An empirical examination of the value of patented inventions using German employee inventors' compensation record

The study of technological change and innovation has been seriously hindered by the scarcity of appropriate measures of innovative output. This dissertation examines the returns to patented inventions using a novel data source: the compensation records of German employee inventors. In Germany, compensation for employee inventors is mandated by the German Employee Invention Act of 1957. The guidelines of the act stipulate that employee's compensation is to be proportional to the private asset value of the invention to the firm. To comply with the act, the firm must calculate the contribution to firm profits of every invention made by an employee inventor. The resulting employee inventor compensation records not only provide a novel source of estimates of invention value, but also provide the timing of the returns for the patented invention.;This dissertation uses estimates of invention value from each invention's employee inventor compensation record to construct the largest dataset of individual estimate of patent values yet assembled. The dataset consists of information on 1172 inventions originating from Germany during the 1977--1982 period. It is restricted to inventions that were granted a U.S. utility patent, in order to make use of information from U.S. patent records. Using these data has enabled me to conduct the first research that explores the value of an individual patented invention along two dimensions: the total returns earned over the patent's life and the timing of the returns.;The paper finds that the returns earned by most patents dissipate rapidly, with only a small fraction of total returns earned at the end of the patent term. The vast majority of the returns are earned by the highest value patents, which can be identified relatively early in the patent term. These findings suggest that shortening the patent term (with the possible exception of patents for pharmaceuticals) or moderate maintenance fees would have little impact on the incentives to innovate. Several new explanatory variables are developed that proxy for the strength of patent protection and the scope of the patent's claims, which prove to be highly informative for predicting value.