Review of Accounting Studies, 5, 351–375 (2000)
2000 Kluwer Academic Publishers, Boston. Manufactured in The Netherlands.
Reputation Without Repeated Interaction: A Role
for Public Disclosures
STEVEN T. SCHWARTZ
RICHARD A. YOUNG
The Ohio State University
The Ohio State University
Abstract. This paper conducts an experiment to investigate the economic effect of public disclosures within a
multi-move adaptation of the Prisoner’s Dilemma game. The game, which has multiple equilibria, is characterized
by: (1) a stochastic endpoint, (2) random, repeated pairings with anonymous partners, and (3) public disclosures
concerning the current partner’s previous strategies. In the experiment,cooperation is improved by the disclosures.
In addition, subjects cooperate more frequently when encountering a player who has tended to cooperate in the
past, and less frequently when encountering a player who has tended to defect in the past. Delayed disclosure
leads to levels of cooperation only slightly less than those obtained with timely disclosure.
Keywords: experimental economics, Prisoner’s Dilemma, disclosure, reputation
This paper reports on an experiment designed to investigate the economic impact of public
disclosures. The setting is a modiﬁed repeated Prisoner’s Dilemma in which subjects are
randomly assigned a new, anonymous partner at the start of each round. We demonstrate
that public disclosure of past choices increases the frequency of cooperation which, as
implied by the Prisoner’s Dilemma setting, improves efﬁciency. We further demonstrate
that even untimely disclosures can be effective in inducing cooperation among agents.
The accounting and economics literatures have identiﬁed settings where increases in ef-
ﬁciency are achieved through coordination. First, agents may have the ability to credibly
commit to their future behavior, perhaps by signing contracts that are (somehow) enforce-
able. Second, agents may be in a situation where they transact with each other directly and
repeatedly, so that future decisions may be made implicitly conditional on past behavior.
While efﬁciency gains might be achievable through contracting or repeated interaction,
there are many instances where contracting is costly or illegal, and where transactions are
not repeated among the same set of agents. The contribution of this paper is it demonstrates
that, in the absence of contracting or repeated transactions, public disclosures can be useful
for increasing efﬁciency.
The basic structure of our experiment is a modiﬁed Prisoner’s Dilemma game with ran-
domly paired subjects and a stochastic endpoint. There are three experimental treatments
that differ only by the type of disclosure. In the ﬁrst treatment, no public disclosures are
made (no disclosure). In the second, a player receives the history of how his or her partner