Structural Analysis and the Measurement of Demand for Farm Products
TEN YEARS have passed since publication of Haavelmo's first article on simultaneous-equations approach.' These years have produced an abundance of theoretical literature and have greatly increased capital requirements of those who would engage in measurement of economic relationships. Measurement without theory has (quite properly) been drummed out of journals. However, measurement with theory has been painfully slow come forward. This may be due in part a sort of natural selection, in that those who have been most forehanded in acquiring new methodology have (with few exceptions) been chiefly interested in methodology per se. For them personally, empirical applications of new may be even less inviting than research using simpler methods. For, paraphrase Haavelmo,2 economists will have revise their ideas as to not only the level of statistical theory and technique but also the amount of tedious work that will be required, even for modest projects of research. There is no point in lamenting division of interest between methodologists and applied workers. But it appears that burden of testing new econometric tools under operating conditions must be taken up by latter group. Specific cases of breakdown or evidence of poor design may then direct attention of methodologists toward improving their product or modifying their claims for it. My own experience suggests that advertising has been much too derogatory of a long-established competitor. During past few years my work in United States Department of Agriculture has involved a considerable amount of statistical demand analysis. The object of this work has almost invariably been obtain numerical results which made sense in terms of commodities and classes of economic agents involved -that is, results of structural significance. In all but a few cases I have used singleequation methods for estimating desired coefficients. I accept proposition that many economic phenomena must be explained in terms of two or more simultaneous relationships. However, single-equation methods appear be both practically and theoretically appropriate for estimating many structural relationships in field of food and agriculture. The first section of this paper is an appraisal of applicability of single-equation methods statistical demand analysis for farm products. The second section deals with an older and simpler problem in structural analysis. This is adjustment of least-squares results for effects of measurement errors in independent or predetermined variables, object being obtain best estimates of coefficients of reversible (hence, structural) demand functions. The common element in two sections is emphasis upon estimation of coefficients or parameters of reversible demand relations. The usefulness of single-equation methods for predicting future values of a variable (given unchanged structure) has not been disputed by proponents of simultaneous-equations approach. Their applicability structural analysis has, I believe, been underestimated during past decade, and some reaffirmation of their value in this area is needed.
- DOI
- 10.2307/1924877
- Volume
- 36 (1)
- Pages
- 57
- Export
- BibTeX
- Sources
- openalex crossref