115
N O N E X P E C T E D - U T I L I T Y T H E O R Y
an individual must either prefer the “safer option” in
both
choices
or
the “riskier
option” in
both
choices, yet many people choose
s
1
** over
r
1
** and
s
2
** over
r
2
**.
This is the common ratio effect and, as in the common consequence effect, relative
to the predictions of EU, there is an “inconsistency” in the risk attitudes revealed
across their choices.
B
e
s
t
C
o
n
s
e
q
u
e
n
c
e
Worst Consequence
r
s
Indifference curve for
the more risk-averse agent
r
Figure
4.1c Different degrees of risk-aversion in EU.
r
1
**
r
2
**
s
2
**
s
1
**
p = 1
p < 1
Figure
4.2 Common ratio prospects.
116
S T A R M E R
Viewed in the context of the triangle, this inconsistency is suggestive of a sys-
tematic pattern: relative to the predictions of EU, choices between prospects located
in the bottom right-hand corner appear more risk-prone than should be expected
given preferences revealed for choices located leftward and/or upward in the tri-
angle. Any conventional theory seeking to explain these standard violations of EU
will therefore need at least one quite specific property: indifference curves deter-
mining preferences over pairs of prospects located near the right-hand corner of a
given triangle—e.g., {
s
2
**,
r
2
**}—will need to be relatively flat (reflecting more
risk-prone behavior), compared with indifference curves determining choices
over pairs of prospects, like {
s
1
**,
r
1
**}, near to the left-hand edge of the triangle.
All of the proposed conventional alternatives to EU are able to generate this prop-
erty, though they do so in a variety of ways.
4.1.1. THE “FANNING-OUT” HYPOTHESIS
Having observed this apparent connection among different violations of inde-
pendence, Mark Machina (1982) proposed an analytical extension of EU (termed
“generalized expected utility analysis”), along with a specific hypothesis on the
shape of nonexpected utility indifference curves. Analytically, he noted that under
expected utility, where
V
(
q
)
5
íU
(
x
i
)
p
i
, the utility values
U
(
x
i
)
5
•
V
(
q
)/ •
p
i
are
the probability derivatives of
V
(
?
). He then showed that standard expected utility
results (e.g., risk aversion
5
concavity of
U
(
?
)) also hold for the probability de-
rivatives
U
(
x
i
;
q
)
5
•
V
(
q
)/ •
p
i
of smooth nonexpected utility preference functions
V
(
?
), so that
U
(
?
;
q
) can be thought of as the “local utility function” of
V
(
?
) about
q
. For example, the property “concavity of
U
(
?
;
q
) at every
q
” is equivalent to
global risk aversion of
V
(
?
).
Given the existence of phenomena like the common ratio and common conse-
quence effects, Machina hypothesized that the local utility functions
U
(
?
;
q
) be-
come more concave as we move from (first order) stochastically dominated to
stochastically dominating distributions. Loosely speaking, this essentially empir-
ical assumption (which Machina calls “Hypothesis II”) implies a tendency for
agents to become more risk averse as the prospects they face get better; in the
context of the triangle, it means that indifference curves become steeper, or “fan
out,” as we move northwest. Figure 4.3 illustrates the general pattern of indiffer-
ence curves implied by Hypothesis II. Notice that they are drawn as wavy lines:
generalized expected utility theory requires indifference curves to be smooth but
does not imply that they must be linear (though they may be). It is very easy to see
that this fanning-out property generates implications consistent with the common
consequence and common ratio effects. Since indifference curves are relatively
steeply sloped in the neighborhood of prospect
m
,
m
lies on a higher indifference
curve than
q
or
r
. Flatter indifference curves in the bottom right-hand corner of
the triangle are such that
t
lies on a higher indifference curve than
s
. Hence, for an
individual whose indifference curves fan out we can construct prospects over
which we will observe a common consequence effect (e.g.,
m
s
q
and
t
s
s
) and
a common ratio effect (e.g.,
m
s
r
and
t
s
s
).
Do'stlaringiz bilan baham: |