This paper investigates and compares changes in the employment structure in manufacturing in Finland, France, Germany, Japan, the United Kingdom and the United States in the period 1975-1995. Using data on employment by skill level and type and several measures of an industry's technological advancement, we find a positive relationship between skill upgrading and R&D intensity. Since technology has some characteristics of a public good, it is important to include into the analysis spillovers explaining the changes in employment structure. Including knowledge spillovers lowers the estimates for R&D intensity substantially and shows a significant impact of knowledge spillovers on skill upgrading. In addition, splitting the sample into high-tech and low-tech industries reveals that the joint impact of spillovers in explaining skill upgrading in low-tech industries is more important than the impact of R&D itself. Furthermore, using three different definitions of skill, we find different point estimates for the impact of technology variables on skill upgrading. The results reveal that both the development of technology (through R&D) and its application (through spillovers) is skilled-biased, and not just the development as is often assumed in previous work. (C) 2002 Elsevier Science B.V. All rights reserved.