Intertemporal Tradeoffs for Gains and Losses: An Experimental Measurement of Discounted Utility

This paper is the first to measure utility in intertemporal choice and presents new and more robust evidence on the discounting of money outcomes. Our measurement method is parameterfree in the sense that it requires no assumptions about utility or discounting. We found that intertemporal utility was concave for gains and convex for losses, consistent with a hypothesis put forward by Loewenstein and Prelec (1992). Utility in intertemporal choice was close to utility in decision under risk and uncertainty, suggesting that there may be one unifying concept of utility that applies to all of economics. The existence of one concept of utility is important for applied economics, because it largely reduces data requirements. Discount rates declined over time, but less so than has been observed in previous studies that assumed linear utility. Of the main discounted utility models, Loewenstein and Prelec’s (1992) generalized hyperbolic discounting model best fitted our data. The widely-used quasi-hyperbolic model fitted the data only slightly better than constant discounting. Finally, we obtained evidence of an asymmetry in discounting between gains and losses, which, in contrast with earlier findings, cannot be explained by a framing effect.