One of the central features of a democratic regime is the accountability of elected decision makers to their
electorate. In the literature, accountability is seen to be based either on sanctions or on justifiability. Fearon
(1999, 55) defines accountability in the first sense by stating that Y is accountable to X if Y is obliged to act on
behalf of X and X is empowered to sanction or reward Y. However, accountability can also be understood as a
relationship based on decision-makers’ obligation to justify their decisions (March and Olsen 1995; Philp 2009).
We compare these two forms of accountability in a decision making experiment with student subjects. In the
experiment, subjects engage in a three player trust game (Berg et al 1995) where two players act a first movers
and the third player is a responder. The first movers decide how much they send to the responder of their
initial endowments. The amount sent is tripled, and the responder then decides how much he or she returns to
each first mover. We see the trust game as an analogy between the electorate (first movers) who invests (e.g.
votes) in a candidate (responder) who responds by contributing benefits to the electorate.
The experiment follows a 2 (opportunity to punish) x 2 (requirement of justification) factorial design.
Punishment means the first movers’ opportunity to punish the responder, whereas justification means that the
responder is required to justify her or his decision to the first movers. We ask how holding the responder
accountable to the first movers influences the amount he or she returns. Further, we study whether there is a
difference in the retuned amounts between two forms of accountability, costly punishment and justification.
We expect to see smallest amounts retuned in the baseline treatment and highest in the treatment with both
punishment and justification. Treatments with either punishment or justification are expected to fall in between these two.
The trust game has been extensively studied but we are not aware of previous studies that would examine the
influence of the responders’ need to give ex post justifications for their allocation decision.
