The Instrumental Voter Goes To the Newsagent

Published date01 July 2007
Date01 July 2007
Subject MatterArticles
Valentino Larcinese
This article studies the impact of instrumental voting on information demand
and mass media behaviour during electoral campaigns. If voters act instru-
mentally then information demand should increase with the closeness of an
election. If mass media are prof‌it-maximizing f‌irms then information supply
should be larger in electoral constituencies where the contest is expected to be
closer, delivery costs are lower and customers are on average more prof‌itable
for advertisers. The impact of the size of the electorate is theoretically unde-
termined. These conclusions are derived within a formal model of information
demand and supply, and then tested, with good results, on data from the 1997
general election in Britain.
KEY WORDS .British elections .Information acquisition .mass media
.media bias .rational ignorance
1. Introduction
There is a vast literature that explains voting behaviour using the tools of
rational choice theory. This approach to the study of elections assumes instru-
mental voting: citizens care about public policies and voting is the instrument
used to inf‌luence policy choices, or at least to increase the probability of obtain-
ing the preferred option.
I would like to thank Evelina Larcinese for her invaluable help in collecting the data and Tim
Besley, Ian Preston, Imran Rasul, Cheryl Schonhardt-Bailey, David Str
omberg and Cristiana Vitale
for useful comments and suggestions. I also benef‌ited from presenting this work at the workshop on
mass media economics at STICERD, at meetings of the Public Choice Society, the European Public
Choice Society, the European Economic Association, the Econometric Society, and in seminars at
LSE, SOAS, Royal Holloway, Pompeu Fabra and Universitat Autonoma de Barcelona. The usual
disclaimer applies.
Journal of Theoretical Politics 19(3): 249–276 Copyright Ó2007 Sage Publications
DOI: 10.1177/0951629807077569 Los Angeles, London, New Delhi and Singapore
This theory poses some problems, especially since the probability of being
pivotal in large elections is normally so low that it could be considered negligi-
ble in an optimization process. This criticism can be overcome if we are ready
to compromise on what we intend by a rational act. In a weak sense, agents
behave rationally according to their perception of the reality, which could be
different from the ‘objective’ state of facts. The probability of being pivotal in a
large election is clearly very low, but it is not zero, and the subjective perception
of the probability of casting a decisive vote may not necessarily coincide with
the inf‌initesimal numbers that appropriate but cumbersome calculations would
deliver (see for example Uhlaner and Grofman, 1986). Moreover, voting can be
seen as a ‘low cost–low benef‌it’ activity (Aldrich, 1993): it is therefore possible
that even small changes in this probability might have an effect on incentives to
participate in an election.
If we accept this argument then turnout should be larger in closer elections,
when the probability of casting the decisive vote is higher. Unfortunately,
empirical analysis does not deliver any clear-cut conclusion. Foster (1984), after
reviewing a number of studies on the closeness–turnout linkage in US elections,
concludes that ‘the perceived probability of a tied election at the state level is
not a powerful or reliable factor in explaining across-state voter participation
rates in presidential elections’. Grofman et al.’s (1998) study on US Senate and
House of Representatives elections instead f‌inds evidence of higher turnout
among registered voters in closer contests. However, other recent studies based
either on aggregate data (Kunce, 2001) or on survey data (Matsusaka and Palda,
1999) show a weak relationship between closeness and turnout. Using poll data,
Kunce (2001) also shows how ‘the extent to which pre-election perceptions
matter depends directly on how one measures the likelihood of a close contest’.
It seems fair to say that evidence is, at best, mixed.
This article will consider another implication of instrumental voting: when
elections are closer then information on candidates and platforms should be
more valuable, since the probability that a vote matters is higher. Although
Downs (1957) himself hints both at the ‘paradox of voting’ (low incentives to
vote) and at ‘rational ignorance’ (low incentives to gather political information)
as closely related consequences of instrumental voting, the second of the two
paradoxes has received less attention, in particular concerns for what the predic-
tive implications of comparative static analysis.
Thus, information acquisition should be related to the probability of being
pivotal when voting. If the suppliers of political information are aware of this,
we should expect their behaviour to be inf‌luenced by marginality. In this sense,
the behaviour of the mass media will provide a different and new test of theories
of instrumental voting and of the role of marginality as an incentive for partici-
pation in election. This allows us to exploit information no used so far for this

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT