Despite limited uptake, index insurance is often seen as one of the most remarkable innovations of the past decades to help smallholder farmers manage risks and boost their productive investments. In this paper, we use a Bayesian hierarchical model to aggregate evidence from eight experiments and assess the external validity of their results. We find that interventions expanding access to index insurance increase productive investments by 0.06–0.11 SD on average (equivalent to 8%–16% of the control mean). Heterogeneity in treatment effects across studies is high but imprecisely estimated, and controlling for basic, pre-registered household characteristics has no impact on these estimates. Effects on those who take up index insurance are larger but even more heterogeneous. We conclude that the existing evidence base offers limited insights to predict effects in new settings, and that policymakers should be cautious before investing in the widespread promotion of index insurance.