Elicitation Designs and Time Discounting: Comparison of Predictive Power on Behavior

Abstract

A growing number of empirical studies have aimed to identify dynamic inconsistency by combining diverse sets of elicitation designs and tested the extent to which this important component of individual heterogeneity predicts behavior; however, relatively little consensus on which designs are the most reliable for predicting behavior. Using a representative sample randomly assigned across elicitation designs, we gather eight measures of dynamic inconsistency by elicitation design and compare their power to predict behavior. We examine a multiple price list or a titration, whether to ask on amount or delay, and time horizons at two time periods starting either at different points in time or at the same point. We find that the elicitation design combined with (1) a titration, (2) asking subjects on amount, and (3) time horizons starting at different points in time predicts behavior well. Finally, we discuss the implications in the domain of savings and debt.