Estimating technical change by nonparametric methods

Abstract

A nonparametric method using dynamic data envelopment analysis is developed here to specify and estimate a dynamic production frontier. By this technique technological progress measured by the time shift of the production frontier and the scale economies measured by increasing returns to scale are estimated from the observed input output data, which may sometimes be nonstationary. An application to US computer industry exhibits substantial technological progress and scale economies over the period 1987–1998.