Eliciting utility curvature and time preference

In both standard and behavioral theory, as well as experimental procedures to elicit time preference, it is commonly assumed that a single utility function is used to evaluate payoffs both under risk and over time. I introduce a novel experimental design to examine this assumption, by transposing the well-known Holt-Laury risk preference experiment from state-payoff space into time-dated payoff space. I find that the curvature of utility elicited directly from choices over time is significantly concave, but far closer to linear than utility elicited under risk. As a result, the effect of correcting discount rates for this curvature is modest.