Interactions between large macro models and time series analysis

Building large models, with little dynamics, was long considered to be an alternative to small dimensional time series models involving many lags. The advantages of one modelling methodology are compared to others; such as the size of the model, the use of economic theory, and simultaneity in specification. The question of how to evaluate the possible relative advantages of these alternatives is discussed. The conclusion is that in the future, time series models have to become larger, that is, involve more variables and that some lessons can be learnt from the construction of current large econometric models. Copyright © 2002 John Wiley & Sons, Ltd.