Estimation and hypothesis testing using dynamic models of firm behavior
Abstract
Recent applications of dynamic firm models have utilized data aggregated across commodities and firms. This approach has several shortcomings: (i) restrictive conditions must be applied to ensure exact aggregation across firms, (ii) the estimated dynamic structure is unlikely to represent any of the sub-aggregates, and its is likely to differ from the summed sub-aggregate structures, and (iii) degrees of freedom considerations dictate relatively simple models which do not capture complex dynamic relationships. This thesis develops a panel data set from business records of Indiana dairy farms for the years 1971 to 1982. It distinguishes two capital stocks (machinery and dairy herd), two outputs (crops and milk), and four inputs (hired labor, crop, livestock and other inputs). A cost of adjustment model, derived from a quadratic value function, is fitted to this data. The implied parametric restrictions are rejected implying that intertemporal profit maximization is not consistent with observed behavior. A less restrictive, disequilibrium model is accepted at a 0.025 level of probability. This is a model of partial and interrelated adjustment among commodities with long run behavior characterized by a quadratic profit function. Estimates of the adjustment coefficients suggest: (i) the own-adjustment coefficient for the dairy herd (0.448) is twice as large as that for machinery, (ii) expansion in any capital stock reduces the short run demand for the other stock, and (iii) expansion in machinery temporarily reduces crop production as well as short run demands for hired labor, and crop and livestock inputs. Sample means of adjustment speeds (i.e., that portion of desired adjustment accomplished in a year) were computed over the 1971-82 period. Most of them differed markedly from the own-adjustment coefficients, because cross-adjustment effects, which are taken into account in the computation of adjustment speeds, were large. Extreme cases are milk production and the dairy herd; their adjustment speeds were equal to $-0.03$%. Adjustment in crop production and machinery was not significantly affected by disequilibrium in the dairy herd; their adjustment speeds equalled 90% and 23%, respectively. Hired labor, crop and livestock inputs were significantly affected by disequilibrium in capital stocks, with adjustment speeds equal to 79%, 57%, and 67%, respectively.
Degree
Ph.D.
Advisors
Hertel, Purdue University.
Subject Area
Agricultural economics
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.