The localization of technological knowledge in the international semiconductor industry

Is knowledge localized? Why? This thesis studies the spatial diffusion of technological knowledge through the analysis of patent citations of important semiconductor innovations. The study first examines the variation in the localization of knowledge (a) by country and region and (b) by technology. The thesis then investigates why knowledge is localized to particular geographic communities by looking at the factors and mechanisms which influence knowledge flows. Finally the study investigates the knowledge acquiring and sharing patterns of foreign firms located in the United States. The geographic characteristics of knowledge flows are analyzed through patent citation analysis using the case-control method. The factors influencing localization are investigated by applying regression analysis. Field research, consisting of interviews with engineers in the industry, is used to provide a rich context for the interpretation of the results of the patent analysis. The thesis finds that the localization of knowledge is not a \u27given\u27 but is the outcome of the interaction of the nature of technology and geographic regions. Semiconductor knowledge in the US (particularly in Silicon Valley) tends to be localized, while in Japan and Europe there is no evidence of localization. Design knowledge diffuses locally but fabrication knowledge exhibits no such effects. Data on the inter-firm and inter-regional mobility of engineers and entrepreneurs indicate that the flow of ideas embodied in engineers is a critical explanation of the variation in localization. The thesis also finds that foreign firms in the US use predominantly local knowledge in their innovative activities. In fact, foreign firms use local knowledge, significant!y more than domestic firms. The study also finds that foreign firms also contribute knowledge locally. Through these findings, this dissertation furthers our understanding of some of the very basic questions regarding the spatial dimensions of knowledge flows. The results of this thesis are important to our understanding of the meaning of borders, and therefore, of firm and country competitiveness