Information and Strategic Voting: Coordination
in a Small World
Emily Clough
The relationship that information plays in shaping political
decisions and outcomes has long been an area of interest to political scientists.
Using a game theoretic model of strategic voting, Cox (1997) suggests one
way that information may have a profound impact on voters’ ability to coordinate
on M+1 parties in a M-seat district: polls, or some other source of public
information about the relative standing of the candidates, must be available
in order to help voters understand which candidates are likely to win. While
there are a number of situations in which these conditions are filled, it
is possible to think of a number in which they are not: historically, for
instance, or in developing countries. What happens in situations in which
these conditions are not filled? Can voters coordinate effectively? How does
this affect the shape of the party system? One way of looking at this problem
formally is to create a model using an agent-based approach; this approach
allows the modeler to specify informational conditions with greater
flexibility than a game theoretic model. This paper presents the results
from an agent-based model in which voters gather information about the voting
tendencies of their neighbors in a network. The network is a “small world”
network, i.e. it is highly clustered and has a short characteristic path
length (Watts 1999); in other words, social networks are clustered, but there
are a few key agents who provide connections across these clusters, making
the length of the path between any two agents fairly short. Empirically,
we could think of this as a situation in which there is a great deal of segregation
of communities, but a few people who are still able and willing to communicate
between these communities; it is possible to think of a number of concrete
instances where this is the case. Watts (1999) demonstrated that employing
these networks can have a profound impact on the outcomes of models and provide
a better understanding of how information is shared and spread; however,
political scientists have not
really begun to integrate these networks into their models. Using a small
world network to model coordination through strategic voting gives us a better
sense of how coordination may occur when polls are not available.