Tuesday, August 25, 2015

Do participants try to strategize in strategy-proof mechanisms? Alex Rees-Jones surveys medical students about the NRMP.

One of the papers I heard at the recent SITE conference at Stanford was this one, reporting a survey of medical students engaged in the NRMP.

Suboptimal Behavior in Strategy-Proof Mechanisms:Evidence from the Residency Match 
Alex Rees-Jones
The Wharton School, University of Pennsylvania
August 10, 2015

Abstract: Strategy-proof mechanisms eliminate the possibility for gain from strategic misrepresentation of preferences. If market participants respond optimally, these mechanisms permit the observation of true preferences and avoid the implicit punishment of market participants who do not try to “game the system.” Using new data from a flagship application of the matching literature—the medical residency match—I study if these potential benefits are fully realized. I present evidence that some students pursue futile attempts at strategic misrepresentation, and examine the causes and correlates of this behavior. These results inform the assessment of the costs and benefits of strategy-proof mechanisms, and demonstrate broad challenges in mechanism design.

From a survey of graduating medical students: "I find that 17% of students self-assess their preference reporting strategy to be nontruthful, with 5% directly attributing this nontruthful behavior to strategic considerations."

1 comment:

dWj said...

cf. second-price auctions, e.g. Kagel, Harstad, and Levin (1987).

BTW, the link isn't working for me.