Knowledge and productivity in the world's largest manufacturing corporations

w This paper develops a model I inking firm knowledge with productivity. The model captures three characteristics of firm knowledge (capital, diversity and relatedness) that are tested on a sample of 156 of the world's largest corporations. Panel data regression models suggest that unlike knowledge diversity, knowledge capital and knowledge relatedness explain a substantial share of the variance of firm productivity. Relatedness matters because it lowers coordination costs between heterogeneous activities. Consequently, the traditional econometric specification has repeatedly underestimated by 15 percent the overall short-run contribution of intangible assets to firm productivity. This underestimation becomes fiercer in high-technology sectors. (C) 2007 Elsevier B.V. All rights reserved.