leaders were able to move HIPPA through the Senate with a Unanimous Consent
Agreement (UCA), negotiated before floor debate began, which established the date of
final passage.10 Due to the UCA, obstruction was ended, the Senate’s more moderate
language inserted into the bill, and final consideration was a relatively smooth process.11
6
Haveman, Judith. “House GOP Releases Plan on Health Care.” Washington Post. March 9, 1996.
7
HIPAA was introduced in the House on March 18, 1996, reported from Ways and Means on March
25, and passed on March 28. HIPAA was also referred to the House Commerce, Judiciary, and Economic
and Education Committees. These committees were given additional time to consider the bill, but the
final passage vote was held before the extended reporting deadline with no reports given. HIPAA passed
on a recorded vote, with Republicans voting 299-1 in favor and Democrats 149-38 against.
8
Note that this bill was the target of multiple procedural tabling and cloture motions.
9
Dewar, Helen. “Striving to Lead, Dole Stumbles Into a Week of Setbacks.” Washington Post. April
21, 1995.
10
A Unanimous Consent Agreement (UCA) is an agreement between party leaders, and agreed to by
all 100 senators, which covers any or all of floor activity, passage procedure, amendment limitations,
and setting the time of final consideration prior to floor consideration.
11
The bill was ultimately amended by substituting the language of S. 1028, the Health Insurance
Portability and Accountability Act of 1995, after the enacting clause. H.R. 3103 passed through the
Senate with no referral, but S. 1028 was referred to, and reported by, the Committee on Labor and
Human Resources.
3
HIPAA was brought up and passed unanimously on April 23, the date specified in the
UCA, and ultimately signed into law on August 21, 1996 by President Clinton. The
actions by party leaders to negotiate the UCA before floor debate began allowed members
to present filibusters but not have them undermine passage, as it would have without
this action by leaders.
The Hatch Act Reform Amendments of 1988, H.R. 3400 in the 100th Congress,
presents a different story of obstruction and leadership.12 The bill, a recurring item
largely supported by congressional Democrats and opposed by President Reagan, sought
to allow civil servants to interact with campaigns for elected office.13 The bill progressed
quickly through the House, passing in a contentious vote on November 17, 1987.14
Senate proceedings advanced much less smoothly for the Hatch Act Reform Amendments of 1988. After introduction the bill remained in the Committee on Governmental
Affairs until finally reported to the floor and placed on the calendar on July 7, 1988.15
The committee deliberations were contentious, but the lack of party-line divisions within
committee would indicate that some possible amending activity should resolve conflicts
before the bill was disposed of on the floor.16 However, the bill remained on the calendar
12
Causey, Mike. “Hatch Act Hearings.” Washington Post. May 14, 1987.
13
The Hatch Act Reform Amendments was introduced by William Clay (D-MO) on November 1, 1987.
For further detail, see both Pianin, Eric and Judith Havemann. “Bill to Ease Hatch Act Advances.”
Washington Post. October 7, 1987; and Havemann, Judith. “Administration Assails Bill To Ease Hatch
Act Restrictions.” Washington Post. June 4, 1987.
14
The final vote was Democrats voting 242-3 in favor and Republicans 109-63 against.
15
This reporting came only after committee substitute amendments and an order to report issued by
the full chamber.
16
Phillips, Don. “Senate Panel Approves Mild Hatch Act Revision.” Washington Post. May 20,
4
without ever reaching the floor or even being subject to a motion to proceed.17 Thus, the
Hatch Act Reform Amendments of 1988 died a quiet death with no visible explanations
after being reported from committee (Krutz 2005).
The Hatch Act Reform Amendments of 1988’s failure is not due to filibusters
or poison pill amendments. This is apparent from the bill’s lack of floor consideration.
Rather, it is due to the lack of floor exposure for the bill. That is, there was no previously
negotiated UCA to help structure floor consideration. Majority leader George Mitchell
(D-ME) and minority leader Dole could not procure a UCA and therefore elected to not
proceed on the floor. To negotiate a UCA, Mitchell and Dole required the consent of all
100 members, and such consent was not attainable for this bill. Senators Helms (R-NC)
and Roth (R-DE) sent Republican leader Dole holds stating their intent to object to
the bill should a motion to proceed come to the floor. Thus, these holds undermined
the possibility of unanimous consent, and the leaders decided that the possibility of
passing the bill with this objection was not worth the cost of extra time on the floor
should Helms and Roth actually follow through with their threats. These holds removed
the possibility of a UCA, and ultimately undermined any efforts to move the legislation
forward.
These two examples above make apparent a hidden truth in traditional depictions
of Senate proceedings, which tend to focus on the filibuster as the tool of choice for
1988.
17
Causey, Mike. “The Hatch Act Hangs On.” Washington Post. September 28,1988.
5
obstruction-minded senators. In order to understand why some bills do become laws
while others do not, examining the entire process of obstruction – and not just the
filibuster – is essential. While the presence or absence of filibusters provides great insight into how changes within institutions and incentives shape behavior and outcomes,
characterizing filibusters as deciding the fate of legislation is incomplete. Fully conceptualizing what bills become law requires understanding member actions at earlier stages in
bill consideration. This project, like those focusing on the filibuster, takes seriously that
members have both an incentive to pass legislation and insist on their preferred policy
being considered. With both of these premises in mind, I begin with an explanation for
why a legislature would allow individual members to object to bills on a bill-by-bill basis
rather than insisting on centralizing authority and passing bills through the chamber.
I then develop a theoretical explanation for why certain bills may be obstructed and
articulate expectations for when objection will be observed. I then test these expectations with new data I collected on holds during the period of Republican leaders Howard
Baker (TN) and Robert Dole (KS). Finally, I explore how the legislature reacts to these
instances of obstruction and if senators receive payoffs from engaging in obstructionist
behavior.
2
Senate History: Why Holds?
Senate organization limits the ability to centralize chamber control and greatly
enhances individual senators’ power. The President of the Senate, a natural central-
6
izing figure, is constitutionally mandated to be the non-senator Vice President of the
United States (Smith 2014; Lynch and Madonna N.d.). Floor party leaders did not
exist until the 1920s, with no procedural recognition until the 1930s, limiting the ability
of organized parties to control debate (Koger 2006; Sinclair 1989).18 Limiting debate
proves extraordinarily difficult due to the removal of the Previous Question (PQ) motion in 1806 (Gamm and Smith 2002).19 This difficulty in ending debate eventually
led to the implementation of cloture to allow for a vote to end debate on a motion or
measure (Wawro and Schickler 2006, 2004). In short, managing debate is possible, but
requires a high organizational costs, a possible supermajority vote, and a time penalty
for implementation (Oppenheimer 1985; Sinclair 2002).
Implementing cloture and the majority leader’s right of first recognition results in
a system where unlimited debate is still possible, but can be overcome through either
membership action or leader recognition. Difficulties in actually implementing debate
limitations led Senate party leaders to attempt ex ante limitations on debate through
Unanimous Consent Agreements (UCAs) beginning in the 1940s and 1950s. UCAs are
provisions for consideration of a bill negotiated between the two party leaders and agreed
18
The right of first recognition is that in the event that multiple senators rise to be recognized by the
chair at the same moment, the chair will recognize a party leader if she is one of those seeking to be
recognized. A lack of centralized control can result in cycling between options, both as a splintering of
existing coalitions and outright moving between bills (Poole and Rosenthal 2007). The ability of leaders
to gain recognition at the beginning of debate also limits cycling between bills and amendments (Arrow
1951; Riker 1980; Schickler 2001; Rohde 1991; Koger 2010).
19
PQ motions are utilized in the post-18th Century House to end debates and bring bills to a vote.
This is not the only interpretation of the lack of a PQ motion, or even its original interpretation within
the Senate. For a detailed history of the Senate PQ motion, see Binder and Smith (1997) and Binder
(1997).
7
to by all 100 senators before floor debate ever begins. These limitations can cover any
part of consideration, including the total time for floor debate, amendments allowed on
the floor, and the mechanics of final passage consideration, among others (Smith 1989;
Smith and Flathman 1989). These agreements overcome the heavy time penalties of
implementing cloture, the difficulties in managing recognition procedures, and vastly
limit the impact of the Senate’s other procedural rules. Through UCAs leaders can
effectively create something similar to the House’s special rules and control debate and
passage before consideration ever begins (Cox and McCubbins 2005, 2007).
The downside to utilizing UCAs is that it does require truly unanimous consent,
and one member unwilling to abide by the agreement can undermine all efforts to limit
debate. This is why analyzing holds is essential to understanding Senate obstruction. A
hold is a threat to object to a specific piece of legislation sent by a senator – or group of
senators – to a party leader. Taken differently, holds are a threatened filibuster before
an observable filibuster is actually possible.20 Holds clearly prevent party leaders from
easily negotiating a UCA. Figure 1 presents an example of a hold from Jesse Helms
(R-NC) in the 104th Congress. As can be seen here, the hold is a statement that a
senator, Jesse Helms in this case, will obstruct proceedings in some fashion if the Senate
chooses to proceed to consideration of this bill.
The Senate agenda is far more vast than the time available even in a world where
obstruction is curtailed (Jones and Baumgartner 2005; Krutz 2005). Compounding
20
This distinction will be made in more detail later in this paper, when the connection between holds,
UCAs, and filibusters is elaborated.
8
Figure 1: Example Hold
9
this inability to consider all legislation are the extra time constraints given through
the filibuster, recognition procedures, and lack of a central agenda setter. Legislation
must pass the Senate both for the purposes of continuing governmental operations and
reelecting incumbent senators. Leaders thus allow holds as a mechanism through which
they can screen bills before floor time can be consumed.
