The Compromise Algorithm in the Swedish PhD Admissions Problem

Abstract: What mechanism should be designed to allocate PhD applicants to universities in Sweden? We introduce the Swedish PhD admissions problem, and it is influenced by the college admissions problem (Gale and Shapley 1962) and the student placement problem (Balinski and Sönmez 1999). In order to “solve” this problem, we design a novel mechanism, namely the compromise algorithm. We propose three theorems from this algorithm, i.e. equivalence theorems. The equivalence theorems specify the equivalence relations among stability, worse and responsiveness. Additionally, we find a positive result that the number of fields determines the strategy-proofness of the algorithm; meanwhile, the student optimal stable matching and the university optimal stable matching can be treated as special cases of our model when we restrict the number of fields. Generally, the compromise algorithm generates a stable matching that falls in between the student optimal stable matching and the university optimal stable matching.