107
N O N E X P E C T E D - U T I L I T Y T H E O R Y
both completeness and transitivity. Completeness entails that for all
q
,
r
:
either q
f
r or r
f
q or both
where
f
represents the relation “is (weakly) preferred to.”
Transitivity requires that for all
q
,
r
,
s
:
if q
f
r and r
f
s
,
then q
f
s
. Continuity
requires that for all prospects
q
,
r
,
s
where
q
f
r and r
f
s
,
there exists some p
such that
(
q
,
p
;
s
, 1
2
p
)
|
r
, where
|
represents the relation of indifference and
(
q
,
p
;
s
, 1
2
p
) represents a (compound) prospect that results in
q
with probability
p
;
s
with probability 1
2
p
. Together the axioms of ordering and continuity imply
that preferences over prospects can be represented by a function
V
(
?
) which as-
signs a real-valued index to each prospect. The function
V
(
?
) is a representation of
preference in the sense that
V
(
q
) •
V
(
r
)
⇔
q
f
r
: that is, an individual will choose
the prospect
q
over the prospect
r
if, and only if, the value assigned to
q
by
V
(
?
) is
no less than that assigned to
r
.
To assume the existence of some such preference function has seemed, to many
economists, the natural starting point for any economic theory of choice; it
amounts to assuming that agents have well-defined preferences, while imposing
minimal restriction on the precise form of those preferences. For those who en-
dorse such an approach, the natural questions center around what further restric-
tions can be placed on
V
(
?
)? The independence axiom of EU places quite strong
restrictions on the precise form of preferences: it is this axiom which gives the
standard theory most of its empirical content (and it is the axiom that most alter-
natives to EU will relax). Independence requires that for all prospects
q
,
r
,
s
,
if
q
f
r then
(
q
,
p
;
s
, 1
2
p
)
f
(
r
,
p
;
s
, 1
2
p
), for all
p
. If all three axioms hold,
preferences can be represented by
V
(
q
)
5
•
p
i
u
(
x
i
)
(1)
where
q
is any prospect, and
u
(
?
) is a “utility” function defined on the set of
consequences.
The concept of risk is pervasive in economics, so economists naturally need a
theory of individual decision making under risk. EU has much to recommend it-
self in this capacity. The theory has a degree of intuitive appeal. It seems almost
trivially obvious that any satisfactory theory of decision making under risk will
necessarily take account of both the consequences of choices and their associated
probabilities. These are, by definition, the dimensions relevant in the domain of
risk. EU provides one very simple way of combining probabilities and conse-
quences into a single “measure of value,” which has a number of appealing prop-
erties. One such property is monotonicity, which can be defined as follows: Let
x
1
, . . . ,
x
n
be consequences ordered from worst (
x
1
) to best (
x
n
). We may say that
one prospect
q
5
(
p
q
1
, . . . ,
p
qn
) first-order stochastically dominates another
prospect
r
5
(
p
r
1
, . . . ,
p
rn
) if for all
i
5
1, . . . ,
n
,
(2)
n
n
•
p
qj
$
•
p
rj
j = i
j = i
Do'stlaringiz bilan baham: |