Constructing intersectoral innovation diffusion networks with input-output: how to get relative flows? An illustrative application to six OECD technological systems for the middle '90s

The paper aims at addressing some methodological issues in applying Social Network Analysis indicators to the investigation of nation-wide intersectoral innovation ows matrices. The majority of SNA techniques require dichotomization of the original matrices and suitable relativization procedures, in order to avoid size-biases. The relativization procedures used so far suffer from some limitations, as they either alter the meaning of SNA indicators or do not take into account the composition of countries' final demand. In order to overcome these limitations, we propose two new different methods and compare them with the existing ones on the basis of their rationale. Rather than with respect to a single cut-off, the comparison is carried out by working out SNA indicators distributions. An illustrative application is carried out by comparing the technological systems of six structurally different OECD countries in the mid-'90s. In so doing, the robustness of their conventional innovation ranking is tested and discussed.