Four score and seven years from now: the “date/delay effect” in temporal discounting

We describe a new anomaly in intertemporal choice: the “date/delay effect:” Future outcomes are discounted more when time is referred to as a delay (e.g., ‘in 6 months’) than as a calendar date (e.g. on October 17th). The effect is demonstrated in four experiments, using both choice and matching response modes. Moreover, hyperbolic discounting is found only when time is referred to as a delay (discount rates are constant when time is referred to in terms of calendar dates). We conclude by suggesting that Rubinstein’s (2003) ‘similarity’ hypothesis as a potential explanation, and then consider some of its practical implications.