Targeted Health Warnings Amidst Extreme Inequality: Evidence from a Regression Discontinuity Design in South Africa

Abstract

We make use of a quasi-experimental design and panel survey data of South African adults to, in a context of extreme inequality, estimate and examine the impact of a blood pressure-related health warning on objective and subjective health and behavioural outcomes. We additionally analyse effect heterogeneity by awareness and insurance status. While we find that the warning reduced the probabilities of reporting good self-assessed health and frequently drinking alcohol both by 4 percentage points, we observe notable heterogeneity by health insurance coverage but not awareness status. Among the insured, we estimate a significant negative effect of 13.9mmHg on systolic blood pressure. Our findings suggest that targeted health warnings may have the potential to improve health outcomes even in a context of extreme inequality, however the effectiveness of such interventions may depend on access to quality health services and resources.

Keywords: South Africa, blood pressure, hypertension, health warning, regression discontinuity design, health economics

JEL Classification: I10, 12, I13