3
Obstructing Bills
The question here is not “who holds?” but rather “what is held?” Senators prefer
to have policy made at their ideal points, regardless of institutional rules or passage
procedures.21 The difficulty for this is that each senator insisting on their ideal points
creates unstable political environments (Gilligan and Krehbiel 1987; Krehbiel 1991).
Previous research has established that more ideologically extreme members hold more
often, but this is insufficient to develop a broader understanding of the effects of holds on
legislation (Howard and Roberts N.d.). Additionally, we know that the implementation
of cloture altered senator’s considerations about utilizing filibusters and that further
changes in the institution created different incentives for senators (Wawro and Schickler
2006, 2004; Koger 2010; Binder and Smith 1997). What I undertake here is to look
at how obstruction is utilized on a bill-by-bill basis. Thus, the central question of this
21
This remains true no matter which motivation for policy one assumes in the discussion above.
Previous studies have questioned whether policy should be designed to appeal to a minimum winning
coalition of members, to all co-partisans, to all members of a chamber, or to some other subset of
members (Austen-Smith and Banks 1988; Baron and Ferejohn 1989). For the purposes of this discussion
I remain agnostic about how policies should be designed outside of conforming to some predetermined
threshold.
10
theory is what aspects of a piece of legislation make it more and less likely that it is
targeted for obstruction?
Stating that obstruction is in many ways the defining feature of the U.S. Senate
provokes relatively few arguments, but this statement masks the process of obstruction.
Both popular and scholarly attention to obstruction, particularly in the case of the
Senate, has focused on public displays of dilatory intent – filibusters in the congressional
world (Binder, Lawrence and Smith 2002; Binder 2003; Krehbiel 1998; Koger 2010).
Scholars of legislatures the world over understand that most legislative bargaining occurs
in places either difficult or impossible to observe among the public. Parties often organize
in closed sessions and much bargaining among members is unobservable without private
information (Aldrich 1995). Thus, understanding what bills are targeted by obstruction
requires developing an understanding of not only what bills are likely to have publicly
observable obstruction but what bills are likely to have any type of obstruction observable
to party leaders.
3.1
Obstruction and the Legislative Process
Establishing a theoretical understanding of what bills should be subject to holds
requires first establishing the process of obstruction and bill consideration. That is,
laying out the legislative process and the assumptions made about this process. I make
two assumptions about this process. First, senators desire to have policy made at their
ideal point and will introduce legislation which accomplishes this goal. Second, I assume
that committees carefully vet legislation within their defined jurisdiction while only
11
Figure 2: Process of Obstruction
Bill Introduced
Committee Referral
Report
Not
Hold
Pull Bill
Accommodate
Obstruct
Not
No Hold
Not Accommodate
Obstruct
Obstruct
Not
working within their jurisdictional area. An implication of the assumptions is that bills
reported out of committee have a chance of passing the Senate and becoming law.22
With these assumptions and their implication in mind, the legislative process can
be laid out as shown in Figure 2. Bills are introduced by their sponsor and referred to
a committee. The committee can then choose to report the bill or not, and if the choice
22
Another possible implication of this is that committees are unable to log roll across bills within
different committees on an individual level, and this process can only operate through the chamber
referring bills to multiple committees.
12
Not
is to not report the bill dies at that stage. After a bill is reported from committee it is
received by the two party leaders.23 At this point, leaders can attempt to organize their
parties and broker an agreement outside of any observable bargaining structure before
a bill is placed on the calendar.24 After a bill is reported from committee it is placed
on the official Senate calendar, and is expected to come up for debate on the floor at a
future point in time.
When the official calendar lists a bill senators view it as viable legislation and then
face a choice of whether to procedurally obstruct through holds. A senator, looking at
an individual bill, chooses whether to place a hold on that bill or not on the specific
piece of legislation.25 A senator can choose to place multiple types of holds, including
statements of objection with no recourse, specific and general amendment requests,
and simply requests for more time related to bill consideration. This distinction is
undoubtedly important, but I choose not to focus on it at this point for purposes of
clarity while returning to it in later sections.26 Regardless of a senator’s objectives in
placing a hold, if they choose to place a hold they have registered some type of objection
23
The bill will also be received by both the President of the Senate – the Vice President – and the
Senate pro tempore. While these leaders factor into decisions on the floor, they rarely play a role in the
off-floor bargaining and obstruction stages. Thus, I assume that the chamber leaders in this context
are equivalent to the two party leaders.
24
This would be akin to attempted distributive payoffs before the obstruction process can begin
(Arnold 1990; Lee and Oppenheimer 1999).
25
This operates on a bill-by-bill basis, with a senator targeting an individual bill. Hold letters, such
as in Figure 1, do not have to target only one bill. However, senators wishing to target multiple bills
within a single hold letter must state declare objection to each bill separately. Thus the integrity of
this process is preserved, eve if a senator can direct holds to more than one bill at a time.
26
Howard and Roberts (N.d.) provide extensive treatment of this data and how hold types are
distributed across congresses.
13
to the bill.
Leaders then choose a strategy with which to respond to senators’ holding behavior. If a hold is not observed, the bill is introduced on the floor.27 Senators can choose
to object on the floor at this point through filibusters and dilatory amendments, but
leaders expect no objection to come up. That is, obstructive behavior should be unlikely
given the lack of previous objection through holds.
When senators place holds on a bill, the choice leaders face becomes more complex.
Leaders have three general strategies, and the choice a leader makes centers around the
possibility of gaining a unanimous consent agreement and managing floor time. First, a
leader can choose to withdraw a bill targeted by a hold. Most of the time this will involve
pulling the bill from its spot on the calendar and proceeding with the next bill. These
are bills on which obstruction is not likely to be overcome or the cost of overcoming it
will be too high for the benefits from passing the bill. If leaders choose to withdraw a
bill at this stage the bill dies and consideration ends.
The other two options involve leaders moving forward with the held bill in some
form. Leadership’s second option is to move forward with the bill after offering some
amendments to holding members. This takes one of two observable forms. Leadership can either work with holding senators to guarantee that their amendment will be
27
This is undoubtedly an oversimplification, as bills not targeted by holds are occasionally not brought
up on the floor. The purpose of this simplification is to focus on leaders looking at individual senator
behavior, rather than the larger political climate. Interest groups, caucuses, and time pressures may
push items off the agenda, but the focus here is on the connection between senator behavior and leader
response.
14
considered on the floor, or they can modify a UCA to include the amendment a member requests. Guaranteeing consideration of amendments on the floor requires more
careful on floor management than altering UCAs, but it does not require leadership to
spend time and effort to craft a UCA responsive to an individual senator’s wants and
concerns.28
Finally, leaders can proceed with the bill in its current form and choose to ignore
the hold. Leaders choosing this strategy may believe either that there will be no actual
objection on the floor or that such objection can be overcome.29 Obtaining a UCA at
this point is possible, but the leader will more likely move forward with the bill without
a UCA guiding debate.
After a leader has chosen to proceed by either allowing amendments or ignoring
a hold, the bill is placed on the floor for consideration. At this point a member can
choose to obstruct on the floor through additional amendments or filibusters among
other activities.30 Obstruction should be more likely when leaders choose to ignore a
hold than when amendments are offered. This said, on-floor obstruction is possible with
either strategy leaders choose. After this process of obstruction is complete the bill will
pass or not given the results of the previous stage and its level of support. What should
28
The form this takes hinges upon the ability of the leader to craft the details of content within a
UCA relative to other members of the chamber.
29
This may be due to both the nature of the bill, as some bills such as budgets are must-pass, and
the nature of the holder, as holders with extreme ideologies are less likely to be successful in stopping
passage (Howard and Roberts N.d.; Krehbiel 1998).
30
This can include simply expanding the number and duration of actually related speeches and
introducing a greater number of amendments. In short, obstructionist activity is primarily considered
a pure delay tactic, but it need not be only this.
15
be obvious here is that each stage is fundamentally important in its own right, but does
exist as part of a larger whole within the process of consideration.
3.2
Holds in Incomplete Expectation
When viewed with actors possessing both perfect information and the ability to
act without costs or harm, the puzzle of obstruction disappears. However, the Senate
does not operate in this environment. If it did, two primary reasons we could expect
obstruction on legislation to appear and be prevented appear in Figure 2. First, a senator
naturally prefers legislation appealing to her, which may appear as policies being close to
rather than far away from her ideological position.31 On balance, this senator prefers that
bills be exactly at her ideal point so that she receives maximum benefit. Additionally,
she may desire to be payed off to not obstruct and behave so that she receives these
payoffs. This senator has the ability to object to any bill not at her ideal point or on
which she would like to be payed off. While difficult for one member, the problems
stemming from this ability for any member to object to any bill dramatically increase
when the number of senators increases toward the realistic Senate. Thus, if holds were
a costless activity we would observe them on nearly all pieces of viable legislation.
Second, leaders value managing the floor. While leaders inarguably have personal
policy preferences, furthering the agenda of the party is valued in addition to these
policy preferences. Knowing that policy must be passed for her party’s goals to be
31
For the sake of simplicity and brevity, I assume a single dimensional world for these thought exercise
examples. The same logic applies for multidimensional spaces as well.
16
realized, the leader values expediency in the legislative process. The leader will then
work out compromise bills by negotiating log rolls across bills and between members, or
will pull all contentious legislation. Thus, if leaders had perfect information holds would
be prevented through compromises across senators or preventing consideration.
