SIMPLE NONPARAMETRIC TESTS OF TECHNOLOGICAL CHANGE: THEORY AND ApPLICATION TO U.S. AGRICULTURE

Several nonparametric tests of technological change are proposed conditional on alternative maintained behavioral assumptions. The tests are simple as they require verification of axioms via binary comparisons that are analogous to those of the revealed preference theory. Our model allows a broad interpretation of technological changes, yet trivial rationalization of the data with either profit-maximization or cost-minimization behavior is excluded. The tests are applied to study the U.S. aggregate agricultural data, supporting the hypotheses of rankable and neutral Technological Variations. Key words: neutral technological variations, nonparametric analysis, rankable technological var-iations, revealed technological superiority. Nonparametric analyses of production data are aimed at testing the consistency of ob-served behavior with a particular optimization rule and certain technological restrictions, with no need for maintained assumptions about the form of the production function. These nonparametric methods have been de-veloped by Afriat, Hanoch and Rothschild, and Varian (1984). They recently gained spe-cial popularity in analyzing technological changes and productivity in general, and in the agricultural sector in particular (e.g., Faw