US PATENT CITATIONS DATA AND INDUSTRIAL KNOWLEDGE SPILLOVERS

The aim of this paper is to empirically evaluate the US interindustry knowledge spillover using the NBER patents data file (1963–1999). Reputing the patent backward citations as a good proxy of the patent's knowledge spillover, we proceed by building a time series to each US manufacturing industry patent citations and their lags. Then, we generate the time series of the external and the internal knowledge flow indices, showing that traditional sectors are more technology-dependent from the others than the new one. Here, in the spirit of Pavitt [Pavitt, Research Policy 13, 343–73, 1984]. We derive a new taxonomy of innovation focusing on the ideas instead of the goods production in order to obtain the innovation linkage and trajectories. Once we determined that each sector's most cited patents are typically belong to the ‘new’ sectors, we evaluate the high- and low-tech sectors innovation effect on the whole economy innovation process. Confirming that the high-tech sectors, and its R&D expenditures, are the most important, we conclude that it is the giant's shoulders, substance of the whole economy.