THE HARVARD HOUSING LOTTERY : RATIONALITY AND REFORM Susan

This paper studies allocation schemes Harvard has used to assign undergraduates to residential houses. The main conclusions are as follows. First, students appear to behave remarkably rationally. Second, simulations suggest that, given rationality, Harvard's allocation scheme sometimes allocates fewer students to one of their top three choices than would occur under an alternative system that generates truthful revelation. * We are indebted to Ling Hui Tan and Gabriel Loeb, as well as to Ataman Ozyildirim, Kathleen McDill, Phillip Swagel, Elaine McCormick Watt and Kristen Williams for research assistance. We would also like to thank the Harvard Housing Office for providing data, Evelyn Taylor for assistance in typing, John Pratt for directing us to the relevant literature, and Vijay Krishna, Richard Zeckhauser, AI Roth and anonymous referees for detailed and very helpful comments on earlier versions. This project was partially supported by the Harvard Institute for Economic Research.