The Utilization of Agricultural Land: A Theoretical and Empirical Inquiry

Abstract

OR purposes of economic analysis, two important characteristics of agricultural land are its heterogeneity and the possibility of its use for the production of alternative outputs. One or both of these characteristics are frequently recognized in theoretical studies of land utilization. David Ricardo and his followers used the heterogeneity of land as a cornerstone in their theoretical analysis, but ignored the possibility of alternative uses. Later, Jevons and other economists extended the theory to encompass both characteristics. With the increasing application of linear programming methods, alternative land uses, and occasionally land heterogeneity, are being emphasized in empirical studies of land utilization for individual farms and small groups of farms. However, empirical studies of land utilization for the economy as a whole are seldom, if ever, constructed with an explicit recognition of these two important characteristics of agricultural land. These studies are either formulated without an explicit theoretical framework, or built around a statistical version of the neo-classical demand and supply analysis. This paper is aimed at the development of a theoretical analysis based upon maximizing behavior which can be empirically implemented and will generate numerical predictions of shortrun land utilization patterns for the economy as a whole. Particular emphasis is placed upon land heterogeneity and alternative uses. New methods are developed for the study of agricultural supply relations. Factors such as technological change and input substitution, which are of great importance for a long-run analysis, are largely ignored in the present short-run formulation. The analysis is developed within the institutional framework of agriculture as practiced in the United States. Land utilization decisions are made by a large number of farmers, each of whom possesses a relatively small parcel of land. Generally, we may assume that farmers act independently, i.e., the decisions of each are unaffected by those of the others. The decisions of different farmers are similar only insofar as the factors conditioning their decisions are similar. The dissimilarity of conditioning factors, however, is of particular importance. National land utilization patterns are the result of decisions made by individual farmers who possess different types of land and are subject to a diversity of other conditioning factors. A mass of existing data has been brought together for the purpose of using the theoretical analysis to make predictions of the land utilization patterns of individual decision-making units for the I955 crop year. Each decision-making unit is assumed to allocate its land among alternative uses in such a way as to maximize its expected return, subject to a number of technical and institutional constraints. Limited linearity assumptions are postulated, and the decision problem of each unit is expressed as a linear programming problem. Predictions of national and state allocation patterns are obtained by aggregating the patterns predicted for individual decision-making units. The present analysis is both descriptive and recursive. It is descriptive rather than normative in that the question asked is what will be done rather than what should be done. In this respect it differs from other current applications of linear programming to agriculture, which are intended to determine how individual farmers should allocate their land. It is recursive in that the values of the variables for a particular crop *This paper contains some of the results of a general study of natural resource utilization being undertaken as a part of the research program of the Harvard Economic Research Project. The author is indebted to the members of the staff of the project for aid in the preparation of this paper. He owes a particular debt of gratitude to the staff members who aided in the arduous task of gathering and rectifying the data. Mrs. Virginia McK. Nail provided expert help in all phases of this task. Jan Basch, Richard H. Day, Barbara King, and Ronald J. Wonnacott each made substantial contributions in one or more of its phases. An earlier version of the theoretical part of this paper was presented at a joint meeting of the Econometric Society and the Regional Science Association in Cleveland in December I956, and was reproduced in Papers and Proceedings of the Regional Science Associaotion, Vol. m.