The Rate of Obsolescence of Technical Knowledge--A Comment

Abstract

THE main point of Dr. Bosworth's note is to show that the rate of obsolescence of technical knowledge is not a constant as I assumed in my estimates of technical knowledge in Schott [2]. This point is of course correct. In fact, it should be stressed that any attempt at estimates of technical knowledge are suspect for a multitude of reasons, some of which are discussed in my original article, and I am happy to see Dr. Bosworth drawing attention again to one of the problems. Indeed the main defence for attempting to estimate technical knowledge is that such estimates avoid the assumption that technology increases at some constant exponential rate. This latter assumption is very common despite general recognition that it is incorrect. Nevertheless, whether or not technical knowledge estimates (and their associated difficulties) represent an improvement over the constant exponential growth of technology assumption remains to be seen. My present feelings are that for empirical work based on short and medium term economic models the use of technical knowledge estimates may represent an improvement, despite the inherent estimation problems; but for such things as long-term growth models the constant exponential growth of technology assumption may be preferable. In any case, the main point that should be incorporated in all models, whatever their time horizon, is that technology is not costless. Unfortunately, Dr. Bosworth's estimates of the rate of obsolescence cannot be used to revise the technical knowledge estimates presented in Schott [2]. There are several reasons for this. First, as he points out, we both use different definitions of technical knowledge. My technical knowledge concept refers to some body of techniques which are actually being used in production or sold on the market. These techniques may be embodied in machinery or people or products, or just be some method of doing things. Once these techniques are no longer being used or sold they are termed obsolete. On the other hand, under Dr. Bosworth's definition, obsolescence occurs when a particular technology is no longer embodied in goods currently being sold. A measure of this obsolescence rate is derived from patent renewal data. It is not clear that patent renewal behaviour will reflect the obsolescence rate as defined by Dr. Bosworth, and his estimates, as I understand them, strictly explain the decay rate of patents. Various non-trivial assumptions are required for the decay rate of patents to be used as a measure of technical obsolescence in the sense that obsolescence occurs when a particular tech