How Middle-Movers come out on Top: A Theoretical and Experimental Analysis

A theoretical model nds a prominent middle-mover advantage in an openly sequential procedure. The equilibrium analysis also nds, both, a rstand last-mover disadvantage. These results provide general intuition about the underlying incentives in a sequential process. Additionally, examples are used to show how a sequential process can be manipulated to construct settings with, either, a rstor last-mover advantage. The equilibrium predicts an asymmetric distribution of expected payo s across positions. This distribution guarantees the maximum possible payo for the middle-mover. A laboratory experiment is used to test whether or not human behavior aligns with the theory's unintuitive expected payo distribution. Data from the experiment unambiguously support a middle-mover advantage as well as, both, a rstand last-mover disadvantage. Also, observed payo s better re ect the theoretical prediction in later rounds, thus suggesting learning behavior. JEL Classi cation: C72; D71; D72; L25 ∗Department of Economics, University of California-Irvine, Social Science Plaza 3151, 92697-5100. Ph#623-221-4446. U.S.A. E-mail: [email protected]. †I would like to thank Donald Saari for his guidance, support, and constant encouragement on developing this project. Special thanks go to Don-Squad members, in particular: Heidi Tucholski, Daniel Jessie, Reuben Kline, Tomas McIntee, Johnathan Cook, and George Ng. I greatly appreciate the time spent by Katri Sieberg and Michael McBride to re ne this idea. I would also like to thank Louis Narens, Olga Shvetsova, Stergios Skaperdas, Anthony McGann, and Igor Kopylov for detailed comments and bene cial discussions. This research was funded by fellowships from UC Irvine's Institute of Mathematical Behavioral Sciences and the Department of Economics, as well as a grant from UC Irvine's Experimental Social Science Lab. Additionally, I was supported by working as a graduate student researcher for Michael McBride on an Army Research O ce MURI grant # W911NF-11-1-0332.