Learning about monetary union: An analysis of bounded rational learning in European labor markets

Abstract

This paper addresses the issue of rational learning in European labor markets where individuals have to form expectations about the future. It extends the work of Anderton, Barrell, and In't Veld (1992) and Anderton, Barrell, In't Veld and Pittis (1992) where forward-looking wage equations are introduced into the National Institute's Global Econometric Model, GEM. The model was developed by the Institute, and it is now jointly maintained with the London Business School. We extend the analysis of learning undertaken by Hall and Garratt (1992) on the London Business School model of the U.K. economy, where bounded rational learning was first introduced into large-scale models. Agents are assumed to form expectations about future prices using a Kalman filter-based variable parameter learning mechanism. This learning mechanism is compared to model-consistent expectations and a fixed-parameter adaptive expectations mechanism. Ours is the first study to use multiple learning rules, and the first to use learning rules in labor markets in large estimated nonlinear macro models. We analyze an oil price rise and realignments of the franc, lira, and pound within the ERM, and compare the three expectations environments.