THREE ESSAYS ON LOW FERTILITY AND POPULATION AGING

Abstract

This dissertation investigates the contemporary issues of low fertility and population aging inSouth Korea, China, and the United States. The first and second chapters examine the linkage between health, work capacity, and well-being among older people using Health and Retirement Study (HRS). The third chapter sees the effect of the cash transfer on fertility in South Korea. Using the methodology of CMR (Cutler, Meara, and Richards-Shubik, 2013), the first chapter estimates the extent to which older adults can potentially extend their work lives due to the improvement of health status (health capacity to work) in South Korea and China. Besides, we estimate the growth potential when utilizing the capacity to work by linking the results to the National Transfer Accounts (NTA) of each country. Unlike the other advanced countries, only a part of the elderly receives old-age pensions in both countries, which further enables us to simulate the health capacity to work for the eligible pension group and the other who is not. The estimated health capacity to work is also reported for various sub-groups. The results suggest that the two countries seem to have much smaller economic gains from the health capacity to work, compared with other advanced countries. The results also imply that the unused health capacity to work is concentrated only on wealthy older people. The results for China suggest that while the elderly in rural areas tend to work more regardless of their pension status, urban residents show very similar patterns to Korea. We explain the results by linking them to Korea and China's labor markets and their old-age support systems. All these results indicate that it might be misleading to interpret the results in Coile, Milligan, and Wise (2017) as a major outcome of the statutory retirement age for the economies in countries with the less developed pension system. The second chapter carefully examines the effects of Internet use by older Americans ontheir physical and mental health conditions by utilizing the panel study of Health and Retirement Study (HRS). Although there is a growing body of studies on the effect of Internet use on health, the results are inconclusive. First, we investigate if using other tools such as a map or phone can also have a future effect on health. The results show that the Internet and other tools are all significant for improving health status in the near future when they are included separately as controls. Thus, the positive correlation of Internet use with mental health in the previous literature could be due to an individual's unobserved heterogeneity of using these tools, as it is hard to believe that other tools (map or phone) can also improve health conditions. Second, we employ a dynamic panel design with instrument variables to address the selection issue. When we use the previous Internet use as an instrument variable, the Internet variable remains significant for change of self-reported health status and cognitive skills, but the effects on depressive symptoms are not significant, suggesting that Internet use can have a distinct effect on general health status, not a specific health condition (depressive symptoms). Third, some older people may not use the Internet owing to a physical disability or a health issue. In this case, difficulty using the Internet might be just a proxy for the person's existing health status. To address this issue, we re-estimate the model controlling for the Instrumental Activities of Daily Living (IADL) and their interaction terms. The results suggest that a person's inability to access Internet use due to disability or serious health limitations has little to do with the positive linkage between Internet use and health outcomes. Finally, we estimate all models by individuals' education level, race, age group, and asset level. The results are more significant for the wealthier, more educated, and White. The third chapter examines the fertility effect of a cash transfer program designed toboost fertility in South Korea. Many Korean local governments have provided childbirth grants to families for all new births within the districts since 2002. The cash amounts and timing of adoption in the municipal districts (counties) vary substantially with no oversight from the central government. This study constructs yearly county-level panel data for 2000-2017 and uses panel fixed effects (FE) to control for counties' time-invariant specific characteristics. The results show that the program positively affects fertility rates in rural areas, but not in urban areas. The fertility effects are very small, however; a 100 percent increase in a cash payment is associated with raising total births per 1,000 people by only 0.025 in the rural areas. This small effect suggests that the cash support program may not be a cost-effective instrument. Political rather than economic motivations might explain the rapid adoption of the pronatalist program over a short period (2005-2009), raising an additional question about its economic significance. This research also calculates how much each additional childbirth costs the government, and the results show that more than 30 times of current cash grants are needed for extra childbirths. Thus, although the results are statistically significant, they do not show much meaningful economic significance. This study further examines whether the cash transfers have unintended consequences for neonatal health outcomes, such as birth weight. We employ a simultaneous equation model (SEM) to deal with simultaneity between childbearing and birth outcomes. The results show virtually no impact on infant health. The insignificant impacts on birth outcomes might be due to the insufficient cash amounts as exogenous and sizable income may be needed for significant effects.