Time discounting for primary and monetary rewards

There is considerable evidence indicating individuals are highly impatient. In particular, they exhibit higher discount rates in the short run than in the long run (Frederick et al., 2002). This result has significant implications for how we model intertemporal decisionmaking (Strotz, 1955–56). In most of the experiments, short-term discount rates are measured using monetary rewards delivered at different moments of time (e.g. Thaler, 1981; Kirby, 1997). Unfortunately, since money is fungible, it might be inadequate for this purpose. In particular, access to credit decouples money from consumption, which poses difficulties for how to interpret intertemporal choices (Cubitt and Read, 2007). An alternative is to elicit discount rates using primary goods, which are less fungible. However, goods have important disadvantages: first, immediate consumptionof a goodeasily leads to satiation; second, comparisonsbetween individuals are more complicated as their desire for a good might vary; third, goods introduce additional uncertainty with respect to the future (e.g., I might not know how much I will want a piece of chocolate next Monday afternoon); and fourth, some goods are not easily divisible in small-enough units to provide accurate discount-rate measures. In this paper, we test whether discount rates elicited with these two different approaches are correlated. This exercise is important for two main reasons. First, a significantly positive correlation between the two suggests that we are indeed measuring an underlying characteristic of individuals, which gives support to the use of models of present-biased preferences such as the quasi-hyperbolic discounting model (Laibson, 1997) and dual-system models (Bernheim and Rangel, 2004; Fudenberg and Levine, 2006). Second, the existence or not of this correlation can enlighten us on the ultimate source of this present-bias: whether it comes from an impatience for the utility coming from consumption or whether it is an impatience toward instant gratification, even when the source of that gratification is not instant consumption (as in the case of the monetary reward). In spite of money being fungible, there are reasons to think that discount rates elicited with money and chocolate are correlated. In particular, neurological data suggests that individuals derive utility from receiving money irrespective of whether it is immediately used or not. For instance, the same limbic areas are activated when individuals make intertemporal choices with monetary rewards (McClure et al., 2004) and with an immediate-consumption good (McClure et al., 2007). If money is seen as a carrier of reward that