Laboratory Tests of Equilibrium Predictions with Disequilibrium Data

Abstract

We examine a common practice used within previous studies of laboratory markets, testing equilibrium models using some data from markets that have not reached an equilibrium. We examine the effect of this practice when it is applied to some laboratory markets where an appreciable number of them eventually satisfied an operational definition of an equilibrium. Our data suggest that, for our markets, significantly different equilibrium test results would be obtained when using all data available in market periods analysed in previous studies rather than using only equilibrium data. For our markets, choices made in disequilibrium are quite different from those in equilibrium.