| Sumario: | The article presents some results from a pseudo data analysis of a small hypothetical process model that casts doubt about the usefulness of the approach. In the article, the authors discusses the statistical properties of the errors, the question of the appropriate norm for approximation errors; the alternative functional forms; and the issue of single versus multiple equation methods of estimation. An illustrative analysis of a small process model and the resulting elasticities are presented. The purpose of the "pseudo data approach" is to condense the technical information in a complex process analysis model into a single equation. This single equation summarization is supposed to give estimates of several price and substitution elasticities as well as input-output coefficients which can presumably be used in other modeling exercises. The authors states that their examples illustrates the fact that the elasticity estimates obtained from the single equation summarization of the data generated from the process model are likely to be unstable and thus, not of much use.
|
|---|