Examining Legislative Cue‐Taking in the US Senate

• Published date: 01 February 2015
• DOI: http://doi.org/10.1111/lsq.12064
• Author: Anand Edward Sokhey,Janet Box‐Steffensmeier,Josh M. Ryan
• Date: 01 February 2015

JANET BOX-STEFFENSMEIER

The Ohio State University

JOSH M. RYAN

Bradley University

ANAND EDWARD SOKHEY

University of Colorado at Boulder

Examining Legislative Cue-Taking

in the US Senate

Weexamine congressional cue-taking theory to determine its extent, conditionality,

and various forms in the US Senate. Using a novel data-collection technique (timed

C-SPANfootage), we focus on temporal dynamics via event history analysis. Examining

the effects of senator characteristics across 16 votes from the 108th Congress, we f‌ind that

committee leadership and seniority generally predict cue-giving, while other types of

characteristics predict cue-giving on certain types of votes. Our results underscore the

importance of considering the order and timing of voting when studying congressional

behavior.

“On agriculture, all the Republicans would come into the chamber and ask, ‘How did

Conrad vote?,’ because everyone knew Conrad Burns was the expert on agriculture

issues.”

—Senior Legislative Assistant to Senator JohnThune

Pushed by a new Democratic president and their colleagues in

the House, the 111th Senate was expected to deal with health care

reform, economic issues, two wars, and the conf‌irmation of Obama’s

appointees. These issues were the most well-known, but this Senate

also had to deal with less salient legislative items, as do all Congresses.

For example, in the midst of securing f‌inal passage on its version of

health care reform in December of 2009, the Senate faced a number of

other pressing items that the New York Times listed as “must-pass”:

raising the debt ceiling, seven spending bills (or an emergency appro-

priations bill), and renewing expiring provisions of the Patriot Act, the

federal estate tax, highway construction programs, and unemployment

assistance.1In addition to the bills themselves, the Senate also had to

deal with each bill’s accompanying—and sometimes sizable—set of

amendments.2

bs_bs_banner

LEGISLATIVE STUDIES QUARTERLY, 40, 1, February 2015 13

DOI: 10.1111/lsq.12064

© 2015 The Comparative Legislative Research Center of The University of Iowa

Senators are expected to cast votes on procedures, amendments,

and f‌inal passage on virtually every issue on the public agenda, whether

big or small, prominent or inconspicuous. And, despite the substantial

increase in the congressional workload, senators must be careful not to

cast votes that could potentially alienate their constituents, party,relevant

interest groups, or donors and lobbyists. In the 105th Congress (1997–

98), senators recorded 633 f‌loor votes, while their colleagues in the

House logged 1,187 votes (Ornstein, Mann, and Malbin 1997). On many

major issues, existing research suggests party-line votes or leadership

instructions provide guidance. But, much of what Congress does is

neither politically divisive nor particularly salient. The normal business

of the institution receives less attention but constitutes the bulk of the

congressional workload. In the different information environments of

Congress, senators sometimes must rely on their fellow members’ votes

as cues to help them make their choices; this requires that some senators

act as cue-givers by providing signals to other senators, while other

senators act as cue-takers, receiving signals and making voting decisions

based on them.

While cueing likely occurs in both chambers, we focus on the

Senate for two reasons. First, the Senate provides competing theoretical

expectations: on one hand, it seems a fertile environment for cue-taking

behavior because senators have less specialized knowledge; on the other,

cue-taking may be less likely as senators have greater access to

information-gathering resources than their House counterparts. Put dif-

ferently, because there are fewer members in the Senate, each senator

serves on more committees than their House counterparts. However,

senators are generally considered to have less in-depth knowledge about

specif‌ic issues than members of the House. We see greater potential for

cue-giving and cue-taking opportunities given the “breadth versus

depth”/more general versus less specialized dynamic of the chamber.

Second, the unique voting procedure of the Senate provides a

straightforward way of testing cue-taking theory. Data availability is a

factor given our desire to test theoretical claims empirically. Although the

electronic voting board in the House captures the order of voting, that

information is not released to the public, and note taking is not permitted

in the House or Senate Galleries. In the House, C-SPAN coverage does

not capture enough of the voting board or action by the members to

back-out complete voting-order information. However, C-SPAN cover-

age does permit us to retrieve such information in the Senate.

We explore the extent to which cue-taking behavior occurs, the

conditions under which it is most likely (i.e., controversial vs. consensual

votes), and which senators serve as cue-givers. Our main results tell a

14 Box-Steffensmeier, Ryan, and Sokhey

simple story: the order of voting matters in the Senate. Senate voting

behavior on normal, everyday issues depends largely on the ability of

some senators to leverage their expertise, providing cues for others who

may face uncertainty. This type of voting allowing members to manage

workloads and cast votes consistent with their preferences, particularly

when engaging in information-gathering activities, may be costly or

ineff‌icient. In the sections that follow, we begin by looking at voting and

information in the Senate before reviewing the literature on cue-taking

theory, noting its contributions and limitations. We then discuss the

critical importance of examining timing when analyzing cue-giving and

cue-taking, making the case for our unique dataset.3Finally, we present

the evidence on cue-taking from a series of event history models—an

approach that is ideally suited for examining temporal questions.

Voting and Information in the Senate

One Senate staff member noted that “[i]t is literally impossible for

every single senator to hold in their head every single issue that comes

before any body as diverse and complex as the Senate.”4Indeed, senators

are faced with such an immense amount of work, and such multifaceted

and complicated issues, that they are often required to use information

shortcuts when trying to decide how to vote (Porter 1974), or they may

simply decide to ignore a certain set of issues if those seem to offer little

payoff (Woon 2009). As the public agenda has expanded, so have the

demands placed on members of Congress. One consequence of this is the

increasing scarcity of f‌loor time to discuss and debate bills and amend-

ments, as evidenced by the rise in complex unanimous consent agree-

ments as tools for managing the business of the f‌loor (Smith and

Flathman 1989) and the increasing power of the hold and f‌ilibuster as

tools of obstruction (Binder and Smith 1997; Koger 2010; Smith 1989b;

Wawro and Schickler 2006).

The time and information demands Senators face are, for multiple

reasons, likely greater than those faced by House members. First, unlike

the House, the Senate does not have a germaneness rule that prevents

surprise amendments or limits the number of possible issues on any

given vote.5Second, a majority coalition in the House can limit the

number, scope, and type of amendments offered through its use of rules

generated by the Rules Committee. This tool is typically used by the

majority party to further its own agenda and restrict the rights of the

minority party (Cox and McCubbins 2005). Finally, the House operates

under strict time limits which may reduce the number and scope of

amendments proposed.6For these reasons, the Senate lacks tools to limit

Legislative Cue-Taking 15