ENERGY SUBSTITUTION IN THE U.S. FOOD SUPPLY INDUSTRY: AN ECONOMETRIC TEST IN MULTI-FACTOR PRODUCTION FUNCTION.

摘要

The purpose of this dissertation is to investigate whether there exists substitutability of complementarity between energy and non-energy inputs in the U.S. food supply industry--namely, the agricultural, fishing, and food processing industries.;The basic function forms of the investigation are the Transcendental Logarithmic (TL), the Genealized Leontief (GL), and the Constant Elasticities of Substitution (CES) production and cost functions. Because random deviation from profit maximization would affect factor markets, the parameters of the models have been estimated using Zellner's two-stage GLS (generalized least squares) method. In order to obtain conclusive results, checks for neoclassical restrictions--monotonocity and convexity--for TL and GL approaches have been performed. To investigate the significance of the model estimates, goodness of fit has been compared in terms of pseudo-R('2) and likelihood test statistics at the .01 significance level.;Finally, in accordance with the combined results of the above tests, we have found that the PES between capital and energy, between capital and labor, between capital and material, between labor and energy, and between energy and material inputs indicate substitution in all industries. However, we have also found the PES's between labor and material indicate complement only in agriculture and the fishing industry.;The estimation of technical substitutability between energy (E) and non-energy inputs (K, L, M) has been done in the manufacturing process. All studies have found energy (E)-labor (L) and energy (E)-material (M) are substitutable. But Berndt-Wood and Magnus find energy-capital to be strong complements, while Griffin-Gregory and Pindyck find strong evidence of substitutability. There are three reasons for these conflicting conclusions: First, the complementarity results were based on single-year or time series data, while the substitutability results are from cross section investigations. Second, the substitutability results omitted the material input (M) in their model. Finally, substitutability results refer to a capital input as non-labor value-added, while complementarity results refer to a capital input composed only of equipment and structures, called "reproducible capital." In order to obtain conclusive evidence, four input factors (K, L, E, M), pooled cross section time series data, and reproducible capital as capital input are used in the model for this study.