Working Papers
1. Does Increasing the Generosity of Public Health Insurance Affect Medical Expenditures? Evidence from China’s Critical Illness Insurance Program
A large body of literature has studied the impact of gaining public insurance coverage, but we know little about how the generosity increase affects health care expenditures and utilization, especially among middle-aged and elderly people, who are at high risk of illness. China launched the Critical Illness Insurance (CII) program in 2012 with the goal of lowering high out-of-pocket costs for the rural and urban insured population without formal jobs. I use unique data on CII phased rollout dates across cities combined with the data from China Health and Retirement Longitudinal Survey of individuals over 45 years old, to study the impact of CII. Due to the large mass of zeros and highly right-skewed observations in the medical costs data, I use two-part models in a generalized difference-in-differences and event-study framework to examine the effects of CII on medical costs. Similarly, with the presence of a large number of zeros and over-dispersed data in the measure of health care utilization, I apply the negative binomial model with generalized difference-in-differences and event-study framework. For the rural insured people, I find CII had a negligible effect on the likelihood of using inpatient care but decreased out-of-pocket inpatient costs by 45% for those who did. Individuals over 60 years old, individuals from poor households, and those who are chronically ill experienced the more considerable reductions in out-of-pocket inpatient costs. I do not find significant evidence showing that CII affected the urban enrollees' medical costs or health care utilization, which is expected given the higher level of pre-existing benefits available to urban residents.
2. Maternal Education and Early Childhood Outcomes in China (with Xiaozhou Ding).
In this paper, we study how maternal education affects children’s early childhood health outcomes and the development of social and motor skills. We take advantage of the higher education expansion in China, which creates credible exogenous variation in access to col- leges that improves mothers’ educational attainment, to examine these effects through an instrumental variable approach. Our results show that increases in years of schooling beyond the nine-year compulsory education level significantly improve children’s outcomes. We find the probability of an infant having low birth weight is reduced and the time it takes for a child to start speaking, counting, and walking is shortened. We investigate several mechanisms that could explain these results and find that mothers’ schooling is strongly associated with assortative marriage and rural-urban migration. Suggestive evidence also shows mothers with more schooling are likely more aware of how to effectively invest in their children.
3. High-Speed Rail Accessibility and Housing Prices: A Case Study of Jiangsu Province in China (with Shiyu Cheng), Under Review.
This paper investigates the impact of improved High-Speed Rail (HSR) accessibility on housing prices by comparing properties near new HSR stations to those near existing ones, using complex-month-level data from 2015 to 2018 in Jiangsu Province, China. Since city governments can determine the placement of new HSR stations within their localities but lack influence over site selection in other cities, we consider the reduction in distance to the nearest HSR station resulting from new station openings outside the city boundaries as exogenous. Employing a Difference-in-Differences (DD) approach, our findings reveal a significant 20% reduction in housing prices when station distance decreases, i.e., HSR accessibility improves due to the establishment of new HSR stations outside the local city. The effects of station openings within city limits are statistically insignificant. A 1-kilometer reduction in station distance corresponds to a 0.5% decline in housing prices. These negative impacts are observed across varying distances— ranging from 106 to 15 kilometers—from the nearest HSR station.