Something New in Medical Residency Matching Markets

Abstract

Worldwide medical residency markets commonly employ variants of the two-sided central clearinghouse designed by Roth and Peranson in 1999. In the NSW physiotherapy residency matching market, a one-sided and computationally efficient matching mechanism is used – the Kuhn-Munkres algorithm. The mechanism is new for medical matching markets, with no publicly known application and no existing literature. A crucial contribution of the thesis is presenting the algorithm and starting a discussion around the Kuhn-Munkres algorithm in matching. The thesis models the iterative working of the Kuhn-Munkres algorithm. I show that the Kuhn-Munkres algorithm is rank-efficient, outcome unfair, procedurally fair and not strategy-proof. Comparing the Roth-Peranson and Kuhn-Munkres algorithms on efficiency, fairness and incentive properties, the thesis concludes that there is no settled winner between the two algorithms. The competition eventually comes down to the trade-off between cost reductions and market complexities.