Measuring generative appropriability: Experiments with US semiconductor patents

Abstract

This study presents a novel approach to quantify Generative Appropriability (GA) – a firm's capability to exploit current inventions to generate follow-on inventions (cumulative GA) and preclude rivals from them (preclusive GA). Unlike prior research, our approach relies not only on patent citations but also on texts. In the first step, we preprocess the texts of the focal patents and their citations. In the second step, we measure knowledge flow by similarity measurements. In the third step, we calculate cumulative and preclusive GA by means of constructed formulae. To test our approach, we select particularly active firms in patenting US semiconductor inventions. First, results from knowledge flow confirm our initial concerns on measuring knowledge flow by means of patent citation count. Second, results from GA show different patterns among the semiconductor firms and firms which are well-positioned in both perspectives and firms which lack cumulative GA but show good values in preclusive GA.