OLD PAPERS
Anticipation in Observable BehaviorAbstract:
This paper asks the natural question: how is utility from anticipation reflected in behavior?
I consider a general model of decisionmaking where rationally formed anticipation enters the
agent’s utility function in addition to physical outcomes, and allow for interactions between
these two payoff components. The paper explores three types of behavior made possible by
utility from anticipation, and proves that if a decisionmaker who cares about anticipation is
distinguishable from one who only cares about physical outcomes, she has to exhibit at least
one of these phenomena. First, the agent can display informational preferences because she is
not indifferent to insecurity, or because she cares about future disappointments. I prove that an
agent who is indifferent to insecurity always prefers full to partial information if and only if she
is disappointment averse, but a stronger condition is needed for her to prefer more information
to less. Second, the agent can be time inconsistent because anticipatory feelings pass by the
time she has to “invest” in them, and this time inconsistency can be reflected in intransitivity of
choices. Third, the agent can be prone to self-fulfilling expectations. In developing these ideas,
I also deal with several modeling difficulties when expectations enter
utility.
An older, more general, and significantly messier version of "Utility
from Anticipation and Personal Equilibrium," Economic Theory.
Ego Utility and Information AcquisitionAbstract:
Based on extensive psychological evidence and the experience of most of us, it seems
obvious that people intrinsically care about the perceptions of themselves, not only because it helps in making decisions. This paper explores some of the consequences of
this motivation in a model where the agent derives utility from both financial outcomes and her beliefs about her ability (`ego utility'). The model can explain a variety
of anomalous-looking behavior, like refusing to consider new information about past
judgments, putting off making judgments, and holding on too long to losing decisions.
Several applications are discussed, with particular attention to how to provide incentives to employees with ego utility.
Revised January 2001.
Emotional Agency: The Case of the Doctor-Patient
RelationshipAbstract:
This paper identifies an array of complications in doctor-patient communication that arise
when the patient suffers from anxiety. I assume that the patient derives utility from health
outcomes as well as the anticipation of the exact same outcomes, and that the doctor wishes to
maximize the patient’s utility. The doctor privately (and probabilistically) observes a diagnosis,
which affects the optimal treatment. She then sends a message to the patient, who chooses
a treatment. While formulated in terms of medical care, this formal model of “emotional
agency” also applies to many other situations. If the doctor cannot certifiably convey her
diagnosis, communication is endogenously limited to a treatment recommendation, which the
doctor distorts toward the treatment that is optimal when the patient is relatively healthy.
Paradoxically, more emotional patients get less useful recommendations, even though basing
their treatment choice on better recommendations would make them less anxious. If the doctor
can certifiably convey her diagnosis, she does so for good news, but unless she needs to “shock”
the patient into behaving correctly, she pretends not to know what is going on when the news
is bad. If the patient visits two doctors in a row, the second doctor reveals more information
than the first one. Results from an original survey of practicing physicians confirm that doctors
care about patients’ emotions, and alter their recommendations and other communication in
response to them.
An earlier version of "Emotional Agency" (QJE 2006), focused on the doctor-patient relationship
Revised January 2004.
