An experimental test of the under-annuitization puzzle with smooth ambiguity and charitable giving

International audienceIn a life-cycle model with a bequest motive, we study the impact of smooth ambiguity aversion to uncertain survival probabilities on the optimal demand for annuities. We implement a theory-driven laboratory experiment. First, a subject's ambiguity attitude is elicited in a simple experimental setting able to make the smooth ambiguity model operational. Then, in a two-period annuity-bequest decision problem, the subject's bequest in the second period is presented as a donation to a previously chosen charity, contingent to the subject being active after the first period. In line with the theoretical predictions, we find that ambiguity-averse (resp., loving) subjects invest less (resp., more) in annuities than ambiguity-neutral ones. Furthermore, subjects' contingent donation to the chosen charity increases in their investment in annuities only for sufficiently high levels of warm-glow altruism