J.P. Eggers, Stern School of Business, New York University
Administrative Science Quarterly, March 2012, Vol. 57, No. 1, pp. 47-80
Abstract: This study theorizes about the behavioral and knowledge creation implications of betting on the losing technology in a competing technology situation and focuses on three main outcomes. First, in a situation with competing technological options, firms that invest initially in the losing technology will be less successful subsequently in building new knowledge in the winning technology because their experience with failure will lead them to update their expectations of the industry and choose to pursue less risky alternatives. Second, two classic risk-reducing strategies—investing in both technologies or entering after uncertainty is resolved—will not be completely effective. Firms investing in both technologies are likely to suffer the incentive and coordination-driven innovation penalties of generalists, while late entrants will suffer learning disadvantages. Third, the possession of key and relevant complementary assets—upstream and downstream—will positively moderate the observed inertial effect on firms that backed the failed technology and generalists that backed both technologies, as these complementary assets will increase incentives to adapt to the winning technology. I find empirical support for my hypotheses using a novel data set on the evolution of the global flat panel display industry from 1964 to 2003 to investigate the technological competition between plasma and liquid crystal displays. Results show that firms initially pursuing plasma generated less subsequent knowledge in liquid crystal displays, and that firms betting on both technologies were also slow to build knowledge in liquid crystal displays. Meanwhile, firms with upstream and downstream complementary assets were able to moderate, but not overcome, this barrier to knowledge creation. The findings have implications for our study of technological evolution and adaption, for learning from failure and reinforcement learning, and for the relationship between resource partitioning and adaptation.