Adv Behav Econ pdf

206
F R E D E R I C K E T A L .
treatments).
40
Thus, as of yet there is no clear evidence that hypothetical rewards
are discounted differently than real rewards.
41
What Is Time Preference?
Figure 6.2 reveals spectacular disagreement among dozens of studies that all pur-
port to be measuring time preference. This lack of agreement likely reflects the
fact that the various elicitation procedures used to measure time preference con-
sistently fail to isolate time preference, and instead reflect, to varying degrees, a
blend of both pure time preference and other theoretically distinct considerations,
including: intertemporal arbitrage, when tradeable rewards are used; concave util-
ity; uncertainty that the future reward or penalty will actually obtain; inflation,
when nominal monetary amounts are used; expectations of changing utility; and
considerations of habit formation, anticipatory utility, and visceral influences.
Figure 6.2 also reveals a predominance of high implicit discount rates—
discount rates well above market interest rates. This consistent finding may also
be due to the presence of the aforementioned various extra-time-preference con-
siderations, because nearly all of these work to bias imputed discount rates 
upward—only habit formation and anticipatory-utility bias estimates downward.
If these confounding factors were adequately controlled, we suspect that many in-
tertemporal choices or judgments would imply much lower—indeed, possibly
even zero—rates of time preference.
Our discussion in this section highlights the conceptual and semantic ambigu-
ity about what the concept of time preference ought to include—about what prop-
erly counts as time preference per se and what ought to be called something else
(for further discussion see Frederick 1999). We have argued here that many of the
reasons for caring when something occurs (for example, uncertainty or utility of
anticipation) are not time preference, because they pertain to the expected amount
of utility consequences confer, and not to the weight given to the utility of differ-
ent moments (see figure 6.3 adapted from Frederick 1999). However, it is not ob-
vious where to draw the line between factors that operate through utilities and
factors that make up time preference.
Hopefully, economists will eventually achieve a consensus about what is in-
cluded in, and excluded from, the concept of time preference. Until then, drawing
attention to the ambiguity of the concept should improve the quality of discourse
40
It is hard to understand which control eliminates the differences that are apparent in the raw data.
It would seem not to be the demographic differences per se, because the hypothetical condition had a
“substantially higher proportion of non-white participants” and “non-whites on average reveal dis-
count rates that are nearly 21 percentage points higher than those revealed by whites” (Coller and
Williams 1999, pp. 121, 122).
41
There has been considerable recent debate outside of the context of intertemporal choice about
whether hypothetical choices are representative of decisions with real consequences. The general con-
clusion from this debate is that the two methods typically yield qualitatively similar results (see
Camerer and Hogarth 1999 for a recent review), though systematic differences have been observed in
some studies (Cummings, Harrison, and Rutstrom 1995; Kroll, Levy, and Rapoport 1988).

---