8
competing in the election, and involves the decision to target a particular group (gi),
whereas the second one takes place between parties and voters.
Consider two parties, A and B, both facing the decision of mobilizing (or not)
support from a particular group. As suggested by Geer and Shere (1992), the interplay
between Party A and Party B can be thought of as a prisoner’s dilemma. Parties can
choose to cooperate among themselves by making a commitment not to target a particular
group of voters, for instance low income voters. Or they can choose not to cooperate and
make an effort to mobilize them for their platform. In principle, both parties would be
better off cooperating – that is, not to target all segments of voters. By ignoring a certain
group (g1) of the electorate, they save mobilization resources that can be devoted to other
groups (gi≠1) where the rate of return – the probability that the mobilization effort will
bear fruit – is likely to be higher. As in the standard prisoner’s dilemma, provided that
party A cooperates, B is tempted to defect and maximize its support among the members
of g1. Symmetrically, A faces a similar temptation. The expected outcome, as indicated
by Geer and Shere, is that both parties defect, thus incurring large mobilization efforts to
gain the support of g1.
In turn, this interaction among parties speaks to a second game, namely one that is
played between voters and parties. Members of g1 can cooperate with part A or B by
showing up at the poll and casting their votes, or they can defect by staying at home. In
turn, parties can cooperate with voters by designing policy platforms specifically targeted
to benefit g(1), or they can defect by designing policy platforms that benefit other groups
in society, g(i≠1). As in the standard prisoner’s dilemma, cooperation makes both actors
better off: parties gain electoral support, whereas g(1) voters would obtain policies
Convention