Consistency versus Responsiveness

• Date: 01 March 2017
• DOI: 10.1177/1065912916670271
• Published date: 01 March 2017
• Author: Adam F. Cayton
• Subject Matter: Articles

Political Research Quarterly

2017, Vol. 70(1) 3 –18

© 2016 University of Utah

Reprints and permissions:

sagepub.com/journalsPermissions.nav

DOI: 10.1177/1065912916670271

journals.sagepub.com/home/prq

Article

Why do Members of Congress change positions over

time? Legislative responsiveness to constituents is funda-

mental to democratic representation and presents an

interesting theoretical puzzle. On one hand, normative

expectations of representatives in a democratic system

hold that when constituents’ attitudes and economic cir-

cumstances change, lawmakers should adjust their behav-

ior accordingly. Regardless of whether one believes that

the representative’s role is to obey public opinion (Dahl

1956; Key 1964) or to act in constituents’ interests regard-

less of what they think (Burke 1774), democratic theory

and conventional wisdom hold that legislators should

respond to changing conditions by adjusting their behav-

iors, including the policy positions taken on behalf of

constituents. On the other hand, politicians face compet-

ing pressure to be leaders, support the party line, and risk

being labeled “flip-floppers” if they change positions

(Hayes, Hibbing, and Sulkin 2010). Voters may view this

behavior as unreliable (Bender and Lott 1996; Lott and

Bronars 1993) or even dishonest, and partisan activists

may prefer a more consistent candidate in the next pri-

mary (Masket 2009). Remaining steadfast against a tide

of public opinion and changing economic reality and

switching positions to adapt to new circumstances both

come with political risks. In this paper, I present a theory

of how members of Congress balance these risks and test

it using indicators of economic well-being between 2007

and 2010.

In addition to their implications for democratic respon-

siveness, the findings of this study help us understand

how aggregate responsiveness occurs (Stimson, Mackuen,

and Erikson 1995; Wlezien 1995) even though individual

members of the House rarely reverse positions (Clausen

1973; Poole 2007; Poole and Rosenthal 1997; Stratmann

2000). Stimson, MacKuen, and Erikson find that much

aggregate policy change happens between elections

through “rational anticipation.” This study sheds light on

who rationally anticipates a change in the public mood

and adjusts their positions accordingly and under what

circumstances they do so.

I test the theory that members of the House are more

likely to reverse a previously held position when their

district is affected by an issue in a way that makes their

prior vote less compatible with current conditions. For

example, if a representative votes to terminate federal

unemployment benefits in one year, but then unemploy-

ment increases in her district, in the next year, she will be

more likely to support unemployment benefits. This

expectation seems intuitive in light of the extensive

670271PRQXXX10.1177/1065912916670271Political Research QuarterlyCayton

research-article2016

1University of West Florida, Pensacola, USA

Corresponding Author:

Adam F. Cayton, Department of Government, University of West

Florida, 11000 University Parkway, Building 50, Pensacola,

FL 32514, USA.

Email: acayton@uwf.edu

Consistency versus Responsiveness:

Do Members of Congress Change

Positions on Specific Issues in

Response to Their Districts?

Adam F. Cayton1

Abstract

While democratic theory suggests that representatives should be willing to adjust their issue positions to adapt to

new circumstances, politicians face serious political risks from “flip flopping.” How do members of Congress balance

these risks? Using an original data set of district economic conditions and opinion from 2007 to 2010 and sets of

repeated roll call votes, I leverage the exogenous shock of the Great Recession to explain position change on three

major economic policies. I find that position change occurs in response to the constituency on final passage votes, but

that partisan pressures exert greater influence, especially on procedural votes. This novel test of responsiveness has

implications for the nature of policy representation and the mechanisms behind aggregate responsiveness.

Keywords

representation, legislative responsiveness, roll call voting, economic policy, U.S. House

4 Political Research Quarterly 70(1)

literature on the electoral connection, but at present, only

an alternative model of responsiveness on specific posi-

tions has been tested. Characteristics of the district may

also cause lawmakers from some contexts to be more

responsive than others. Pressure from constituents is not

the only kind lawmakers face. On votes for which party

leaders want a unified caucus, lawmakers should be more

likely to take the party position, possibly at the expense

of responsiveness to the district.

