Individual discount rates: A meta-analysis of the experimental evidence

Abstract

We examine 434 estimates of the individual discount rate reported in 27 published studies. The estimates vary substantially across studies with reported mean at the value of 0.4. We detect presence of selective reporting in the discounting literature using a meta-analytical methods. Our results suggest that relevant discounting literature overestimates the discount rate approximately twofold. We apply Bayesian model averaging to explain heterogeneity in the estimates. Discount rate estimates are influenced mainly by the experimental design used for its elicitation. We confirm domain independence and the effect of time horizon presented within the discounting task. Our results support the external validity of experiments conducted on student samples and the utilisation of hypothetical rewards in experiments.