Matching Candidates and Jobs in Government Recruitment under Multidimensional Sorting: An Empirical Study of Matching Games Without Individual-level Data

I study matching mechanism design in government recruitment, focusing on a primary method for recruiting bureaucrats in China: the National Civil Service Exam (NCSE). The NCSE is a centralized matching mechanism where each candidate can apply for one position before taking a meritocratic exam, causing strategic application. An econometric challenge arises in studying matching games in the absence of individual-level data. I propose a job- matching model with two-sided heterogeneity and demonstrate its nonparametric identification using position-level data. I apply the framework to estimate an empirical model of the NCSE. In the counterfactual analysis, I introduce a strategy-proof ‘first-exam-then-apply’ mechanism. Simulations reveal that the NCSE can outperform the alternative mechanism when candidates’ characteristics are multidimensional.