Building Presidential Coalitions Among Cross-Pressured Members of Congress

• Published date: 01 March 1988
• Author: Cary R. Covington
• DOI: 10.1177/106591298804100104
• Date: 01 March 1988
• Subject Matter: Articles

BUILDING PRESIDENTIAL COALITIONS AMONG
CROSS-PRESSURED MEMBERS OF CONGRESS
CARY R. COVINGTON
University of Iowa
PRESIDENT’S success with Congress rests in large part on factors
beyond his immediate control: the partisan and ideological com-
-Z.~L position of Congress, and, to a lesser extent, his public popular-
ity (Bond, Fleisher and Northrup, 1986). However, no matter how favora-
ble a balance these factors form for a president, they do not endow him
with a legislative majority that he can assume will form naturally to sup-
port his positions. Rather, he must invariably take action to create his
majorities.
The key to a president’s successful coalition-building efforts often de-
pends on marginal sources of support within the Congress. Many mem-
bers are caught between the conflicting demands of the president on the
one hand, and constituents, interest groups and others on the other hand.
As a consequence, these members are an unreliable source of support for
the president. They may frequently be predisposed to help him, but un-
willing to take a clear position that contradicts the view of these other
groups.
To maximize the support that cross-pressured members give to the
president, he must identify and encourage the use of alternative means
and contexts that allow them to support his positions without alienating
their other supporters. The most common form of support that presi-
dents seek from these members is favorable votes on unrecorded votes,
such as teller and voice votes. Since the members’ actions cannot be
documented on such votes, they cannot easily be held accountable for
their actions by those who hold opposing positions.
Ideally, to observe how presidents build support among cross-
pressured representatives, we would study how members voted on these
unrecorded votes. Of course, the characteristic that makes these votes
an attractive context for engaging in this kind of behavior also renders
Received: July 14, 1986
Revision Received: December 12, 1986
Accepted for Publication: December 16, 1986
NOTE: I wish to extend my thanks to all those who helped create the data set upon which
this paper is based. A grant from the University of Iowa for the summer of 1983 ena-
bled to me visit the Lyndon Johnson Presidential Library to collect the records on
presidential support scores compiled by the White House Office of Congressional Re-
lations from 1961 to 1967. The archivists at the Johnson Library were very helpful
in locating the relevant files. Sara Brandes, Gregory Markko, and Laura Behrens coded
and aided in the analysis of the data. Jim Grifhorst and Chia-Hsing Lu provided invalu-
able assistance in turning the raw data into variables appropriate for analysis. Of course,
none of these individuals bear any responsibility for the analysis or interpretations der-
ived from the data.

---

48
them impervious to empirical analysis. Therefore, while this study would
most appropriately focus on, has implications for, and draws it sig-
nificance in large part from this broader range of unrecorded congres-
sional voting behaviors, data limitations restrict it to the analysis of
roll-call votes.
Consequently, this study examines the period 1961-67 to determine
the extent to which cross-pressured Democratic members of the House
of Representatives helped Presidents Kennedy and Johnson build majori-
ties on roll-call votes through the use of alternative forms of support.’
1
The remainder of the paper is divided into four sections. In the first sec-
tion, I will demonstrate that presidents have engaged in efforts to build
majorities through appeals to cross-pressured members. To that end, I
will identify the forms through which cross-pressured members can pro-
vide support and minimize their opposition to a president on a roll-call
vote. Then I will present evidence showing that those methods were em-
ployed by the Kennedy and Johnson administrations. On the basis of that
evidence, I will develop hypotheses about how different groups of mem-
bers of Congress should behave if those methods were used. The second
section will describe the data and methods used in the study. In the third
section, I will present evidence that tests those hypotheses. Finally, on
the basis of those findings, I will offer a number of conclusions and im-
plications for our understanding of how presidents conduct their rela-
tions with Congress.
THEORETICAL CONSIDERATIONS
The President’s Congressional Context
A variety of actors pressure members of Congress as they decide
whether and how
to vote: constituencies, fellow representatives, institu-
tional and regional party leaders, interest groups, the media, and the mem-
bers’ own staffs (Dexter 1969; Clausen 1973; Kingdon 1973). These
pressures have historically tended to break Congress into three groups:
The president’s core support, his core opposition, and cross-pressured
members (Fleisher and Bond, 1986). When the voting cues that represen-
tatives from the president’s party receive from these other sources agree
with the president’s agenda, the members are consonantly pressured and
can usually be counted on to vote consistently for the president’s posi-
tions. They form his core supporting coalition. The president’s primary
task with these members is to keep them informed of his position on key
issues and mobilize them when those measures come to a vote (Holtz-
man 1970: 250). Conversely, representatives from the opposition party,
especially those who are subject to voting cues that are consistently an-
1
The choice of house and years used in this study was determined by practical considera-
tions. The analysis is base on records collected within the White House by the presi-
dents’ Office of Congressional Relations. Similar records for the Senate for the same
years, and both houses for other administrations are know to exist, but are not yet
collected or prepared for analysis. Thus, the current boundaries of this study can and
will be expanded in the future.

---

49
tagonistic toward the president’s agenda, will be discounted as virtually
unapproachable and consigned to his opposition coalition.
The key to a president’s success rests with the cross-pressured
representatives, those who are faced with conflicting voting cues. This
bloc has historically consisted primarily of southern Democrats, who are
pulled by their constituency, salient interest groups, and/or their personal
ideology toward conservative positions, yet at the same time are pres-
sured by a common party bond, often manifested in appeals by institu-
tional or regional party leadership as well as by the administration, to
support the more liberal initiatives of their party’s president. In such sit-
uations, the members’ desire to be reelected usually leads them to give
precedence to the wishes of their own core of supporters within their
constituencies (Kingdon 1973; Clausen 1973). However, differences ex-
ist among cross-pressured members in the intensity and salience of those
competing pressures and the depth of allegiance each feels to the presi-
dent. Consequently, the degree of support each provides the president
can vary. Those who, despite conflicting demands, still find it possible
to vote frequently with the president are more likely than nonsupporters
to engage in alternative means of showing support as well.
T’he Role of Visibility in Determining Support
Visibility refers to ease with which an outside actor (e.g., an interest
group, an electoral constituency) can determine that a member voted
against its interests and for the president’s postion on a piece of legisla-
tion. The less visible the support requested by the president, the less likely
that outside group is to associate the member with an unpopular posi-
tion, and so the more likely that member is to go along with the presi-
dent’s request. Visibility varies according the type of motion being
considered, and the form of participation the members use to register
their position.
Some types of motions are more visible to outsiders than are others.
For the purposes of this paper, it is not necessary to establish the relative
visibility of all motions. Rather, it is sufficient to demonstrate that, in
general, votes cast on the final passage of a bill are more visible than votes
cast at earlier stages in the consideration of that bill. A cross-pressured
member who votes for the president’s position on final passage clearly
contradicts the wishes of those pressuring the member to oppose the presi-
dent. Especially before the recent proliferation of interest group voting
indices that systematically analyze all roll-call votes to assess a member’s
s
position on a particular topic, final passage votes were likely to define
that member’s position on the issue for most outsiders. In contrast, other
interim roll-call motions draw less attention and/or establish a less clear
position for the member on the bill. These votes, while often critical to
the ultimate success or failure of the president’s legislation, can be lost
in the shuffle, or explained away by representatives on grounds divorced
from the substance of the legislation. Two examples illustrate the use of
this ploy. Theodore Sorenson reports that President Kennedy used ap-

---

50
...