An Extension of the Quasi-Hyperbolic Discounting Model: The Cost of Keeping Track as a Predictor of Dynamic Choice Reversals

Abstract

Although there is overwhelming evidence documenting dynamic inconsistency due to present bias (i.e. static choice reversals), there is a significant and surprising gap in the experimental literature detailing other predictors of time inconsistency. In addressing this gap, Haushofer (2015) proposes a theory that predicts decreasing impatience and dynamic choice reversals for conditions in which an agent incurs “a cost of keeping track”. According to his theory, an agent incurs a cost of keeping track when he must remember to accomplish a task or future transaction and when there is a penalty for forgetting (Haushofer, 2015). Also, importantly, his theory predicts that reminders will incur the cost of keeping track for the agent and will thus make the him time consistent. The present study provides the first empirical test of Haushofer’s (2015) theory. In each of two separate calls of this study, participants chose between a smaller, sooner monetary reward and a larger, delayed reward. According to Haushofer’s (2015) theory, participants assigned to a No Reminder Group incurred a cost of keeping track, while participants in a Reminder Group did not. My hypothesis was two-fold. First, I predicted that those in the No Reminder Group would demonstrate more dynamic choice reversals in the forward direction (switching from the later reward to sooner reward) than those in the Reminder Group. And second, I predicted that there would be a significant effect of reminders on dynamic inconsistency in the forward direction. The results confirmed the first hypothesis and documented forward dynamic inconsistency for 12.5% of those in the No Reminder Group compared to 6.3% of those in the Reminder Group. And although a marginally significant effect of reminders on dynamic inconsistency in general was found, there was no significant effect for dynamic inconsistency in the forward direction; therefore, the second hypothesis was not confirmed.