Testing the representative agent assumption: the distribution of parameters in a large-scale model of the EU 1972–1998

Abstract

This study considers the validity of an assumption of the representative agent often made in economic models, that the behaviour of an economic group is adequately represented by each member of the group having the identical characteristics of the average. It can be tested whenever sufficient data are available to estimate sets of parameters of a model with and without the assumption that they are equal to some average. The specific test of the assumption is done for a set of 720 estimated parameters associated with the effects of technological change on industrial employment for 18 regions of Europe. The main finding is that the parameters are significantly different across industries and across regions and that an assumption that all responses to technological change are the same is not justified. Similar results are found for other parameters studied. This finding implies that the assumption that each industry has the same responses as an average, chosen from a literature review, is likely to be misleading. The data strongly suggest that behaviour is differentiated by industry and region.

Acknowledgements

The authors thank Donald Robertson, Milan Scazni and the participants of the Charles University Prague round table ‘Assessing the Environment: Methods on quantification of the economic impacts and externalities’ for useful comments and discussion over earlier versions of the paper. The participants of the Economics Department seminar at the Catholic University of Chile also provided useful discussion. We are grateful to Ron Smith and Hashem Pesaran for their comments and discussion of the paper. Any remaining errors are the sole responsibility of the authors.