Applications of Panel Data Models in Health Economics

This thesis contains three independent studies that examine health economics related issues using a Bayesian hierarchical random coefficients model, IV-probit model and time-varying grouped fixed effects model. The research uses both macro level data and large hospital administrative data at individual level. We find health care is a necessity. Patients in public hospitals stay shorter than those in private hospitals due to the funding pressure imposed by the government. Shorter length of stays will lead to worse health outcomes reflected as higher readmission rate. The information can be used by hospitals and policymakers to improve the patient quality of care.