The Timescale of Economic Model How Long is the Long Run?

Abstract

Although models of economic growth have been intensively studied in recent years, relatively little attention has been given to the underlying timescale of these models.² While in many cases we know how the major variables of the models change over time, in very few cases do we know how quickly they will change. Yet the speed of change is a prediction of the model, and by examining this we have a further test of the model’s properties. For example, in many cases it is shown that all paths converge to a long-run equilibrium, but we also want to know how soon the paths will reach the vicinity of this equilibrium. The speed of convergence makes a great deal of difference to the way in which we think about the model. Alternatively, where a model gives rise to oscillations, we need to have some idea as to their probable period. If we throw away information about the time dimension, we are reducing still further our limited understanding of the relationship between these models and the real world.

I should like to thank F. H. Hahn, R. E. Hall, G. de Menil, D. M. G. Newbery, M. Rothschild, K. Shell, J. E. Stiglitz, and J. H. Williamson for helpful comments on an earlier version of this paper. They bear no responsibility for any remaining errors. The calculations in the paper were carried out at the Massachusetts Institute of Technology Computation Center.