As mentioned above, reality presents a situation distinct from either of these perfect information, no cost environments. Holds are indeed observed on legislation, but
not on all bills. According to the data presented below, approximately 15% of Senate
originated bills reported out of Senate committees receive holds. Thus, holds occur on a
regular basis, but are not utilized as a blanket strategy by members. This study pursues
the specifics of bills that are targeted, rather than focusing on why members may choose
to hold any randomly selected bill.32 Understanding what drives this variation in hold
usage provides a key insight into how senators view bills.
3.3
A More Complete Explanation
In order to explain what bills are targets of holds, we need to separate the bills
themselves from the membership. Some bills are clearly higher priorities than others,
such as the difference between yearly budgets and reforms clarifying legislative language.
Similarly, some bills are more widely supported than others, both within and external to
Congress. My explanation must account for these factors to discern between legislation
32
In what follows I pursue the question of why bills are held, rather than why members choose to
use holds. Thus, I build on – but differ in focus from – Krehbiel (1985), who demonstrates through
comparative statics that an individual member would choose to hold a given bill as part of a log roll
across multiple bills.
17
that is more and less likely to be targeted by obstruction.
I begin with a senators’s incentives to utilize obstruction. As previously stated,
senators desire to have policy made at their ideal point, and are more likely to favor
a bill as its content approaches this ideal point.33 Thus, a senator will be more likely
to obstruct a given piece of legislation as the ideal point of the policy moves further
away from the senator’s ideal point. If, as in the discussion above, we assume that
the sponsor introduces policy which maximizes his or her benefits from the policy, then
the distinction is between the ideal point of sponsor and the ideal point of the possibly
obstructing senator. As the distance between the sponsor’s ideal point and obstructing
senator’s ideal point decreases, a bill should be less likely to be held.
Hypothesis 1 : As the ideological distance between a bill sponsor and prospective
holder increases, a hold will be more likely on the sponsored bill.
More widely supported bills should be more likely to pass the chamber. These bills
are more likely to have greater support on both procedural motions and final passage
votes. This legislation should also be more likely to be the target of holds, however.
To see why this is true, assume that a bill following the process shown in Figure 2
has a percentage of senators potentially greater than a majority in favor of this bill.
This makes the bill likely to be approved by the chamber median, and the bill may be
able to reach the supermajority pivot’s threshold of approval (Krehbiel 1998). Opposing
33
This relationship is commonly thought of through either a linear loss function (−|xi − xp |) or a
quadratic loss function (−(xi − xp )2 ), where xi is the policy ideal point of the senator and xp is the
point of the policy. For the purposes of this examination, either formulation is functional.
18
senators know that if the bill comes to the floor they will have to obstruct through actions
such as filibusters in order to prevent passage – and may lose even with a filibuster. This
will cause these opposing senators to place a hold on this more widely supported bill
when it is placed on the Senate calendar in order to prevent it from reaching the floor
at all. These opposing senators hope that leaders seeing a hold may decide that the
bill is not worth managing potential floor objection, even if this objection is likely to be
overcome. Thus, while more popular legislation should be more likely to be put on the
floor, it should also be more likely to be the target of holds.
Hypothesis 2 : As visible support for a bill increases, holds will be more likely.
Wawro and Schickler (2006) demonstrate that obstructionist activity, observed
through filibusters, becomes more powerful as time remaining in the legislative session
decreases. Thus, at first glance it seems that bills introduced later in Congresses should
be more likely to be held. However, this misses is that obstructing bills has a costly
component related to time, even for private obstruction such as holds. To illustrate
this, take a senator who has ten policies she wishes to be enacted into law. Early in
a Congress, placing a hold on a bill does not radically diminish the probability that
her policy goals are made into law as time is still plentiful. As time to consider her
policies decreases, the probability that a preferred bill is considered begins to depend
more heavily on her own obstructionist actions. If the senator chooses to engage in
obstructionist activity, her own legislation has a higher probability of becoming part of
the obstructed backlog due to her own actions. Without changing any of the direct costs
19
of obstruction, such as public visibility and leader recalcitrance, the costs of engaging in
obstruction increase as time decreases.
Hypothesis 3 : Holds will be more likely on bills introduced early in a Congress.
Finally, bill content presents useful information to senators beyond sponsor details. Separate classes of legislation exist simultaneously in Congress. The most obvious
example is ceremonial legislation, such as post office renamings, versus non-ceremonial
legislation, such as the Affordable Care Act and HIPAA.34 By definition, legislation that
deals with more issues is more likely to be held, as there are more areas on which senators can alter and object. More senators have interests represented in these bills, and
thus the probability that one chooses to place a hold increases. Bills with broader effects
are more likely to be considered in a few specified and higher profile committees, such
as the Budget Committee.35
Hypothesis 4 : Bills referred to higher profile committees will be more likely to be
targeted by holds.
In short, senators are perceived to be knowledgeable about both their own interests
and how exercising their procedural prerogatives alters their own interests. The activity
of interest is whether they choose to place a hold or not, but what drives this is a
combination of bill-specific parameters and institutional circumstance. The specifics of
34
Important to note is that almost no ceremonial legislation reported from committee is subject to a
hold. Only one ceremonial bill reported from committee was subject to a hold in the 95th (1977-1978)
and 97th (1981-1982) Congresses combined.
35
A second possibility is that the most likely bills to pass the Senate are those which are highest
priority. These bills are also often referred to prestigious policy-related committees, implying that the
process of referral location and related hold behavior should appear the same.
20
the sponsor, a bill’s popularity, the time remaining in a session, and the general content
of the legislation should push members to perceive bills differently. They know that when
placing a hold on a bill leaders may react certain ways depending upon these things as
well, and adjust their strategy accordingly (Ainsworth and Flathman 1995). With this
theoretical basis in place, I now move to translating this theory into observable metrics.
4
Research Design
Senate floor obstruction is increasing over the last half century. The number of
filibusters and cloture petitions have risen. (Binder, Lawrence and Smith 2002; Koger
2010; Smith 2014). Additionally, the pursuit of poison-pill amendments has become
more pronounced, with senators seeking to undermine legislation through fracturing
coalitions and punishing implementation (Jenkins and Munger 2003). Parties have become procedural teams on the floor, increasingly obstructing and voting together (Lee
2009). All of this is to say that we know a great deal empirically about the last stages
of obstruction, but little about the stages prior to floor consideration.
Testing the theory presented above requires capturing and analyzing the process
of obstruction prior to floor introduction in a consistent fashion across legislation. I
utilize a dataset of holds I gathered from the Robert J. Dole (R-KS) Archive and a set
of holds gathered by Evans and Lipinski (2005) from the Howard Baker (R-TN) Archive
to explore this process (Evans and Lipinski 2005, N.d.; Howard and Roberts N.d.). This
dataset provides an accounting of holds placed by Republicans in the 95th, 97th, and
21
99th-104th Congresses.36 These data provide a summary of the target of the hold, the
day the hold was registered, and the general type of hold, providing for a general test of
hold dynamics.37
These data can be used to test the hypotheses above. The available data contains
only Republican holds, implying that Sponsor Ideology, measured as first dimension
DW-Nominate, should be negatively related to hold placement. Thus, as Sponsor
Ideology increases, the probability of a hold on a bill should decrease. Treating sponsors
as important actors in bill content and viability requires several controls for sponsors,
including Sponsor Seniority and Sponsor Committee Leader status. Translating visible
support is far more direct, as the number of Cosponsors provides a close corollary to
if a bill is likely to pass (Krehbiel 1998). I also consider time as a steadily decreasing
function, with Months Remaining serving as this proxy.38
The need to treat bill content through committee referral proves far trickier. Several options for this exist, but I begin with the committee classification system provided
by Deering and Smith (1997).39 This approach utilizes both the observed behavior of
36
The 95th Congress only has an accounting of holds through marked calendars generated by Baker’s
office, whereas the remainder of the congresses have the hold letters as seen in Figure 1. I utilize this
congress, but acknowledge the different data structure.
37
For the 97th and 99th-104th Congresses I have additional information based on the letters. Much
of this detail is presented in Howard and Roberts (N.d.), and some of this will be utilized later in this
paper.
38
It is possible, as Wawro and Schickler (2004) find, that the last month is driving this effect. Given
that I am primarily interested in the placement of holds, and that holds are observed at all points within
a congress, I choose to focus on time as a continuous function.
39
Two alternative options exist. The first option, simply controlling for referral to a specific committee
through a series of dummy variables, provides useful information but few general expectations. Second,
controlling for referral to any category of committee provided by Deering and Smith provides great
22
senators through their requests to serve on specific committees and the types of policies
committees produce in order to generate categories of committees.40 A Policy Committee is the highest profile type of committee in which a bill can be considered, thus I
utilize a dummy variable for all bills considered in this type of committee.41
Also important to consider for both content effects and possible popularity measures is the presence of Multiple Referral for bills. These bills should be more likely to
receive holds, as they have been considered by a wider range of senators before reporting to party leaders, and therefore opposing senators may believe these bills are more
likely to pass.42 Finally, I employ controls for institutional circumstances for both the
chamber, Republican Majority and whether a bill has a Majority Sponsor.
5
Results
In order to test the question of what is held and what affects this behavior, I
must rely upon observations about bills rather than members. Thus, in the analysis
to follow, the unit of observation will be an individual bill rather than an individual
richness but does not directly test my hypothesis. Each of these approaches has appeal, but the
approach chosen provides greater ability to develop general expectations with only marginal loss in
specific information.
40
The committee coding system provided by Deering and Smith (1997), which I utilize, is as follows. Policy committees are: Budget, Foreign Relations, Governmental Affairs, Judiciary, and Labor.
