Three Essays on Population Studies

Abstract

This dissertation comprises three essays on population studies. I begin with a paper that is the first to investigate the short- and long-term effects of a recent change in Chinese divorce laws on married women’s and men’s well-being. The 2011 Chinese divorce reform transfers ownership of the family home to the registered buyer, most often the husband, in the event of a divorce. Prior to this legal change, the family home was considered joint property. Adopting a quasi-experimental study design and using data from the China Family Panel Studies, I found that this legal change led to gendered consequences. For a typical Chinese household where only the husband’s name is on the deed of the family home, it decreased women’s well-being while men’s well-being did not change. In the second study, I have examined the effect of the husband’s retirement on the wife’s health in China. The large increase in the probability of retirement at the legal retirement age for Chinese workers in the formal sectors enables me to exploit this discontinuity as a source of exogenous variation in retirement. Using data from the China Health and Retirement Longitudinal Survey, I implement a fuzzy regression discontinuity design to compare the health outcomes of wives whose husbands just retired with those of wives whose husbands are about to retire. Results show that the husband’s retirement significantly improves the wife’s health. In the third study coauthored with Scott M. Lynch, we develop an extension of the Bayesian approach to making multistate life tables (MSLTs) introduced in Lynch and Brown (2005). Among all extant methods, the Bayesian approach developed by Lynch and Brown (2005) offers several key advantages over other approaches, including the ability to incorporate prior information, direct and probabilistic interpretations of estimates, and the flexibility to incorporate model changes to handle idiosyncratic data. However, this approach has been limited to only two states, such as “healthy” vs. “unhealthy”, and cannot handle partially absorbing states. The main contribution of our method is to allow high dimensional state spaces with partially absorbing states.