The production of innovation: Complementarity, organization, and strategy

This dissertation proposes and examines an empirical production function of innovation to gain insights on production itself, the disclosure of information in patents, organization choice and efficiency, and lastly the strategic consequences of different growth strategies. Using patent citations in the Information Technology industries, I measure the degree of complementarity among knowledge inputs. This measurement reveals the persistence of strong complementary inputs across both firms and time in the production of innovation. It is shown that the degree of input complementarity embodied in a patent significantly affects the value of the innovation. This relationship provides empirical insights on first mover advantage and information disclosure in published patents. Next, I measure complementarity in the downstream use of various innovation portfolios. This portfolio level analysis focuses on the selection choices and efficiency of different organization types, including firms, joint ventures, and licensing agreements. It is shown that output complementarity significantly explains both the selection biases and efficiency of these different organizations. Lastly, an agent-based computational model is constructed on top of the empirical patent landscape. Using simple rules-of-thumb, agents create innovations by acquiring and combining knowledge inputs. The agents receive value for these proposed combinations if and only if the combinations represent empirical patents. I use this model to reveal and analyze the strategic concerns with the timing of growth strategies. Also, the model provides insights on the impact of the publication of information, by the patent office, on innovation.