Mixed policy and constituency committees are: Armed Services, Banking, Finance, and Small Business. Constituency committees are: Agriculture, Appropriations, Commerce, Energy, and Environment
and Public Works. Unrequested committees are: Rules and Administration, Veterans’ Affairs, and
Intelligence.
41
Several studies have utilized a similar approach to generating categories of committees, including
Fenno (1973).
42
Additionally, this helps control for consideration in non-high profile committees as well.
23
senator. Furthermore, I restrict my analysis to only Senate originated bills which have
been reported from committee, as discussed above and presented in Figure 2.43 These
are the bills most likely to pass as they have been carefully vetted and approved by
committees.44 Holds speak to a senator’s perception of a bill relative to their own
priorities and perceptions, and focusing on bills more likely to pass the chamber accounts
for this logic.
I begin with testing the basic question of the theory, which is what bills are targeted
by holds. Table 1 provides results for models related to this question. The outcome
variable of the models in this table is if a bill is targeted by a hold, and the unit of
observation in these logistic regression models is a Senate originated bill. Thus, each
coefficient in Table 1 provides the relationship between the associated covariate and the
presence – or absence – of a hold.
Column 1 in Table 1 presents the model for all Senate bills. The effects of public
support for a bill are clear and follow my expectations. Holding all over variables in the
model constant, the addition of one more Cosponsor for a bill increases the probability
that a bill is targeted by a hold, on average. This can be seen graphically in Figure 3.
In short, when a greater number of senators approve of a bill, and the bill is perhaps
43
I focus on Senate bills which must be passed by both the Senate and House and signed by the
president to become laws in the following analysis. These bills, commonly referred to as ”S type bills,”
are by far the most common type of bill considered in the Senate.
44
It is true that senators can, and on occasion do, place holds on bills before the committee reports
them. The rarity of this makes treating these bills as a separate category within this analysis difficult,
however. Thus, I acknowledge that this is possible, but choose to focus on the broader framework of
Senate obstruction.
24
Table 1: Senate Bill Subject to a Hold
All Bills Majority Sponsor
0.025∗
0.027∗
(0.003)
(0.004)
Sponsor Ideology
-0.744∗
-0.915∗
(0.200)
(0.320)
Months Remaining
0.006
0.006
(0.008)
(0.009)
Policy Committee
0.258∗
0.176
(0.104)
(0.113)
Multiple Referral
0.466∗
0.379
(0.188)
(0.204)
Sponsor Seniority
0.045∗
0.060∗
(0.014)
(0.016)
Committee Leader Sponsor
-0.198
-0.324∗
(0.132)
(0.143)
∗
Republican Majority
-0.892
-0.729∗
(0.148)
(0.237)
∗
Majority Sponsor
0.570
.
(0.159)
Dole as Leader
-0.198
-0.282∗
(0.119)
(0.127)
Intercept
-2.391∗
-1.848∗
(0.238)
(0.225)
Variable
Cosponsors
N
Log-likelihood
3096
-1316.792
2513
-1133.138
Minority Sponsor
0.013
(0.008)
-1.331
(0.776)
0.004
(0.024)
0.663∗
(0.305)
0.948
(0.498)
-0.024
(0.042)
0.454
(0.361)
-1.399∗
(0.566)
.
0.380
(0.378)
-2.534∗
(0.624)
583
-176.106
Note: Estimates are from an logistic regression model. The outcome variable is coded 1 if the bill was
targeted by a hold 0 if it was not. Standard errors in parentheses. ∗ = p ≤ 0.05.
25
more likely to pass, holds are more likely to be observed on that bill.
The effects of Sponsor Ideology, measured through DW-Nominate, also follows
the above theory. The expectation above is that senators are less likely to hold bills
whose sponsors are more like themselves, as these bills will be more preferred. Table
2 demonstrates that as a sponsor’s ideology increases, or becomes more conservative,
Republicans are less likely to place a hold on this bill. Figure 4 demonstrates that as
bill sponsors become more conservative, and therefore more like a randomly selected
Republican on average, they are less likely to have holds target their bills. Senators
appear to not want to stop legislation that may be near their ideal point.
Column 1 in Table 1 demonstrates that as Months Remaining between a bill’s
introduction and the end of a Congress increases, that bill is more likely to be the target
of a hold, but this effect is not statistically significant. While a lower percentage of
bills are introduced late in a Congress, and budgets are introduced earlier, these bills
introduced later are less likely to be targeted by holds holding committee consideration
and public support constant. This lack of statistical significance is not due to the
difference between the sessions, as inserting a dummy variable for each session produces
a lower effect with no statical significance.45
The effect of committee type for all Senate bills reported from committee provides
45
Another possibility is that senators view holds as less powerful toward the end of a Congress and
choose not to utilize them. Howard and Roberts (N.d.) demonstrate that outright stoppage holds
placed toward the end of a congress are actually more effective on average, however. This indicates
that, while I cannot state with certainty that my theoretical interpretation is correct, there is something
to the structure outlined above.
26
●●●
●
●
● ●●●
●● ●
●
● ●● ●
●
●
●
●● ●
●
●
●
●●
● ●●
●
●
●
●
● ●● ●
●
●
●
●
●
●● ●●● ●●
●
●●
●
●● ● ●●
●
●
●●
●
●
●
●
●●
●
●
●●
●●
● ●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
● ●
●
● ●
●
●
● ●
●●
●
●
●
●●
●
● ●
●
●● ●● ● ● ● ●● ●
● ●
●
●●
●
●
●
●
● ●
●
●●
●
● ●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●● ●
●
●
●
●
●
●
●
●
●
●
●
●●
●● ● ● ●● ●
●
●
●
●
●●
●
●
● ●
●
●
●●
●●
●
● ●
●
●●
●
● ●
●
●
●
●
● ●
●
●
●
●
● ●●
●
●●
●
●
●
●
●
● ●●
●
● ● ●● ● ●
●
● ●
● ● ● ●●
● ●
●
●
●
●
●
●
●
● ●●
●
●●
●
● ●●●
●
●
●
●
●
●
●
● ●
● ●●
●
●
●
●
●
●
●● ●● ● ● ●
●
●
●
●
●●
●
●
●
●●
●●
●
●
●
●
●
●
●
● ●
● ●●
● ●●
●
●
●
●
●
●
●●
●●
●
●●
●
● ●●●●
●
●●
●
●
●●
● ●
●●
●●● ●
●
● ●
●
●
● ●
● ●
● ●
● ●
●
●
●
●●
● ●●
● ●
●
●
●
●
●
● ● ●
●
● ● ●
●
●
●
●● ●
●●
●
●
●
● ●●
●
● ●●
●
●
●● ●
●
●
●
●
●●
● ●
●
●●
●
● ●
●
●
● ●●
●
●
●
●
● ● ● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●● ●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
0.6
0.2
0.4
Point Estimate
95% Conf. Interval
0.0
Predicted Probability of Hold
0.8
1.0
Figure 3: Effect of Cosponsors on Probability A Bill is Held
●
●
● ● ●●
●
●●
●
●
●
●
●●
●
●
●
● ●●
● ●
●
●
●● ●
●●
●
●
●
●
●●● ●
●
●
●
● ●●● ●
●
●
●●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
● ●●
● ●●
●●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
● ●
●
●●
●
●
●
●
●● ●●
●●● ●●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●● ● ●●
●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●●
●
● ●
●
●
●
●
●●
●
●
● ●
●
●
● ●
●
●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●●
●●
●
●●
●
●
●
●● ●
●
●
●
●
● ●
●
●
●
● ●
● ●
●
● ●
●●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
● ●● ●●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●●
●
●● ●●
● ●●●
● ●● ● ●
●
●
●
● ●●
●
●
●●
● ●
●●● ●
●
●
●
