Random Discounting and Assessment of Intertemporal Projects: a Non-expected Utility Approach

This paper assumes each individual in society has a random discount factor and assesses an intertemporal project using rank-dependent expected utility theory. We consider both the ex ante and the ex post approaches. For the former, we show the social planner's discount factor is a convex combination of those of the individuals under the standard Pareto condition. For the latter, we propose a method for determining the social planner's discount factor distribution from the individuals' distributions, which are possibly heterogenous. We demonstrate that relative to expected utility, overweighing of small probabilities can substantially accelerate the decline of the social discount rate.