Essays on public and private welfare provisions in China

This thesis consists of three self-contained essays that are aimed towards contributing to the understanding of the emerging of the public and private welfare states in devel- oping economies. Three chapters, specifically, focus on how the public policies affect individuals labour market participation and what affects the private provision of the social safety-net in the context of China. The first chapter provides novel empirical evidence for a question: how is the norm of providing old-age support transmitted inter-generationally in China? Intergenerational old-age support within families is an important norm in developing countries, which typically lack comprehensive pension coverage. The transmission mechanism for this norm is potentially influenced by socioeconomic factors internal and external to the family, which the norm may in turn influence. This chapter studies the inter- generational transmission of this social norm in China, focusing on the role of gender. The suggested mechanism behind this transmission is that parents, by their provision of support to their own parents, shape their same-gender children's preference for old-age support. Given that the gender ratio of Chinese children is not random, I develop an instrumental variable strategy using an interaction term of the timing of the ban on sex-selective abortions in China and the gender of the first-born child as the instrumental variable for the gender of the children to alleviate the possible en- dogeneity. The empirical results, using two Chinese datasets, show that parents with more same-gender children provide more support to their ageing parents than parents with cross-gender ones, controlling for their household size. The father effect is more significant in rural subsamples, and the mother effect mainly exists in urban areas. The urban-rural difference in the results may indicate a normative shift accompanying economic and demographic changes. The second chapter presents a theoretical framework for understanding the empirical evidence in Chapter 1. Based on the model of the \demonstration effect" by Cox and Stark (1996), I construct a model describing the intergenerational transmission of social norms in old-age support. The model combines the \demonstration effect" and the same-gender transmission channel. The parents are more likely to influence their same-gender children in terms of providing old-age support, thus they provide more old-age support if there are more same-gender children in the household. The key parameter distinguishing my model from the existing literature is the gender of the future generation. The baseline model concludes that fathers with more sons in their households provide more old-age support to their parents than fathers with more daughters, assuming the number of children are exogenous. Mothers provide more sup- port to their parents with more daughters in their household. The conclusions from the baseline model are shown to be valid under models with generalised assumptions. The last chapter studies how misallocation in labour markets in China can be caused by the provision of public welfare programmes. Providing health insurance with certain geographical restrictions may lead to possible misallocations in the labour market by hindering migration. This chapter tests whether the new rural health insurance introduced in 2003, the New Cooperative Medical Scheme (NCMS), had unintended and negative fects on rural-to-urban migration mobility in China. The NCMS only offers health insurance to people with rural household registration, and rural residents can only benefit from the NCMS if they visit the hospitals near their registered location in the household registration system. Utilising a new dataset collected from provincial yearbooks in China, the results of the event-study approach show that the NCMS does not reduce the percentage of rural residents who are rural-to-urban migrants and working outside their home counties at the county level but does have negative effects on its growth rate. Using the China Health and Nutrition Survey (CHNS), my instrumental variable results find that being enrolled in the NCMS decreases the probability of being a migrant at the individual level. The IV is a time-variant dummy indicating the counties that has relative early NCMS implementations. I also used the CHNS to construct a county-level dataset and replicate the county-level results. Together, the results suggest that the NCMS gradually locks the rural labour force into rural areas and further hinders geographical job mobility in China