●
●
●
●
●
●
●
●
●
● ●● ●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
● ●
●●
●
●
●
●
●
●
●●
●●
●●
●●
●● ●● ● ●
●
●
●
●
●
●
●● ●●
●
●●
●
●● ●
●
●
●●
●
●
●●
●●
●
●
●
●
●
● ●
●●● ●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
● ● ●●● ●
●●
●
●
●
●
●
●
●●
●●
●
●
●
●●
●
●
●
●● ●●
●
●
●
●● ●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●●
●
●●● ●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
● ●●
●
●
●
●● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●● ●●
●
●
●●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
● ●
●
●
●
●
●
●
●
●●
●
●● ●●
●
●
●
●
●
●
● ●
●●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●●
●
●
●
●
●●●● ●
● ●●
●
●
●
●
●
●
●
●
●● ●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●●
●
●
●
●
●
● ●
●
●
●
●
●●
●
●
●
●● ●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●●
● ●● ●
●
●
●
● ●
●
●
●
●
●●
●●
●
●
●
●
● ●●● ● ●
●
●
●
●
●
●
●
● ●
● ●●
●
●
●
●
●
● ● ●●● ●● ●
●
●● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●●
●
●●
●
●
●● ●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●● ●●
●
●
●
●
●
●
●
●● ●●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
● ●
●●● ● ● ● ● ●
●
● ●
●● ●● ●
●●
●
●
●
●●
● ●
●
●
●●
●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●●
●
●
●
● ●●
●●
● ●
●● ●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
● ●●
●
●
●
●
● ●●
●
●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●●
●
● ●● ●
●
●
●
●
●
●
● ●
●
●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●● ●●
●
●
● ● ●●
●
●●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
● ●
●
●
●
●● ●
● ●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
● ●
●
●
●
●
●
●
●●
●
●●
●
●
●
0
●● ●
●
●
● ● ●●● ●
●
●●
●
●
● ●
●
●
●
● ●
●●
●
●●
●
● ●
●●
20
●
●●
●
● ●
●
●
●
● ●
●
● ●
●●● ●
●●
●
●● ●
●
●
●
● ● ●
●●● ●
● ●
●
●
●
●
●●
● ●
●● ●
●
●
● ●●
●
●
●
● ●● ●
●
●●
●●
●
●● ● ●
●●●
●
● ● ●
● ●
●●
●
●
●
●●● ●
●
●
●
●
●
●●
●
●
● ●●
●
●● ●●
●
●
●
●● ●● ●
● ●
●
●
● ●
●
●
● ●●
●
●
● ● ●● ● ●
●
●
● ●
●
●
●
●
●
●
●
● ●
●
●● ●
●
●
●
●
● ●
40
●
●
●
●
●●
●●●
●
●
●
●
●
●
●●
●
● ●
●
●
●
●
●
●
●● ●
●
●
●
●
●
●
● ●
●●●
●
60
Cosponsors
27
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
80
●
100
●
●
●
●
●
● ●
●
● ● ● ●
●
●
●
● ●
●
● ●
●
●
●
●
●
●
●●
●
●
● ●
●
●
●
●●
●
●
●
●
●●
●
●
●
●
●●
●
●
●
●
●●
●
●
●●
●
●
●
●
●●
●
● ● ● ●●●
●
● ●●
● ● ●
●
●●● ● ●●
● ●●
●
●●
● ●
●●
●●●
●
●
●
●
●
●
●
●
●
● ● ● ● ●● ●●
●
●
●●
●
●
●
● ●
●
● ● ●●
●● ●
●
● ●●
●
●●
●
●●● ●
●
●
●●
●
● ●●
●
● ●●
●●
●●●
●● ●
●● ● ●●●● ●●
●
● ● ●●
● ●
●●
●
●●
●
●
● ●
●
●●
●
●●●●●
●●
●
●
● ●●●
● ●
● ●
●●
●
●
● ●
● ●●
●
● ●●
●
●●
●● ●●●● ●●
●●
●
●●
● ● ● ●●
●●
● ●
●
●
●
●
●
●●
●
●
●
●
●
●● ●
●●
●●● ●
● ●
●
●
●●
●
● ●
●
●
●
● ●●● ●
●●
● ●
●● ●
● ●
●
●
●
●
●
●
●● ●
●
●
● ●
● ●●
● ●
●●●
●
●
●
●
●●
●
●
●
●
●●
●
● ●
● ● ●●
● ●
●
●●
●
●●
●●● ●
● ●●● ●●●
●
● ●
●
●
●
●●
●● ●
●● ●
●●
●
● ●
● ●
●
●
●●
●●
●
● ● ● ●
●
● ● ● ●
●
●
●
●
● ●
● ●● ● ●
●
●
●
●
● ●
● ●
●
●●
●
●
●
●
●
●●
● ●
●
●
●
●
●●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●●● ●
●
●
●
●●
● ●
●
●
●
●
●
●
●
●
●
●●
●
● ●
●
●
●
●
● ●
●
●
●● ●
●
●●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
● ●
●
●
●
●
●●
●●
●●
● ●
●
●
●●
●
●
●
●
●
●
●
0.6
0.2
0.4
Point Estimate
95% Conf. Interval
0.0
Predicted Probability of Hold
0.8
1.0
Figure 4: Effect of Sponsor Ideology on Probability A Bill is Held
●
●
●
●
●
●
●
●
●
●
●
●
● ● ●
●
●
● ●
●
●
● ●
●
●
●
●
● ●
●
●●
●●
●
●●
● ●●
●● ●●● ●
●
● ●●● ● ●
● ● ● ●●● ● ●
●
●●● ● ●● ●●● ●
●
●●
●●
●
●
● ● ●●
●● ●●
●
●●
●● ● ● ●●
● ●
●
●
●
●
●
●
●
●●●
●
●●
●●
●
●
●●●●●
● ●
●●●
●● ●
●● ●
●
●●
●●
● ●
●
●
●●
●
● ●
●
●
●● ● ●
●
●
●●●
●
●●
● ●●●
●
● ●
●
●
●
●
●●
● ●
● ●
● ● ●
●●
●● ●●
●●
●●●
●
●●●
●
●●
●
●● ●● ●●
●●
●●
●●
●
●
●●
●
●● ● ● ●
●
● ●
●
●
●●
●
●●
●●●
● ●●
● ●
●
●
● ● ●●●
● ●
●
●
●● ●●●
●●
●
●
●
●● ●
●
●
●●●
● ●●
●
●●
●
●
●
●●
●● ● ●
●●
● ●●● ● ● ●
●
●
●
●
●●
●●● ●●
● ●
●●
●
●
●
●●
● ●●
●
● ●●
●
●
●●●
●
●●
●
●●●●
●●
●●●
● ●●
●● ●
●● ● ● ●
●
●
● ● ●●●
●
●●
●
●
●●●
●
● ●
●
● ●
●●●
●
●
●● ●● ●●●
●
●
●●●● ●
●
●
●
●
●
●●
●●
●●●●
●
●
●
●●●●●
● ●
●●●
●
●●●●●●
● ●●
● ●
●●
●
●
●● ●
●●●● ●
●
●
●● ●
●
●●
●
●●●
●●
●
●
●
●
●●● ●
● ●
●
●
●
●●
● ●● ●
●●●●
●
●●
● ●●
●
●
●
●
●
●
●●
●
●
●●●
●● ●
●
●●
● ●
●●
●●●●●
●
●●
●
●●
●● ●●●●●
●
●●●
●
●● ●
●
●
●
● ●● ● ●
●
●
● ●●
●
●
● ●
●
●
●
●●●●●
●
●●
●●
●●●
●
● ●●
●
●
●
●
●●
●
●●
● ●●
●
●
●●●●
●●
●●
●
●
● ●●
● ●●
●
●
● ● ●●● ●
●
●
●
●
●
●
●
●
●
●
●●
●
●
● ●●
● ● ●●●●
● ● ●
●
●●● ●
●● ●
●
●●● ●
● ● ●
●● ●
● ●
● ●
●●● ●
●
●
●●●●
●
●
●
●●●●
●
●
● ●
●
●
●●●
●●
●
●●● ●● ●●
●
●
●
●●
●● ●
●
●
●
●
●●
●●
●
●
● ●●●
●●
●
●●
●
● ● ●●●●
●
●●● ●
● ●●● ● ●
●● ●
●● ● ●
●●
● ●
●●●
● ●
●● ● ●
●
●● ●●
●
●
●
●
● ●
●●
●
●●
●●
●
● ●●
● ●●●●●
●
●
●
●●
●● ●
●●● ●
●
●
●●
●
●
● ●
●
●●
●
●
●
●●
●●
●
●●
●●
●●●
●
●
●
●
●
●●
●
●●
●●
●
●
●
● ●●
● ●
●
● ●
●
●
●
●●
●●● ●
●● ●●●
●
●●
● ●●
●
●
●●● ●
●
●●
●
●
●
● ●
●
●
●
●
●● ● ●
●
●● ●●
●● ●
●
●●●
●●●●
●●
●
●●
●●
●●● ●
●
●
●
●
●● ●
●●
●
●● ●●
●
●
●●
●
●● ●
● ●
●
● ●●● ●●● ●● ●●
●●●
●●
●●
●
●
●
●● ●●●● ●
●
●●
●
●●
●
●● ●
● ●● ●
●●
●
●
●
●●
●●
●●
●
● ●
● ●
●●●●
● ●●
● ● ●●
●
●
● ●●
● ●●●●
●
●
●
●●●
●●
●
●
●●
● ● ●●
●
●
●●
●● ●
●
●● ●
●
●●●● ●● ●●
●●
●
●● ●
●
●●
● ● ●●
●
●
● ●● ●
●●
●
●
●●
●● ●
●
●●
●●
●
● ● ●●
●
●
●
●
● ●● ●● ●
●
●●
●●●
●
●● ●
●●●
●● ●●●
●●
●● ●
●
●● ●
●● ●
●●
●●●
● ●●
● ●●● ●
●● ●
●●
● ●●● ● ●
● ● ●
●●
● ●● ●● ●
● ●●
● ●● ● ● ●
●●● ● ●
−0.5
●
● ●●●
●
●
●●●●
●●
● ●● ●
●●
●● ●
●● ●
●
●●
●
●
●
●
●
●●
●● ●
●
●●
●
●
●
● ●
●
●
●●
●●
●
●
● ●
●
●
●
● ● ●●●
● ●
●
●●
●
●
● ●
●●
●● ● ● ● ●
●● ●●
●
●
●● ●
●
●● ●
●
●
●
● ●
●
●●
●
●● ●
●
●
●●
●● ● ●
●
● ●●
● ●●
●
●●
●
●
● ●
●●
●●
●
●
●● ●
●
● ●
●
●
●●● ●
●
●
● ●●
●
●● ●
●●● ●
● ● ●
●
●
●
●
●● ●● ● ●●
●
●
●● ●● ●
●● ● ●● ●
●
●
●
● ●●
●
●
●
●●●● ●● ●
●
● ● ●●
●●●
●
●●
●
●
● ●●●
●●
●● ● ● ●
●●
●
● ●●
●● ●
●
●
●
●
● ●●
● ● ●●
●●
●
● ●●
●
●
●●●●
●●
●●●● ●
●
●
● ●
● ●
●
● ●●
●
● ●
● ●● ●
●
●●
● ●●
●
●
●
●
●
●
●●
●●
●
●
●●
●●●
●
●●●●
●
● ●●
● ●●
● ●
●
●
●● ●
●
● ● ●
● ●
●
●
● ●●
●●
●
● ● ●
●
●
●●● ● ●●
●●
●
●
●
● ●
●
●●
●
●●
●●●
●●●●
● ●
●●●
●● ●●
●
●
●●
●
●
●
●
●●●
●
●● ●
●
●● ● ●●●
●
● ●
● ●
●●
●● ● ●
●● ● ●
●
●
●●
●●● ●● ●
●
●
●
● ●●
●
●
●●
●●●●
● ●●
●
●
● ●● ● ●
●
●
●●
●●●●
●● ●
● ●●●
●
●
●●●
●● ●
●
●●
●● ● ● ●
● ●●
●
●
●
●●
●
●
●
●● ●
●
●●●
● ●
●
●
●●
●●
●●
● ●
● ● ● ●● ● ● ● ●
●
●
●
● ●●
●
● ● ● ● ●●●
●
●● ●
●● ●●●● ● ●●● ●
●●
●
●
●
●
●●
● ●
●
● ●●
●
●
●
●
●● ●
●●
●
●●
●
●● ●
●●
●
●
●
●
● ●
●
●
● ●
● ●
●
● ●● ●
●●
● ●
●●
● ●●●● ● ●
● ●
●
●●
●
●● ● ●● ●
● ●
●
●
●
●
●●●●
0.0
Sponsor Ideology
28
● ●
●●
●●
●●
●
●●
● ●●
● ●●
● ●●
● ●
●
● ●●●●
●
●
●
●
●● ●
● ●●
●
●
●●
●
●
●
●●
●●
●
● ●●
● ●
●
●
●
●
●
●●
●
●
●
●
● ●
●
●●
●● ●
●
●
●
●● ●
●
●
●●●
●
●
●
●
● ●●
●
●● ●
●
●
●
●● ●
● ●
● ●
●●
●●
●
●
●
● ● ●●
● ●●
●
● ●
●
●●
●
●
● ●● ● ●●
●
●
● ●●
●
●
●
●● ●
●
●
●
● ●
●●
●●
● ● ● ●●
●●
●
● ●
●
●●
●
●
● ●
● ●● ●
● ● ● ●●
●
●● ●
●
●
●
●
●●
●
●● ● ●
●●
●
●
● ● ● ● ●●
●
●
●● ●
●
● ●
●
●●
● ●●
●
●
●●●● ●
●
●
●
●● ●
● ●●
●
● ●
●
●
● ●
●●
●
●●
●
● ●●
● ●●
● ●
0.5
●
●
●
●
● ● ●
●●
●
●●● ●
●●
●
●
●
●
●●
● ●
●
●●
●
●
●
●
● ●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
● ●
●
●
● ●
● ●
●●
●●
●
●●
●
●
●
●
●
●●
●
●
●●
●
●
●●
●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
clear results, as the effect for Policy Committee is positive and statistically significant.
