Empirical Studies On Health Insurance System Reform In Urban China

Health insurance is recognized as a worldwide problem. So far, more than 160 countries or regions around the world have established their own medical insurance system, but no one country can provide a "standard model", and almost every country's health insurance system has many difficulties to solve.From the early 1990s China launched the medical insurance system reform, and after twenty years of effort, has established the urban basic employee medical insurance(UBEMI) for urban workers (1998), the new rural cooperative medical care insurance (NRCMI) for rural people (2003) and urban basic resident medical insurance (UBRMI) for urban non-employed (2007).In this dissertation, I try to use a unique data set to answer several important questions that are related to the social health insurance in China. The dissertation is structured into seven chapters.In chapter one, I give a brief introduction to the questions this dissertation focus on and my research framework.In chapter two, I briefly describe the Chinese medical insurance reforms in the last ten years, especially the URBMI which was just launched in 2007, but now has covered 195 million urban residents.In chapter three, I examine the problem of adverse selection in URBMI. The results show that adverse selection exists.In particular, individuals with worse health status or higher health risk are more likely to enroll in URBMI than individuals with better health status, and individuals who enroll in URBMI will use more medical services than the uninsured. I also find that employees who already covered by the UEBMI with worst self-reported health status are less likely to buy commercial health insurance than other individuals, and they use more medical services than others who are not covered by commercial health insurance.In Chapter 4,I focus on health status and the utilization of medical services of the'sandwich'group. Some employees are not covered by any health insurance because their employers do not pay their portion of the insurance premium. And at the same time, those employees are not eligible to join NRCMC (because they are urban residents), the free public health insurance (because they are not government employees), URBMI (because they are employed or had been employed before retirement), and finally they generally cannot afford the private commercial insurance. We call these employees "sandwich" group. Compared with the general uninsured, the sandwich is uninsured mostly because of poor performance of their employers, rather than "voluntary" uninsured. The results show that the'sandwich' group suffers from worse self-reported health status, utilizes less medical services, and have less incentive to get health information and medical examination than others.In chapter 5, I study whether and how much urban households can insure against a negative health shock, which is very common in underdeveloped countries and is considered one of the most important causes of poverty in China. A serious illness will not only lead to the loss of work ability thus income of a family, but also very likely resulted in a considerable amount of medical expenses during the treatment of the disease. However, the rise of health care costs and the loss of income do not mean that the family's consumption behavior has to be adjusted accordingly because it can usually protect itself from the negative health shock via a variety of formal as well as informal insurance mechanisms, such purchasing health insurance, reducing savings, selling assets, or borrowing from relatives and friends or getting loans from credit institutions. However, since the majority of serious illness is difficult to predict and the loss of family income may be very large, we assume that most families are unlikely to carry out full consumption insurance meaning that they would adjust their consumption behavior whenever there is a negative health shock. Our empirical results show that (1) consumption insurance is not complete, especially for low and mid income families; (2) the health insurance have no significant welfare effects but increase the medical expenditures.Chapter 6 tries to compare the income redistribution effect of NRCMC and URBMI. The Difference in difference model and fixed effect model estimation results show that the URBMI has reduced the income inequality significantly while the NRCMI has a non-significant effect on income inequality.The last chapter makes some conclusions and proposes some further improvements.