Adaptive learning and equilibrium selection in experimental coordination games: An ARCH(1) approach

We analyze and estimate an adaptive learning model which explicitly considers strategic uncertainty in order to provide a unified explanation of the patterns of equilibrium selection in J. Van Huyck, R. Battalio, and R. Bell's (1990, Amer. Econ. Rev. 80, 234-248; 1991, Quart. J. Econ. 106, 885-910) experimental coordination games. The leaning dynamics follow an autoregressive conditional heteroskedasticity process of order 1 (ARCH(1)). We find that ARCH captures the persistence in players' assessments of uncertainty as they update their beliefs, provides a reduced form for differences of unknown form in prayers' beliefs without relying on any assumptions on their probabilistic structure, and yields a parsimonious parametric description of the learning dynamics. Journal of Economic Literature Classification Numbers: C92, C73, C51, C25. (C) 2000 Academic Press.