These committees are the most requested by senators, and offer the broadest range of
policies within bills. Thus, bills referred to these higher profile committees are more
likely to be targeted by holds on average. These bills are both more likely to become
laws and attract attention to their content.46 Regardless, it is apparent that bills are
viewed differently based on the profile of committees in which a bill was considered.
Multiple Referred bills are more likely to be targeted by holds than bills considered
in a single committee, all else constant. That is, even controlling for the type of committee a bill is considered in and the level of public support, bills considered within more
than one committee are more likely to be obstructed by Republicans. While this effect
is not surprising, the mechanism behind this result is difficult to pin down due to lack of
observationally equivalent bills. It is probable that either the greater winnowing effect
from multiple committees or the increased popularity from more members approving the
bill causes this effect to appear.47
The controls also present interesting results. A bill with a Committee Leader
Sponsor is not more likely to have holds placed on this bill, on average. Bills with
institutionally powerful senators pushing for passage are no more likely to be obstructed
46
This test does not discern between the motivation senators may have in placing a hold on a bill, as
I treat neither hold type or bill salience among the public in this model. While these explanations fit
both with the results and previous examinations of policies, it is possible that senators may place holds
on these bills in order to further their own interests and force a log roll among other senators (Arnold
1990; Krehbiel 1985).
47
The difficulty of discriminating between these explanations is that I do not know what multiply
referred bills would be like without being considered in more than one committee.
29
than bills with non-committee leader sponsors. However, greater Sponsor Seniority for
bills does make them more likely to be targeted by holds, on average. Bills are less likely
to held by Republicans when there is a Republican Majority than with a Republican
minority. Assuming that each senator introduces roughly the same number of bills, on
average, the party with a greater number of elected officials should introduce more bills
than the other party. This agrees with the intuition above that Republicans should place
fewer holds when in the majority, as a greater number of these bills are more likely to
be closer to their ideal point in expectation.
Additionally, bills with a Majority Sponsor are more likely to be the target of
a hold, on average. This is most likely due to one of two processes. First, the logic
presented above for simply a greater number of majority party bills is a possibility, but
perhaps unlikely to explain this finding. The alternate possibility is that, as Den Hartog
and Monroe (2011) find, majority parties and majority party members have a fundamental advantage in Senate legislative consideration. Bill opponents know that majority
sponsored legislation is more likely to pass the chamber due to this advantage, and will
be more likely to place a hold on this bill.
This corroboration of previous findings related to majority party advantage led to
a splitting of the data between majority party sponsored and minority party sponsored
legislation.48 These results are presented in Columns 2 and 3 of Table 1. The primary
interest here are similarities – and differences – between Republican senators targeting
48
This procedure is functionally equivalent to interacting Majority Sponsor with all other variables
in the model.
30
legislation across majority and minority sponsored legislation.
The effects of public support for a bill are consistent across both majority and
minority sponsored legislation, and corresponds with the finding for all bills. Adding
one additional Cosponsor to a bill increases the probability that said bill is targeted by
a hold, on average, for both majority and minority sponsored legislation. This effect
only reaches statistical significance for majority sponsored legislation, however. More
popular legislation is more likely to be targeted for obstructionist activity.
Similar to the results for all bills, Sponsor Ideology is negative for both majority
and minority sponsored legislation. Again, this effect does not reach statistical significance for minority sponsored legislation. As the sponsor becomes more conservative
Republicans are less likely to hold this bill. Months Remaining is also similar to the
results for all bills, with the effects for majority and minority sponsored legislation not
reaching statistical significance for either set of bills.
How senators look at legislation, as measured by the type of committee in which a
bill is considered, varies across sponsor majority status. Majority sponsored legislation
considered in a Policy Committee is more likely to be the target of holds, but does not
reach statistical significance. The effect for minority sponsored legislation is positive and
statistically significant. Thus, bills considered within high profile Policy Committees are
more likely to be the target of holds, on average.
Multiple Referral for legislation is positive for both types of legislation, but does
not reach statistical significance for either majority or minority sponsored bills. Interest-
31
ingly, Sponsor Seniority is positive and statistically significant for majority sponsored
legislation, but negative and not statistically significant for minority sponsored bills.
While having a Committee Leader Sponsor was not statistically significant and negative
for all bills, it is statistically significant and negative for majority sponsored bills but
not for minority bills. The effect of having a Republican Majority is consistent across
models, which is expected as this does not change with sponsor partisanship. The remaining covariate, Dole as Leader, provides a distinction between the period of Baker
and Dole. This effect is varies across all models, meaning that having Dole as leader
can alter the probability a bill is targeted by a hold, but may vary across institutional
circumstance of a bill sponsor. While I cannot determine why this effect exists, it does
hint at a difference between the two periods.49
These results provide explanations mostly consistent with the hypotheses presented above. Legislation that has a greater amount of public support is more likely to
be targeted by Republican holds. Bills with more ideologically conservative sponsors are
less likely to be held across sponsor status. On average, bills referred to Policy Committees are more likely to be held. The only inconsistent finding is that Months Remaining
for bill consideration does not affect the probability a bill is targeted by a hold. In
short, I find that my results generally fit with the theory presented above. Given this
49
Scholars have discussed differences in leadership strategies across both Senate chamber and committee leaders, but the effects of this have not been empirically demonstrated (Smith 1989; Evans 1991).
Also possible is that this is a rough proxy for time. Well established in the literature is that obstruction
has been increasing over time (Binder and Smith 1997; Binder, Lawrence and Smith 2002; Koger 2010;
Smith 2014). Changes in membership during this period increased polarization and potentially changed
incentives (Theriault and Rohde 2011; Theriault 2013). I cannot discern between these explanations
given the data and tests used, but a consistent difference exists between the two periods.
32
understanding of what bills are targeted, I move on to if senators receive any benefits
from holding bills.
6
The Effects of Holds on Bills
Senators desire more than to notify leaders of some vague objection to a bill
when placing a hold. Often, senators have specific goals when utilizing this procedural
prerogative. These goals can affect both the process of obstruction and how successful
a senator is in achieving her objectives. This section addresses if holds supply senators
with their desired outcomes across the range of stated objectives.
To begin, I return to the process of obstruction presented in Figure 2. Senators
choose to hold or not, after which leaders choose a strategy given this choice, and finally
senators choose to obstruct on the floor or not. While this process fits with all holds
across all bills, holds may affect bills in different ways. The reason for these differences
are that holds are not interchangeable with one another. Senators can not only object
to bills in an effort to stop passage in a fashion similar to objection through classic
depictions of filibusters, but can also express desire to change the content of the bill or
simply request more time and notification of motions to proceed. Howard and Roberts
(N.d.) and Evans and Lipinski (2005) develop classifications of holds depending on a
senator’s expressed goal in their hold letter sent to the party leader, and this section
will follow the six-part system of Howard and Roberts (N.d.).50 These differences in
50
Howard and Roberts (N.d.) develop a six part coding system for Republican holds observed under
Dole. The system is as follows. 1) Outright Hold: Letters that threaten to object to a unanimous
33
holds allow me to address if members were given what they wanted in their hold request
during the process of bill consideration.
