Adv Behav Econ pdf

205
T I M E D I S C O U N T I N G
at a particular time than when they compare two outcomes occurring at different
times, or specify the value of delaying or accelerating an outcome. In one study,
for example, two groups of students were asked how much they would pay for a
$100 gift certificate at the restaurant of their choice. One group was told that the
gift certificate was valid immediately. The other was told it could be used begin-
ning six months from now. There was no significant difference in the valuation of
the two certificates 
between 
the two groups, which implies negligible discount-
ing. Yet when asked how much they would pay (have to be paid) to use it 6
months earlier (later), the timing became important—the delay group was willing
to pay $10 to expedite receipt of the delayed certificate, while the immediate
group demanded $23 to delay the receipt of a certificate they expected to be able
to use immediately.
38
Another important design choice in experimental studies is whether to use real
or hypothetical rewards. The use of real rewards is generally desirable for obvious
reasons, but hypothetical rewards actually have some advantages in this domain.
In studies involving hypothetical rewards, respondents can be presented with a
wide range of reward amounts, including losses and large gains, both of which are
generally infeasible in studies involving real outcomes. The disadvantage of hy-
pothetical choice data is the uncertainty about whether people are motivated to, or
capable of, accurately predicting what they would do if outcomes were real.
To our knowledge, only two studies have compared discounting between real
and hypothetical rewards. Kirby and Marakovic (1995) asked subjects to state the
immediate amount that would make them indifferent to some fixed delayed
amount (delayed reward sizes were $14.75, $17.25, $21, $24.50, $28.50; delays
were 3, 7, 13, 17, 23, and 29 days). One group of subjects answered all 30 permuta-
tions for real rewards, and another group of subjects answered all 30 permutations
for hypothetical rewards. Discount rates were 
lower 
for hypothetical rewards.
39
Coller and Williams (1999) asked subjects to choose between $500 payable in 
1 month and $500
1
$
x
payable in 3 months, where $
x 
was varied from $1.67 to
$90.94 across 15 different choices. In one condition, all choices were hypothetical;
in 5 other conditions, one person was randomly chosen to receive her preferred
outcome for 1 of her 15 choices. The raw data suggest again that discount rates
were considerably lower in the hypothetical condition, although they suggest that
this conclusion is not supported after controlling for censored data, demographic
differences, and heteroskedasticity (across demographic differences and across
38
Rating tasks (and probably pricing tasks as well) are subject to anchoring effects. Shelley and
Omer (1996), Stevenson (1992), and others have found that a given delay (for example, 6 months)
produces greater time discounting when it is considered alongside shorter delays (for example, 1
month) than when it is considered alongside longer delays (for example, 3 years).
39
The two results were not strictly comparable, however, because they used a different procedure
for the real rewards than for the hypothetical rewards. An auction procedure was used for the real-re-
wards group only. Subjects were told that whoever, of three subjects, stated the lowest immediate
amount would receive the immediate amount, and the other two subjects would receive the delayed
amount. Optimal behavior in such a situation involves overbidding. Since this creates a downward
bias in discount rates for the real-rewards group, however, it does not explain away the finding that
real discount rates were higher than hypothetical discount rates.

---