Neoclassical Modeling of Nonneutral Technological Change: An Experimental Appraisal

The results of an experimental evaluation of the performance of neoclassical models, using alternative indexes of the pace of introduction of innovations to measure the effects of technical change on the patterns of factor input usage, are reported. The analysis maintains that production processes should be viewed as engineering activities, so that neoclassical characterizations of them and of any innovations to them are best treated as approximations. Two large-scale process analysis models provide the description of the "true" production technology with and without access to six specific process innovations. The analysis considered both input aggregation and the index of the extent of technical innovation. The results suggest that indexes which are explicitly tied to the technical features of the innovations involved perform better in isolating the direction of the input effects associated with each innovation. Moreover this conclusion was found to hold with differing levels of input aggregation.