Essays on Development Economics Applied to China Open Access
Downloadable ContentDownload PDF
This dissertation focuses on development issues in China. The first two chapters analyze the effect of labor mobility restrictions on education in China. First, I explore an extension of Regression Discontinuity (RD) design when individual treatment status is missing. Neither of the empirical estimation methods for traditional sharp nor fuzzy design is applicable in this situation. I propose a weighted sharp design method that allows RD to work when only the aggregate treated proportion is available. The idea is to rescale the raw discontinuity in the outcome variable using the aggregate treated proportion. This rescaling method is similar in spirit to Angrist and Kruger's two sample IV estimation. With this method, the ``quasi-experimental'' RD design could be applied to more empirical cases, especially for research on developing countries where the observability of individual level data is a prevailing problem.Then, I provide an empirical application to the weighted sharp RD design by analyzing the effect of labor mobility restrictions on human capital investment decisions in China. The Hukou system restricted most rural-urban migration in China for over 50 years. Under this system, rural residents could permanently migrate to urban areas by acquiring higher education, which gives them additional incentives to invest in human capital. In this dissertation, I test the hypothesis that mobility restrictions, combined with selective migration policies, encouraged education. I use a 1998 Hukou policy reform as a natural experiment, which granted urban residency to a group of rural individuals based on the birth date. Findings suggest that human capital investment for rural residents decreased sharply when they were allowed to permanently migrate to urban areas. Nonparametric estimates show a statistically significant drop in high school attendance rate, ranging between 8.7 to 30.5 percentage points, depending on specifications and assumptions. There are also signs of a decrease in the middle school graduation rate and an increase in the dropout rate of high school and above. These effects are bigger for males and for those able to permanently migrate to relatively rich areas.In addition to migration and education, this dissertation also studies the rise in rural and urban household saving rates in China from 1995 to 2002. I analyze the rural saving determinants and urban saving determinants each across their respective quantiles, and then compare the difference. While a bigger increase occurs at the lower quantiles of rural saving, urban saving rises the most at the top end of the distribution. Moreover, I decompose the saving shift in each sector into an endowment effect, which could be explained by the change of households' attributes, and a return effect, which is due to the changing coefficients of these factors on saving. I find that the endowment effect, which is mainly driven by rising income, can account for the majority of increase in rural saving. For urban households, however, only a small portion of saving shift can be attributed to change in household characteristics. These results highlights the different saving motives between rural and urban areas and help people better understand the high saving puzzle in China.