WP 2020-12 Do No Harm? The Welfare Consequences of Behavioural Interventions
AUTHORS: Glenn Harrison, Karlijn Morsink, and Mark Schneider
ABSTRACT: The principle of “do no harm” has long been established in the medical field in relation to proposed new drugs, procedures and devices. In that setting, we expect tests of clinical efficacy before the effectiveness of interventions are tested in the field. We propose that behavioural interventions in economics, where possible, be similarly subject to tests of efficacy before being applied in the field, to provide priors that they can be expected to be effective in the field and “do no harm.” We demonstrate one way to do this, by conducting an experiment in which we randomly assign individuals in the lab to a range of behavioural interventions that are typically used to promote index insurance in the field in low-income countries. Based on elicited risk preferences we estimate the expected individual-level welfare gains and losses from insurance decisions, and compare these across intervention arms. Although all interventions significantly increase take-up, and some increase understanding, we find no evidence that our behavioural interventions lead to an average increase in consumer welfare for all interventions.