Website CV Paper
Fields
Applied Micro, Econometrics, Labor, Education
Job market paper
Optimal Multistage Selection: Evidence from Graduate Admissions
This paper uses graduate admissions to study multistage selection under costly evaluation, a problem faced in settings from undergraduate and professional admissions to hiring. Using application files from an economics PhD program, accessed through a privacy-protecting synthetic data protocol, I relate acceptance to success measures based on post-program job placements. I estimate structural models of an admissions committee's decision problem capturing first-round screening and matriculation, develop a test of whether the committee's decisions are optimal given a constraint on how many files it can read, and quantify gains from switching to the models' decision rules. The committee's decisions align with applicants' estimated success probabilities along most dimensions, but some characteristics, such as work experience and TOEFL/IELTS scores of non-native English speakers, appear underweighted. Reweighting applicant characteristics could have increased the admitted cohort's average success probability by 4 percentage points without reading more files, an estimate corrected for the winner's curse.