The Rate of Depreciation of Technological Knowledge: Evidence from Patent Renewal Data

This paper is critical of studies that assume the rate of depreciation of technological knowledge is exogenously given and constant. It argues that the development of rival inventions and/or the existence of a pool of inventions from which spillovers take place impact directly on the size and value of the stock of knowledge. Patent data seem ideal to test such hypotheses as patents represent a store of R&D knowledge, and the declining value of the exclusive right to use an invention is reflected in the failure to renew patent protection. The empirical model includes not only rival invention and spillover effects, but other variables suggested by the existing literature, such as renewal costs and investment activity (representing new niches for inventions). A new quasi-panel patent data set has been constructed, tracing the survival characteristics of each cohort of patents over the period 1950-75. The data allow the first empirical tests of whether the hazard rate from the patent stock is duration dependent, which we demonstrate is linked to the highly skewed distribution of the value of patents. The long sample period also allows an exploration of whether the influences on the obsolescence of technological knowledge have changed over the post-War period.