Incentives for Performance
Control over preferences is the most significant element distinguishing laboratory experiments
from other methods of inquiry. It is of paramount importance that one be able to state that
between two experiments, individual values either do or do not differ in any significant way.
This control can be exercised by using a reward structure and a property right system to induce
prescribed monetary value on (abstract) outcomes.
Outcomes must be measured in order to evaluate the performance of the system. Messages must be
measured in order to identify agent behavioral modes. To accomplish these objectives, we will
draw from Smith's precepts of experimental economics (1982).
When faced with a costless choice between two alternatives, each identical except that one
yields more of the reward medium than the other, the alternative that yields more of the reward
medium will be chosen. In other words, people prefer more of the reward medium to less, ceteras
The institution translates the messages into outcomes and these outcomes allow an individual
to claim a reward which increases (decreases) in the good (bad) outcomes. In
other words, the amount of the reward medium allocated to an individual is tied directly to the
individual/group performance. Not all rewards are salient. Many experimentalists pay subjects
$3.00 "up front" for agreeing to participate and for arriving on time to the experiment.
A second payment, based on the subject's earnings (tied to their experimental outcomes) is made
when the subject leaves the laboratory. This second payment is salient, the first
It is sometimes said that the use of currency (as the reward medium) to induce value on abstract
outcomes in a laboratory experiment may be an artificial procedure peculiar to experimental
methodology and is not the same thing as having "real preferences". Those who raise this question
seem not to realize that all economic systems produce forms of intangible property on which value
is induced by specifying the rights of the holder to claim money or goods. All financial
instruments, including shares, warrant, etc., have value induced upon the instruments by the
bundle of rights they convey. An important part of the property right rules of any institution
is the specification of the conditions under which intangible goods can be redeemed in terms of
other intangibles or commodities. Obviously the reward medium may make a difference, but this is
easily studied as a treatment variable by anyone who is haunted by the thought that it is
important. To argue that preferences based on cash-induced value are somehow different that
home-grown preferences over commodities is also to argue that preferences among intangible
instruments in the field are are also somehow different than commodity preferences.
The reward structure dominates any subjective costs (or values) associated with participating
in the activities of the experiment. In other words, if a subject has many motivations while
making a decision during the experiment, the amount of reward should be high enough to to exclude
all other motivations from entering into the decision. An early path-breaking experimental
study of the binary choice game by Siegal (1961) systematically varied reward level. The results
showed an increase in the proportion of reward maximizing choices when the reward level was
increased for a constant task complexity. Further, when the task complexity was increased,
holding reward level constant, the proportion of reward maximizing choices was reduced. Note that
the level of the reward in order get dominance depends on the subject population; investment
bankers would probably need a much higher level of payments than starving undergraduate students.
Each subject in an experiment is given information only on their own payoff alternatives.
This makes sense in that in the naturally occurring economy, only the person making the decision
knows what that decision is worth to them.
These precepts allow us to study laboratory environments where real agents exchange real messages
through real property rights institutions that yield outcomes redeemable in real money.
Smith, V., “Bidding and Auctioning Institutions: Experimental Results”, Bidding and Auctioning for Procurement and Allocation, University Press, New York: New York, 1976, pp. 43-64.
Smith, V., “Microeconomic systems as an Experimental Science”, American Economic Review, 72:923-955, 1982.
Smith, V., “Economics in the Laboratory”, Journal of Economic Perspectives, 8(1):113-131, 1994.