The Concavity of Time Discount Function: An Experimental Study

This paper presents a new anomaly in the literature of time preference, which is future bias. The experiment with 37 subjects also provides the evidence in favor of an inverse S-shaped time discount function rather than a hyperbolic (and convex) time discount function. The time discount function seems to be locally concave. If time discount function is concave around t, then the decision-maker exhibits future bias. Future bias has seldom been reported in the literature, since it is not easily identified. To detect future bias, this experiment lets subjects compare two future reward options whose interval is shorter than 4 weeks. The new elicitation method is also independent of the confounding factors that are common in the existing experiments. Specifically, the method does not put any assumption on the form of utility function and does not assume that time and reward are separable either. This is a great advantage that enable us to estimate time preference more precisely