The Effectiveness of Public Health Insurance: Evidence From Rajasthan, India

Abstract

Expanding public health insurance programs and contracting private hospitals for service delivery are common policy strategies to meet the goals of universal health coverage, but evidence from lower income countries on their design and function is limited. My dissertation studies the effectiveness of the BSBY government health insurance program that entitles 46 million low-income individuals to free care at public and empaneled private hospitals in Rajasthan, India. We use a unique dataset of insurance claims linked to post-visit patient surveys that allows us to analyze hospital-patient interactions under insurance. In the first paper, we document substantial out-of-pocket payments (OOPP) at private hospitals under insurance across a range of health care services and find that higher risk and less informed patients pay more. In the second paper we investigate whether hospitals are charging in order to compensate for reimbursement rates that are too low. We exploit a policy reform that discontinuously changed hospital reimbursements for different procedures by varying magnitudes to conduct a difference-in-differences analysis of private hospital responses, and find that less than half of the higher public subsidies are passed through to patients in the form of lower OOPP. In the third paper, we conduct an experiment to test whether providing phone-based information to eligible patients about their entitlements under the program can help them hold hospitals accountable and reduce OOPP. We find that the intervention is effective at increasing patient awareness of the program and leads to dramatic reductions in OOPP, but only at public hospitals. Together this research provides evidence that 1) OOPP is not simply due to problems with eligibility, enrollment, or facility choice, but also hospital charging behavior, 2) hospital capture contributes substantially to the observed high OOPP under public insurance in India, and 3) patient-driven accountability interventions may be important but insufficient to improve the effectiveness of public insurance programs.