Composite value index of patent indicators: Factor analysis combining bibliographic and survey datasets

Abstract

The article relates to a patent value composite index that combines twenty different patent indicators according to several dimensions: patent breadth and technology potential, prior art and background of the invention, and filing and procedural aspects of a patent. A novel selection approach of patent indicators and their validation with market value of patents is advanced, whereas the computation of the composite value index is based on the factor analysis methodology firstly proposed in the literature by Lanjouw and Schankerman (2004).This study reports several new findings. Firstly, three common factors are obtained in correspondence to the respective three dimensions of the patent indicators. Secondly, the proposed composite value index can effectively summarize the information conveyed by every single indicator, because the reduction of the goodness-of-fit of the market value model is very limited as compared to the case of the indicators taken separately. Thirdly, a robustness analysis of the composite value index was conducted, relying on post-grant information, such as opposition and renewal decisions, and the results are consistent with the market value model with the composite value index solely considered. More generally, these findings contribute to the research agenda on proposing novel timely indicators of innovation activities.