6.1
Do Holds Stop Bill Passage?
The first type of hold, outright holds which seek to stop passage of the legislation
with no supplied bargaining recourse for party leaders, supplies the first test. To see why
this strategy might both succeed and fail, take two cases. First, the example Jesse Helms’
and Lauch Faircloth’s holds on the Hatch Act Reform Amendments of 1988 which led
off this paper supplies an illustration of successful usage of this strategy. Both senators
stated that they would object to any time agreement or UCA on the floor, and the bill
never reached the floor. The Neighborhood Schools Improvement Act, introduced on
January 14, 1991, presents an example of the lack of success for this strategy.51 After
the bill was reported on April 19, Jesse Helms (R-NC) placed an outright hold on the
bill which stated that he would object to any unanimous consent or time request related
to this bill. The party leadership chose not to honor this outright hold, and proceeded
to floor consideration and eventual passage.
This second example demonstrates that this strategy is not always successful, and
consent request without providing any statement about what could be done to appease the senator. 2)
Specific Amendment Request: Letters that threaten to object to a unanimous consent request unless a
particular amendment is allowed to be offered to the bill. 3) General Amendment Request: Letters that
threaten to object to a unanimous consent request unless an unspecified amendment is allowed to be
offered to the bill. 4) Specific Demands: Letters that threaten to object to a unanimous consent request
unless a particular provision of the bill is removed. 5) Requests for Notification: Letters that ask to be
notified prior to a UCA being entered into without making any requests or demands. 6) Requests for
Delay: Letters that ask that consideration of a bill be delayed in order to gather more information or
hold a committee hearing. For the case studies presented below I will be focusing on types 1 through 4.
51
This bill sought to expand resource provisions for achieving the National Education Goals.
34
the possibility exists that the example of the Hatch Act Reform Amendments of 1998
is an exception to the general process. To test if holds generally, and outright holds
specifically, affect passage of Senate bills, Table 2 presents logistic regression models for
Senate passage of Senate bills. Thus the coefficients this table provide the relationship
between an increase in the covariate and the probability of a bill passing the Senate.
To test the effects of all holds and outright holds as separate categories I estimate two
models with each set of data, with Column 1 in Table 2 relating to all holds and Column 2
relating to only outright holds. The changes between the models are thus an alternation
of the primary covariate of interest, Held Bill, with Column 2 as a subset of Column
1. The theory above, as well as many previous studies on Senate procedure, assumes
that holds lower the probability of Senate passage, so the presence of a hold should be
negatively related to bill passage for both models in Table 2.
Column 1 in Table 2 provides a test of the relationship between holds and passage for all Senate bills referred from committee. Strikingly apparent in this model is
the relationship between the presence of a hold on a bill and its probability of passage.
Controlling for traditional passage effects, including cosponsors, time remaining at introduction, and chamber configuration, the presence of a hold dramatically lowers the
probability that a bill passes the Senate on average. Holds on bills do affect bill passage, significantly lowering the probability that bills become laws both statistically and
substantively.
As discussed above, holds have different purposes including attempting to directly
35
Table 2: Bill Passage in the Senate
Variable
Model 1 Model 2
Held Bill
-0.961∗
.
(0.106)
Killer Held Bill
.
-0.962∗
(0.135)
∗
Cosponsors
0.006
0.004
(0.003)
(0.003)
Sponsor Ideology
0.443∗
0.483∗
(0.130)
(0.129)
∗
Months Remaining
0.025
0.023∗
(0.006)
(0.006)
Policy Committee
-0.110
-0.129
(0.082)
(0.081)
Multiple Referral
-1.047∗
-1.056∗
(0.180)
(0.179)
Sponsor Seniority
-0.013
-0.014
(0.011)
(0.011)
Committee Leader Sponsor
0.068
0.074
(0.098)
(0.097)
∗
Republican Majority
-1.065
-1.056∗
(0.097)
(0.097)
Majority Sponsor
-0.035
-0.078
(0.100)
(0.099)
Dole as Leader
-0.578∗
-0.630∗
(0.091)
(0.092)
Intercept
0.847∗
0.896∗
(0.168)
(0.168)
N
3096
3096
Log-likelihood
-2005.515 -2022.123
Note: Estimates are from an logistic regression model. The outcome variable is coded 1 if the bill
passed the Senate and 0 if it did not. Standard errors in parentheses. ∗ = p ≤ 0.05.
36
kill legislation. These killer holds are treated separately in column 2 of Table 2, which
presents a logistic regression model for Senate bill passage with killer holds as a separate
sub-category of holds. Killer holds should stop passage, as this is the implied goal
when utilizing these holds. The effect for Killer Held Bill is negative and statistically
significant, implying that bills targeted by these killer holds are much less likely to pass
the Senate. Holds do stop the passage of legislation, both for all holds taken together
and killer holds separately.
6.2
Holds Beyond Killing Bills
Holds are not solely designed and submitted to kill legislation.
Howard and
Roberts (N.d.) demonstrate that, depending on the Congress, between 10% and 42% of
holds actually seek to alter legislation through either adding amendments or removing
parts of the bill.52 To see an example of how this looks in practice, take Mitch McConnell
(R-KY) and the The National and Community Service Act of 1990.53 McConnell sought
to have his right to offer an amendment about regulating nonprofit organizations protected in a hold submitted to Republican Leader Dole on February 20, 1990. The party
leadership allowed McConnell to present his amendment to the Senate in deliberation
over the bill after this hold was received. While the amendment ultimately failed through
being tabled, McConnell’s hold was honored in that he was allowed to offer his amend52
These are types 2, 3, and 4 holds in the coding system discussed above.
53
This bill, which sought to streamline community service and volunteer programs and their regulation
in the United States, was Introduced in the Senate on July 27, 1989 and reported from the Committee
on Labor and Human Resources on October 27, 1989.
37
Table 3: Amendment Hold Demands Accommodated
Congress Not Accommodated Accommodated Total
101
69
26
95
(72.63%)
(27.37%)
102
40
11
51
(78.43%)
(21.57%)
103
14
6
20
(70.0%)
(30.0%)
104
62
22
84
(73.81%)
(26.19%)
Total
185
65
250
(74.0%)
(26.0%)
Note: Cells are the number of amendments in a Congress by presented status. Row percentages in
parentheses.
ment on the floor. Thus the question of a hold is not always if a bill passes or if an
amendment is adopted, but rather if a senator is allowed to propose items referenced in
a hold.
Table 3 provides an exploration of if holds requesting the ability to alter bills are
honored, utilizing data on every floor amendment presented in the 101st through 104th
Congresses.54 The data utilized here are all holds regardless of bill type, and not simply
Senate originated bills as discussed above. This table takes each hold requesting either
specific or general amendment requests, along with requests to remove parts of the bill,
and asks if related amendments were presented on the floor. Two aspects of these data
are immediately apparent from this table. First, far more requests are not honored than
are honored. That is, a member requesting the ability to offer an amendment changing
the bill was far less likely to be able to present this amendment on the floor. While I do
54
I thank Chris Den Hartog and Nate Monroe for making their amendments data available to me.
38
Table 4: Final Passage Votes by Amendment Allowed on the Floor
Accommodated Not Accommodated
Total
Yea Vote
21
23
44
(58.33%)
(60.53%)
(59.46%)
30
Nay Vote
15
15
(41.67%)
(39.47%)
(40.54%)
Total
38
36
74
Note: Cells are the number of final passage votes by presented status. Column percentages in
parentheses.
not treat different processes for amendments not being presented, the difference between
presented and not is consistent in every Congress.55
Second, the lack of floor amendments allowed for Republican holds in the 104th
Congress is striking. This is the only Congress in this sample with a Republican majority,
and is moreover the first Congress in more than 30 years to have a Republican House
majority. These structural advantages do not appear to equate to Republicans allowing
co-partisans to change bills, however. This suggests that amendment related holds
were viewed as potentially valuable by individual senators, but majority status does
not necessarily confer disproportionate advantage to individual senators. The allowed
percentage for the 104th is almost exactly the average for this period, with 26.19%
allowed. While these hold requests were regularly utilized by senators, with 250 in an
eight year period, they are only successful 26% of the time.
Given that senators are often not allowed to offer their requests on the floor,
the next question is if this alters votes on bills. Table 4 provides an overview of this
55
These processes include if a bill is never considered on the floor, if no amendments are allowed, or
if a senator’s amendment is not allowed while others are allowed.
39
question, taking all amendment related holds and asking if the senator requesting the
hold voted for or against the bill on final passage.56 The patterns in this table suggest
that there is minimal difference in voting patterns on final passage for senators whose
holds are honored versus those who are not. The observed pattern is the opposite of
what would be expected, with 60.53% of senators whose holds are not honored voting
yes, and 58.33% of senators whose holds are not honored voting yes. In short, it is not
apparent from looking at final passage proceedings that leaders pay a heavy price for
ignoring the wishes of individual senators.
These examples, statistical analyses, and descriptive statistics tell a complicated
story about the effect of holds on legislation and outcomes. Apparent in Table 2 is that
holds do stop bill passage in the Senate, suggesting that holds are indeed a valuable tool.
