240
O ’ D O N O G H U E A N D R A B I N
to period 3, she waits in period 1 incorrectly believing she will do it in period 2.
Unfortunately, in period 2 she prefers waiting until period 3.
21
Proposition
5
has important implications for the literature on smoking guns.
First, proposition 5 implies that smoking guns need not involve the use of external
commitment devices. Even simple behaviors can
sometimes represent smoking
guns. Furthermore, the literature on external commitment devices, provides smok-
ing guns for sophisticates but not for naifs, since naifs would not pay to limit future
choice sets. Proposition 5 implies that smoking guns exist for naifs as well. Finally,
the intuitions above (and in the proof) suggest ways to design experiments attempt-
ing
to find smoking guns, as well as the types of real
world situations without
external commitment devices where smoking guns might be found.
6.
Multitasking
We now begin to explore how our results might carry over to more general set-
tings. Consider a simple extension of our model where the activity must be per-
formed more than once. The basic structure of the model is exactly as in section 2,
but now the person must do the activity exactly
M
$
1 times, and she can do it at
most once in any given period. We let
t
9
(
M
) denote the period in which a person
completes
the activity for the
i
h
time,
and define
Q
(
M
)
;
{
t
1
(
M
),
t
2
(
M
), . . . ,
t
M
(
M
)}. For each period
t
in which the person does it, she receives reward
u
t
and
incurs cost
c
t
, and these can be experienced immediately or with some delay. Us-
ing the interpretations of immediate costs and immediate rewards from section 2,
preferences take the following form.
1.
Immediate Costs
. Given
Q
(
M
), the set of
periods in which she does it, a
person’s intertemporal utility in period
t
is given by equation (1):
(1)
2.
Immediate Rewards
. Given
Q
(
M
), the set of periods in which she does it, a
person’s intertemporal utility in period
t
is given by equation (2):
(2)
U
M
c
t
M
c
t
M
t
t
t
M
M
t
M
M
( (
))
(
)
(
)
(
).
(
)
(
)
(
)
(
)
Θ
Θ
Θ
Θ
Θ
Θ
Θ
≡
−
+
−
∈
−
∉
∈
∈
∈
∈
∑
∑
∑
∑
1
β υ
β
υ
β
υ
τ
τ
τ
τ
τ
τ
if
if
U
M
c
c
t
M
c
t
M
t
t
t
M
M
t
M
M
( (
))
(
)
(
)
(
).
(
)
(
)
(
)
(
)
Θ
Θ
Θ
Θ
Θ
Θ
Θ
≡
− −
+
−
∈
−
∉
∈
∈
∈
∈
∑
∑
∑
∑
1
β
β
υ
β
υ
τ
τ
τ
τ
τ
if
if
21
The proof of proposition 5 essentially involves generalizing these examples for all values of
b
and
d
.