Centralized assignment mechanisms and assortative matching: Evidence from Chinese universities

Student-college academic mismatch has been documented in the literature. This study takes advantage of a unique large-scale policy experiment in China in 2013 in which a variant of the Boston mechanism was replaced with a variant of the deferred acceptance mechanism and compares the quality of assortative matching between student test scores and college quality under the two mechanisms. College admission in China is performed via a centralized matching process based only on one national standardized test. Using administrative data from one province, I find that the restricted deferred acceptance reduces the number of over-matched students (low-scoring students attending high-quality colleges) by 7.1 percentage points (or 51.9%). However, on average, it increases the number of under-matched students (high-scoring students attending low-quality colleges) by 2.7 percentage points (or 12.8%) for those at the top tier based on test score distribution. Moreover, the effects are mainly experienced by students at the bottom 50% of the top tier. The increase in under-matched shares fades over time for men, but persists for women.