Looking into the specific patterns for honoring amendment holds and voting patterns
suggests that the effects are not as clear as it may appear. A large majority of these
holds are not honored by leaders, and senators whose amendment holds are not honored
vote almost identically with senators whose holds are honored. While senators utilizing
holds are able to push for consideration of possible changes, specific provisions they favor
can be ignored by both the leadership and the chamber. Furthermore, obstruction can
be a valuable tool, but its presence may send a signal to the leader about the bill which
a senator does not intend. In short, it is apparent that holds do alter proceedings, but
their direct effect on the content of legislation outside of providing signals about possible
56
There were two votes in which a senator with an amendment request abstained from the final
passage vote. These two observations are not used in this table.
40
objection remains unclear.
7
Discussion
This paper has concerned itself with the place of holds within the legislative process
and how they are utilized. It provides an explanation of two fundamental processes in
legislative deliberation. First, I have established a process through which both leaders
and members knowingly allow obstruction on a bill-by-bill basis. Members are served
well, as they can protect their interests, but leaders also gain by receiving information
about potentially difficult legislation. Second, the theory and models above provide the
first explanation for both holds affecting bill passage and which bills we can expect to
be subjected to holds.
From this discussion it is clear that holds do alter the probability that bills pass
the Senate. This result is perhaps not surprising given what we expect from previous
studies of the filibuster, but this is the first systematic demonstration that holds do lower
the probability a bill is passed by the Senate. Furthermore, this result relates to only
Senate bills reported by Senate committees, a tougher test for obstructionist activity
than sampling from all possible legislation.
My findings also speak to the relationship between individual bills, institutional
circumstances, and opposing senators. It is apparent in the models and discussions
above that senators view bills differently when choosing to engage in obstruction through
holds. The content of legislation appears to structure the decision to hold a bill, but the
41
specifics of this circumstance may depend on the majority status of a bill sponsor. That
support for a bill affects decisions to register a hold is not surprising, but important to
remember when considering how cosponsorship affects bill passage rates. Additionally,
senators seem to be attentive to the identity of a bill sponsor when utilizing holds. Bills
in Republican majority Congresses are targeted less frequently by Republican senators,
and the ideological conservatism of the sponsor affects hold usage. The effects of time
were in the expected direction, but not statistically significant. Holds may be more
effective as Congresses draw to an end, but slowing the entire agenda at a Congresses
end does not drive hold usage. The process behind how support and time may interact,
as well as how supporters and opponents utilize different types of obstruction together,
are worthwhile areas for future research.
More visible forms of obstruction, such as filibusters and pushing for cloture votes,
do receive a great deal of attention. This study makes the case that understanding holds
provides valuable insight into the lawmaking process. Viewing filibusters as primary
evidence of obstruction is a potentially fruitful exercise, but our understanding could
be greatly enhanced with careful analysis of off-floor obstruction. This project pursues
the broad strokes of when obstruction is utilized, but does not deal in detail with how
it is related to other processes and its micro-level effects. Further work is needed to
understand when obstruction is actually successful in its goals and when leaders can
safely choose to ignore a senator’s registered objection. While focusing on the floor is
valuable, pursuing how legislative bargaining works prior – and parallel – to this visible
42
deliberation has the potential to greatly enhance what we know about the American
political process.
43
References
Ainsworth, Scott and Marcus Flathman. 1995. “Unanimous Consent Agreements as
Leadership Tools.” Legislative Studies Quarterly 20:177–195.
Aldrich, John. 1995. Why Parties? The Origin and Transformation of Political Parties
in America. University of Chicago Press.
Arnold, R. Douglas. 1990. The Logic of Congressional Action. Yale University Press.
Arrow, Kenneth. 1951. Social Choice and Individual Values. Wiley Chapman & Hall.
Austen-Smith, David and Jeffery Banks. 1988. “Elections, Coalitions, and Legislative
Outcomes.” American Political Science Review 82:405–422.
Baron, David P. and John A. Ferejohn. 1989. “Bargaining in Legislatures.” American
Political Science Review 83:1181–1206.
Binder, Sarah. 1997. Minority Rights, Majority Rule: Partisanship and the Development
of Congress. Cambridge University Press.
Binder, Sarah. 2003. Stalemate: Causes and Consequences of Legislative Gridlock.
Brookings Insitution Press.
Binder, Sarah, Eric Lawrence and Steven Smith. 2002. “Tracking the Filibuster, 1917
to 1996.” American Politics Research 30:406–422.
Binder, Sarah and Steven Smith. 1997. Politics or Principle: Filibustering in the U.S.
Senate. Brookings Insitution Press.
Cox, Gary and Matthew McCubbins. 2005. Setting the Agenda: Responsible Party
Government int he U.S. House of Representatives. Cambridge University Press.
Cox, Gary and Matthew McCubbins. 2007. Legislative Leviathan: Party Government
in the House. Cambridge University Press.
Deering, Christopher and Steven Smith. 1997. Committees In Congress. CQ Press.
Den Hartog, Chris and Nathan Monroe. 2011. Agenda Setting in the U.S. Senate: Costly
Consideration and Majority Party Advantage. Cambridge University Press.
Evans, Lawrence. 1991. Leadership in Committee: A Comparative Analysis of Leadership
Behavior in the U.S. Senate. University of Michigan Press.
Evans, Lawrence and Daniel Lipinski. 2005. Obstruction and Leadership in the U.S.
Senate. CQ Press chapter 10, pp. 227–248.
44
Evans, Lawrence and Daniel Lipinski. N.d. “Holds, Legislation, and the Senate Parties.”.
Fenno, Richard. 1973. Congressmen in Committees. Scott Foresman & Company.
Gamm, Gerald and Steven Smith. 2002. Policy Leadership and the Development of the
Modern Senate. In Party, Process, and Political Change: New Perspectives on the
History of Congress. Stanford University Press.
Gilligan, Thomas and Keith Krehbiel. 1987. “Collective Decisionmaking and Standing
Committees: An Informational Rationale for Restrictive Amendment Procedures.”
Journal of Law, Economics, & Organization 3:287–335.
Howard, Nicholas and Jason Roberts. N.d. “Holding up the Senate: Republican Holds
Under Dole.”.
Jenkins, Jeffrey and Michael Munger. 2003. “Investigating the Incidence of Killer
Amendments in Congress.” Journal of Politics 65:498–517.
Jones, Bryan and Frank Baumgartner. 2005. The Politics of Attention: How Government
Prioritizes Problems. University of Chicago Press.
Koger, Gregory. 2006. “Cloture Reform and Party Government in the Senate, 1918
1925.” Journal of Politics 68:708–719.
Koger, Gregory. 2010. Filibustering: Political History of Obstruction in the House and
Senate. University of Chicago Press.
Krehbiel, Keith. 1985. “Obstruction and Representativeness in Legislatures.” American
Journal of Political Science 29:643–659.
Krehbiel, Keith. 1991. Information and Legislative Organization. University of Michigan
Press.
Krehbiel, Keith. 1998. Pivotal Politics: A Theory of U.S. Lawmaking. University of
Chicago Press.
Krutz, Glen. 2005. “Issues and Institutions: Winnowing in the U.S. Congress.” American
Journal of Political Science 49:313–326.
Lee, Frances. 2009. Beyond Ideology: Politics, Principles and Partisanship in the U.S.
Senate. University of Chicago Press.
Lee, Frances and Bruce Oppenheimer. 1999. Sizing Up The Senate: The Unequal Consequences of Equal Representation. University of Chicago Press.
Lynch, Michael and Anthony Madonna. N.d. “The Vice President in the U.S. Senate:
Examining the Consequences of Institutional Design.” Forthcoming.
45
Mayhew, David. 1974. “Congressional Elections: The Case of the Vanishing Marginals.”
Polity 6:295–317.
Oppenheimer, Bruce. 1985. Changing Time Constraints on Congress: Historical Perspectives on the Use of Cloture. In Congress Reconsidered. CQ Press.
Poole, Keith and Howard Rosenthal. 2007. Ideology and Congress. Transaction Publishers.
Riker, William. 1980. “Implications from the Disequilibrium of Majority Rule for the
Study of Institutions.” American Political Science Review 74:432–446.
Rohde, David. 1991. Parites and Leaders in the Postreform House. University of Chicago
Press.
Schickler, Eric. 2001. Disjointed Pluralism: Institutional Innovation and the Development of the U.S. Congress. Princeton University Press.
Sinclair, Barbara. 1989. The Transformation of the U.S. Senate. Johns Hopkins University Press.
Sinclair, Barbara. 2002. The ”60 Vote Senate”” Strategies, Process, and Outcomes. In
U.S. Senate Exceptionalism. The Ohio State University Press.
Smith, Steven. 1989. Call to order: Floor politics in the House and Senate. Brookings
Insitution Press.
Smith, Steven and Marcus Flathman. 1989. “Managing the Senate Floor: Complex Unanimous Consent Agreements since the 1950s.” Legislative Studies Quarterly
14:349–374.
Smith, Steven S. 2014. The Senate Syndrome: The Evolution of Procedural Warfare in
the Modern U.S. Senate. University of Oklahoma Press.
Theriault, Sean M. 2013. The Gingrich Senators: The Roots of Partisan Warfare in
Congress. Oxford University Press.
Theriault, Sean M. and David W. Rohde. 2011. “The Gingrich Senators and Party
Polarization in the U.S. Senate.” Journal of Politics 73:1011–1024.
Wawro, Gregory and Eric Schickler. 2004. “Where’s the Pivot? Obstruction and Lawmaking in the Pre-Cloture Senate.” American Journal of Political Science 48:758–774.
Wawro, Gregory and Eric Schickler. 2006. Filbuster: Obstruction and Lawmaking in the
U.S. Senate. Princeton University Press.
46