Social Identity, Political Speech, and Electoral Competition
| Date | 01 January 2006 |
| Author | Eric S. Dickson,Kenneth Scheve |
| Published date | 01 January 2006 |
| DOI | 10.1177/0951629806059594 |
| Subject Matter | Articles |
SOCIAL IDENTITY, POLITICAL SPEECH, AND
ELECTORAL COMPETITION
Eric S. Dickson and Kenneth Scheve
ABSTRACT
Much research in political science suggests that social identity can be an impor-
tant factor in motivating behavior. If voters care about social identity, when
will politicians find it in their interests to make group-based appeals? Do
social identity concerns affect the policy platforms offered by candidates? In
a model of political speech and electoral competition, in which voters care
about both social identities and policy, we demonstrate that social identity
concerns can lead to platform divergence even when the policy dimension is
uncorrelated with identity. For example, policy-motivated politicians can
employ identity rhetoric to obtain ‘slack’ in the policy dimension. Further,
the need for candidates to resort to group-based appeals depends on such
factors as the relative sizes of social groups; the policy preferences of group
members; whether candidates care about policy and if so, their preferred
policies; and the extent of individual identification with groups. The analysis
demonstrates that social identity can have a striking impact on the strategic
conduct of campaigns.
KEY WORDS .campaign rhetoric .elections .ethnic politics .valence
competition
1. Introduction
Issues are important, but they are not as important as the fact that this is an opportunity
to vote for one of your own.
Tony Sanchez, Democratic candidate for governor of Texas,
September 2001.
Buri ’mbuzi aha nkoondo yaayo (‘Every goat at its stump’, meaning every Mwiru vote for
a fellow Mwiru).
Campaign slogan for Patrick Buriiku, local elections in Kampala, Uganda,
April 1998.
Journal of Theoretical Politics 18(1): 5–39 Copyright &2006 Sage Publications
DOI: 10.1177/0951629806059594 London, Thousand Oaks, CA and New Delhi
http://jtp.sagepub.com
We thank Steve Brams, Kanchan Chandra, Keith Darden, Jim Fearon, James Fowler,
Catherine Hafer, Macartan Humphreys, David Laibson, Dimitri Landa, Rebecca Morton,
Sunita Parikh, Nicholas Sambanis, Ken Shepsle, Leonard Wantchekon, Steven Wilkinson, the
editor, and seminar participants at the University of Washington meeting of LiCEP, Duke Uni-
versity, Washington University in St. Louis, the Hoover Institution, Yale University, and New
York University for helpful comments. We also thank Rafaela Dancygier for research assistance.
Why do candidates in democratic elections sometimes make campaign appeals
to voters’ social identities? When and where are such political speeches most
likely to be made? Answers to these questions are central to understanding
the dynamics of electoral competition, as well as the connection between
democracy and social conflict, in many societies. But the formal literature
on electoral competition offers few, if any, insights into either question.
Generally, candidates are assumed to compete for votes by announcing posi-
tions on the policies that voters care about or else by offering voters direct or
indirect monetary transfers. Aside from making such information known,
these accounts suggest no reason that politicians would give speeches gener-
ally or engage in group-based rhetoric in particular. In contrast, the empirical
literatures on public opinion, voting behavior, and campaigns, along with
insights from psychology and sociology, suggest that social identity can
provide an important motivation for vote choice. Thus, a simple answer to
the first question is that candidates seeking office can have an incentive to
make group-based appeals because such appeals can attract votes. This is,
however, an incomplete answer, and the existing literatures do not per-
suasively address the second question concerning the conditions under
which rhetoric will be observed. To answer both questions requires a strate-
gic analysis of electoral competition; politicians’ incentives to engage in
group rhetoric are determined not only by the social identity interests of
voters but also by the strategies of competitors and the structure of electoral
competition.
This article sheds light on both questions by introducing two innovations
to standard game-theoretic models of electoral competition. First, we specify
a utility function for voters that depends both on policy outcomes and on
identity-related payoffs. Second, we explicitly model candidates’ strategic
use of political speech in their election campaigns. Following a substantial
empirical and experimental literature, we argue that political speeches and
advertisements can affect identity payoffs by priming the salience of identity
in the context of a given election, thereby varying the strength of behavioral
prescriptions associated with membership in social groups. In our model,
competing candidates must decide not only what policies they will announce,
but also the level of group rhetoric they will employ.
These two innovations allow us to study the effects of voter social identities
on candidates’ strategic position-taking and speech-making behavior. In the
context of our model, social identity appeals by a candidate can be thought
of as a defensive tactic that serves to shore up support among members of
favorably disposed social groups. The effect of such identity appeals is that
a voter no longer responds continuously to small differences between candi-
date platforms. Rather, a voter sticks with the candidate favored by his or her
6 JOURNAL OF THEORETICAL POLITICS 18(1)
social group until the difference between platforms exceeds some threshold
magnitude that is increasing in the intensity of group-based rhetoric. That is,
using political speech, politicians are able to give themselves a valence advan-
tage with voters from favorable social groups. Such speech can help a politi-
cian keep members of her social group ‘on board’ and can in fact give a
politician a measure of ‘slack’ in choosing a policy platform. A natural
result is that it never hurts competing candidates to engage in maximal
group rhetoric, at least so long as political speech is costless. Empirically,
however, elections exhibit widely varying levels of group rhetoric. Our
model demonstrates that politicians’ level of need for the group-rallying
effect of identity-related speech, and therefore the existence of equilibria
with lower group rhetoric, depends on various factors, such as social
group size, the policy preferences of group members, whether candidates
care about policy and if so, which policies they prefer, and the extent of habi-
tual social group identification among voters. As such, the analysis suggests
under what conditions varying levels of group rhetoric are most likely to be
observed.
The article contains four additional sections. Section 2 motivates the
analysis by discussing relevant literature on identity-related voting behavior.
Section 3 presents a model of electoral competition in which candidates make
speeches that appeal to voters’ social identities, thereby potentially affecting
voting behavior. We discuss equilibria of the model and demonstrate the ways
in which various factors affect the potential for lower levels of identity-related
rhetoric. Section 4 discusses several cases from US and Caribbean politics that
illustrate the logic of the model. The final section contains a summary of the
results and a discussion of possible extensions.
2. Identity-related Behavior and Elections
Virtually all of the seminal empirical work on voting emphasizes the impor-
tance of one type of social identity or another for explaining why citizens cast
the ballots that they do (Lazarsfeld et al., 1944; Berelson et al., 1954; Camp-
bell et al., 1960; Lipset and Rokkan, 1967). The social categories commonly
thought to be important for vote choice include class, religious, linguistic,
ethnic, and partisan groups. Different contributions emphasize the impor-
tance of different social categories and there is often a recognition that in
some political environments multiple categories appear to influence indivi-
dual voting decisions. Nonetheless, there is considerable consensus about
the importance of social identities generally. The empirical foundation of
DICKSON & SCHEVE: SOCIAL IDENTITY, POLITICAL SPEECH 7
To continue reading
Request your trial