Health insurance in low- and middle-income countries increases financial protection and utilization, but health improvements have not been consistently found, prompting debate about the value of subsidized coverage. We revisit this question using a cluster-randomized experiment of a fully subsidized, mobile-based health insurance program in Kenya, combined with weekly surveys collected over eighteen months. Full subsidies combined with trusted, hands-on registration support drove take-up near universal coverage, addressing the weak first stage that has limited prior evaluations, while high-frequency measurement captures health events that standard recall surveys miss. We find that insurance reduces catastrophic health expenditure, shifts responses to illness from foregone care and self-medication toward formal care, and speeds care-seeking after illness or injury. We find suggestive evidence of health improvements through fewer days unable to perform daily activities. These effects would likely go undetected in standard endline surveys, underscoring the value of high-frequency measurement combined with strong take-up

No abstract is available for this item.