|
|
 |
Background Information
Economics has focused primarily on the behavior of individual agents in strategic and
non-strategic environments. Little systematic work has been done to study the behavior
of groups as decision-making agents in markets and other strategic games. This may limit
the applicability of accepted economic principles because many important decisions are
made by groups (e.g., families, boards of directors, legislatures, and committees) rather
than by individuals.
Stephen C. Hayne received his first NSF grant: SBR-7909423 ($15,000), Group
Decision-Making and the Winners Curse. This grant was awarded in August 1997
and is intended for subject payments in initial baseline experiments. From this funding,
22 experiments involving 72 groups (420 subjects) in several treatment conditions have been
conducted. Dr. Hayne and Dr. Cox (University of Arizona) received two more NSF grants,
SES-9818561 (Cox, $142,000) and SES-9819031 (Hayne, $101,500) to further explore the
problems of group bidding.
Prototypes:
Work-in-Progress Results From New Theory:
Working Papers:
-
Cox, J. and Hayne, S. (1998) "Group vs. Individual Decision-Making in Strategic
Market Games," presented at the Economic Science Association in Mannhiem,
(view paper).
This paper compares group and individual economic behavior. The objective of the
research is to learn whether there are systematic differences between decisions made
by groups and individual agents in market environments characterized by risky
outcomes. A quantitative measure of deviation from minimally-rational decisions is
used to compare group and individual behavior in the same strategic environments.
Many naturally-occurring auction markets are characterized by sealed-bids for items
of uncertain value, submitted by firms in which groups have determined the amounts
to bid. Our research involves study of group decisions in the risky strategic
environment of common value auction markets. We report results from experiments
conducted to compare outcomes from individual and group bidding. If group bidding
differs from individual bidding, this could result from groups having more information
or, alternatively, better or worse judgment. This distinction is examined by crossing
an information density (signal sample size) treatment with the bidding entity (groups
or individuals) treatment.
Inspection of the data indicates that the majority of individuals and groups are very
susceptible to the winners curse when they are inexperienced. The more
interesting data are for once- and twice-experienced subjects. These data show that
both groups and individuals deviate more from rational bidding when they have more
information, and that groups are more affected by this information curse
than individuals. The data support the conclusion that, if groups inherently have more
information than individuals, then groups are less rational decision-making agents
than individuals in common value auctions.
-
Hayne, S., Cox, J. Emery, J.C.H. (1997) "GROUP BIDS AND THE WINNERS CURSE,"
(view paper).
An experiment was conducted involving group decision-making through competitive bidding
in a common value auction. Subjects were 420 undergraduate business students who worked
under a group incentive system in an auction market size
of five with groups of five persons (each with their own signal). Tie-breaking can
occur in odd sized groups (if they attempt a voting strategy) and previous research
defines groups as more than 3 persons. The expected loss from naive bidding by groups
in markets of size five was set to be only 50% of the expected loss from naive bidding
by individuals in markets of size three. Since earlier experiments had found that
bidding by experienced individuals in markets of size three produced positive average
profits and no winners curse, this experiment was biased against significant
occurrence of the winners curse. Our results show that experienced groups bid
irrationally and realized large losses
Restricted Results:
|
|