Multi-location Firms as a Medium for the Geographic Diffusion of Knowledge

• Main Author: Blit, Joel Nicolas
• Other Authors: Trefler, Daniel
• Language: en_US
• Published: 2010
• Subjects: multinationals; innovation; knowledge; geography; diffusion; 0501; 0310
• Online Access: http://hdl.handle.net/1807/26349

This thesis groups three papers examining the role of multi-location firms in the geographic diffusion of knowledge. The first chapter examines whether a firm's headquarters can tap into the knowledge pool in a remote location through FDI. Using U.S. patent data, I show that an R&D headquarters in location “A” cites third party patents from location “B” disproportionately (relative to a control group also from location “A”) when the firm has an R&D satellite in location “B”. This “satellite effect” on knowledge diffusion is economically significant, representing 47% of the knowledge flow premium associated with collocation. Furthermore, the effect is particularly strong for new knowledge, as well as in areas of satellite technological specialization. In addition, the results show that firms with stronger cross-location, intra-firm networks experience a larger satellite effect on knowledge diffusion. The second chapter studies the effects of remote satellites on outward knowledge flow with an international development focus. I find that the presence of a foreign MNC subsidiary increases the flow of knowledge from the MNC's headquarters to local firms. This effect is largest in countries and sectors with strong but not world-class capabilities, having both the motivation and absorptive capacity to learn from foreign parties. The results suggest that emerging country governments should promote inward FDI since the knowledge brought by multinationals spills over to local firms and boosts innovative capacity. The third chapter offers a theoretical foundation for thinking about the exchange of knowledge and ideas, and the role of remote satellites. I present a model where ideas are shared through social networks as modeled by repeated agent interactions. The mechanism at once explains three broad empirical findings: why the diffusion of ideas is highly localized, why ideas flow more easily within the firm, and why firms can access remote knowledge by establishing a presence in the remote location. A firm endogenously decides whether to establish a presence in a remote location and if so how much autonomy to award the remote agent. The relative importance of external vs. internal knowledge in the innovative process is a key determinant of the firm's organizational structure.