Technological substitution: A framework of stochastic models

A basic framework for a stochastic substitution model is developed. The model explicitly considers the number of decision makers responsible for the substitution process, and the volume of usage of the old and new technology. The parameters of the stochastic model are adaptively estimated using the extended Kalman filter. This approach is contrasted with the logistic differential equation approach, widespread in the literature, where only the proportion substituted is modelled. The stochastic model is shown to produce a considerable improvement in forecasting accuracy.