Identification of technological structures using patent statistics

This paper studies the use of patent statistics in identifying four aspects of technological structure, namely, the potential knowledge pool, cumulativeness, inter-firm homogeneity in technology levels, and the scope of innovations. The firms are sampled from the chemical (CHEM), the computer (COM) and the electrical and electronic (EE) industries worldwide. Using the proxies defined, we find that (i) the contributions of intra-industry spillover are low, at 12%, 10%, and 9% for the three industries respectively; (ii) they can internalize 15%, 19% and 13% of their previous research efforts respectively; and (iii) a positive relationship between knowledge spillover and technology overlap, and between scope of innovation and number of patents being cited in future.