Transfer Paths and Academic Performance: The Primary School Merger Program in China

In the late 1990s and early 2000s, China's Ministry of Education embarked on an ambitious program of primary school mergers by shutting down small village schools and opening up larger centralized schools in towns and county seats. The goal of the program was to improve the teacher and building resources in an attempt to raise the human capital of students in poor rural areas, although it was recognized that students would lose the opportunity to learn in the settings of their own familiar villages. Because of the increased distances to the new centralized schools, the merger program also entailed building boarding facilities and encouraging or mandating that students live at school during the week away from their family. Given the magnitude of the program and the obvious mix of benefits and costs that such a program entails there has been surprisingly little effort to evaluate the impact of creating a new system that transfers students from school to school during their primary school period of education and, in some cases, making student live in boarding facilities at school. In this paper, our overall goal is to examine the impact of the Rural Primary School Merger Program on academic performance of students using a dataset from a survey that we designed to reflect transfer paths and boarding statuses of students. We use OLS and Propensity Score Matching approaches and demonstrate that there is a large ''resource effect'' (that is, an effect that appears to be associated with the better facilities and higher quality of teachers in the town and county schools) that appears to be associated with the transfers of students from less centralized schools (such as village schools) to more centralized schools. Boarding, however, is shown to have negative impacts on academic performance. However, students who transfer to county school benefit from the transfer no matter where they start and whether they board or not. (Contains 2 figures and 8 tables.)