Matching Impacts of School Admission Mechanism: An Agent-Based Approach

Abstract

Matching mechanisms are critical in determining the assignments of students to schools. We used agent-based modeling (ABM) to simulate the three mechanisms experienced in the fiercely competitive admission systems in China: serial dictatorship (SD), the Boston mechanism (BM), and the Chinese parallel mechanism (CP). We evaluated their multifaceted outcomes under different policy settings, school capacities, and behavioral assumptions. We have replicated their major characteristics found in the analytical models and showed that CP behaves as a hybrid of SD and BM. ABM allows us to distinguish their aggregate effects from distributional effects and their long-term level effects from short-term volatility effects. We found that this kind of “out-of-equilibrium’’ analysis, while mostly absent in the analytical equilibrium analysis, is crucial for practical policy analysis.