The period between 2007 and 2010 is useful for exam-

ining legislative responsiveness to economic changes for

two reasons. First, the housing collapse, financial crisis,

and recession presented an abrupt and severe exogenous

shock to the country, the effects of which varied by con-

gressional district and were uncorrelated with public par-

tisanship, representative ideology, and other political

variables. These conditions facilitate more reliable identi-

fication of constituency effects. Second, this period coin-

cided with the creation of several large data collection

projects that make it possible to construct annual mea-

sures of economic conditions by congressional district.

Using these sources, I examine lawmakers’ individual

responses to changing conditions in their districts and

compare the effects of economic changes in different

political contexts. The analysis shows that members of

Congress often respond to shocks to their constituency by

changing issue positions but are much more likely to do

so when the issue is simple, affects voters directly, and is

less subject to pressure from party leaders. Responsiveness

is confined to votes on final passage rather than proce-

dure, in which partisanship prevails, placing an important

limitation on the importance of responsive position

change. Lawmakers are also likely more responsive when

the party’s initial position is incongruent with problem

salience. These findings may represent the high end of

roll call responsiveness due to the salience of the eco-

nomic crisis. However, salience may increase public

scrutiny in this context. The findings not only contribute

to answering an important question about representation

but also shed light on how the U.S. House responds to an

unfolding economic crisis.

Background

Measuring legislative responsiveness to constituents has

(with good reason) been a central focus of the representa-

tion literature, but most of the work in this area is limited

by the difficulty of measuring changes in legislative

behavior and constituency characteristics over time.

Scholars of representation have accumulated mountains

of evidence that lawmakers vote consistently with the

preferences of constituents, with some variation accord-

ing to issue domain, salience, and district-level factors

such as competitiveness and complexity (Ansolabehere,

Snyder, and Stewart 2001; Bailey and Brady 1998;

Burden 2004; Erikson 1978; Griffin 2006; Hurley and

Hill 2003; Jewell 1983; Miller and Stokes 1963; Page

et al. 1984; see Burstein 2003 and Shapiro 2011, for thor-

ough reviews, and Jacobs and Shapiro 2000, for a dissent-

ing interpretation of the evidence). Voters also appear to

reward and punish their representatives according to the

congruence of their legislative behavior with public opin-

ion (Ansolabehere and Jones 2010; Canes-Wrone, Brady,

and Cogan 2002; Figlio 2000; Stratmann 2000).

This evidence of congruence between public prefer-

ences (or some demographic proxy) and roll call patterns

is usually taken as evidence that lawmakers respond to

their constituents, but the data used are often cross-

sectional, and, thus, only able to show a correlation with-

out revealing the mechanism involved. As Kuklinski and

Stanga (1979, 1091) argue,

[T]he use of data collected at a single point in time . . .

precludes separating agreement resulting from simple elite-

mass sharing of policy attitudes and that due to the actual

response of officeholders. In short, the causal link . . . is

remote at best.

Responsiveness implies change over time, and a respon-

sive legislator will change behaviors following a change

in the wishes or material needs of constituents. Most

dyadic studies of position taking on roll call votes study

congruence and are unable to estimate the extent to which

it is caused by lawmakers changing their positions fol-

lowing changes in their constituencies or already holding

the positions that a plurality of voters support (Stone

1980, 1982 also addresses the conceptual and empirical

problem of separating congruence from responsiveness).

Numerous studies of other legislative behaviors do

explicitly theorize about and observe responsiveness

over time, including work on issue priorities (Sulkin

2005), communication with constituents (Fenno 1978),

aggregate roll call ideology (Glazer and Robbins 1985;

Kousser, Lewis, and Masket 2007; Poole 2007; Stone

1980, 1982; Stratmann 2000), and even issue-specific

scores (Hayes, Hibbing, and Sulkin 2010). Glazer and

Robbins’s (1985) analysis of change in roll call ideol-

ogy following redistricting and Hayes, Hibbing, and

Sulkin’s (2010) analysis of issue-specific interest group

scores and individual legislative priorities are particu-

larly relevant. Both studies measure responsiveness on

positions, but the measures are scores constructed from

aggregated roll call votes rather than specific positions

on individual bills. This study contributes to this litera-

ture by directly measuring “flip flopping” on specific

policy questions and testing responsiveness using an

exogenous shock that is uncorrelated with other social

and political characteristics.