« Back to Results

Is College Admission in China Meritocratic?

Paper Session

Monday, Jan. 4, 2021 3:45 PM - 5:45 PM (EST)

Hosted By: Chinese Economists Society
  • Chair: Yan Chen, University of Michigan

The Institution of Merit: A Study of Chinese College Admissions

Jimmy Chan
,
Chinese University of Hong Kong
Zhi Wang
,
Fudan University
Shunyi Zhao
,
University of Pennsylvania

Abstract

While there is a broad consensus in many countries that places at elite colleges should be allocated by merit, how merit should be measured is a hotly contested issue. Under the current system in China, within province admissions are determined by a college entrance exam. But as tests scores are not comparable across provinces, universities must assign separate admission quotas to different provinces. A generally held belief is that elite universities in Beijing and Shanghai discriminate against out-of-province students.

We investigate how elite Chinese universities choose between applicants from different provinces. We construct a model that captures the key features of the Chinese system. In our model, universities care about student quality, but may disagree over the student quality in different provinces. We show that in equilibrium each university should set admission quotas to equalize the marginal student quality in different provinces. This implies that a university should assign a bigger quota to a province where the university is popular or where the student quality is higher.

We estimate the model using the quotas and cutoff scores of one hundred top Chinese colleges. We find that different universities tend to rank the student quality in different provinces similarly. If one university considers the average student quality in one province is higher than that in another province, then other universities are likely to consider the same. There are sizable disparities in average student quality across provinces. In particular, provinces with higher pre- college education spending, including Beijing and Shanghai, have significantly higher student quality. Higher ranked universities and those directly under provincial control tends to bias in favor of home students, but the magnitude of the bias is small. Counterfactual experiments show that regional inequalities in admissions is driven largely by the disparities in student quality.

How College Choices Disadvantage Disadvantaged Students under Centralized College Admissions Systems

Prashant Loyalka
,
Stanford University
Binzhen Wu
,
Tsinghua University
Xiaoyang Ye
,
University of Michigan

Abstract

We study “undermatching” in a centralized college admissions system. Using unique administrative data from China, we show that, conditional on qualifications, rural students, a typical disadvantaged group in China, are undermatched. They are less likely to be admitted, and, conditional on admission, they are admitted to less selective colleges than urban students. Rural students also choose less selective colleges that have lower tuition, are in-province, and that are lower risk (have larger enrollment quotas). Finally, evidence from discontinuous expansions of student choice sets shows that offering more choices of selective colleges exacerbates rural-urban gaps in college choices and college selectivity.

Matching Mechanisms, Justified Envy, and College Admission Outcomes

Le Kang
,
Peking University
Wei Ha
,
Peking University
Yang Song
,
Colgate University
Sen Zhou
,
Peking University

Abstract

Matching mechanisms are crucial for college admission outcomes. While theoretical and experimental studies have demonstrated that the widely adopted Immediate Acceptance (IA) mechanism is less stable than the Deferred Acceptance and the parallel mechanism, empirical evidence remains scarce. This paper documents the causal effects of changing matching mechanisms on matching stability and student outcomes, exploiting temporal and provincial variations in the reform timing in China. Individual-level data on the Chinese College Entrance Examinations (CEEs) and admission outcomes allow us to construct precise measures of justified envy under reasonable student preference assumptions. We find that changing from IA to parallel mechanism reduced undesirable matching outcomes, including justified envy measures, null admission, and retaking. We also find a nonlinear relationship between the improvement in justified envy measures and parallel choice bandwidth. Finally, congestion in mid-ranked universities could explain the higher levels of justified envy and larger improvements from mechanism change for students ranked in the middle.

The College Admissions Contribution to the Labor Market Beauty Premium

David Ong
,
Jinan University
Man (Amanda) Xie
,
University of Florida

Abstract

Beautiful people earn more. Surprisingly, this premium is larger for men than for women and is independent of the degree of customer contact. Overlooked is the possibility that beauty can influence college admissions. We explore this academic contributor to the labor market beauty earnings premium by sampling 1,800 social media profiles of students from universities ranked from 1 to 200 in China and the US. Chinese universities use only standardized test scores for admissions. In contrast, US universities use also grades and extracurricular activities, which are not necessarily beauty-blind. Consistent with beauty-blind admissions, student’s beauty is uncorrelated with the rank of their college in China. In the US, White men from higher ranked colleges are better-looking. As expected, the correlation is insignificant for White men who attended tech colleges and is highest for those who attended private colleges. We also find that White women and minorities of either gender are not better- looking at higher ranked colleges. Our evidence indicates a college admissions contribution to the labor market beauty premium for US White men, but not for students in China of either gender, White women, or minorities of either gender in the US, or for White men who attended technology colleges.
Discussant(s)
Binzhen Wu
,
Tsinghua University
Yang Song
,
Colgate University
David Ong
,
Jinan University
Zhi Wang
,
Fudan University
JEL Classifications
  • J0 - General