Selection and Behavioral Responses of Health Insurance Subsidies in the Long Run: Evidence from a Field Experiment in Ghana

We conduct a randomized experiment that varies one-time health insurance subsidy amounts (partial and full) in Ghana to study the impacts of subsidies on insurance enrollment and health care utilization. We find that both partial and full subsidies promote insurance enrollment in the long run, even after the subsidies expired. Although the long run enrollment rate and selective enrollment do not differ by subsidy level, long run health care utilization increased only for the partial subsidy group. We show that this can plausibly be explained by stronger learning-through-experience behavior in the partial than in the full subsidy group.