Extending the Frontier: A Structural Model of Investment and Technological Competition in the Supercomputer Industry

A dynamic structural model of innovation in the super- computer industry to evaluate the dependence of technological innovation on market structure shows increased levels of competition are associated with a higher rates of innovation and increased welfare, however, the marginal increase in welfare is decreasing in the number of competitors.

Abstract

How does technological progress depend on competition through investments in innovation? This paper proposes and estimates a structural model of dynamic investment and competition in the supercomputer industry to evaluate the dependence of technological innovation on market structure. Building on recent advances in the structural estimation of investment games, I develop a tractable yet dynamic model of the benefits and costs of innovation, including the dependence of innovation cost on the current technological position of the firm and accounting for the dynamic benefits of technological leadership. The model has two key features. First, it incorporates the possibility of technological leapfrogging commonly found in high-tech industries while controlling for multiproduct firm profits. Second, it uses the adjusted inclusive value metric of Nevo and Rossi (2007) to define quality investment. The latter allows to control for both entry/exit and new product launch without modelling these decisions explicitly. Model estimates facilitate counterfactual comparisons of how the evolution of the maximal computing speed supplied in the supercomputer industry differs under different market structures. Consistent with the importance of a technology “selection effect” (Aghion et al, 2001), increased levels of competition are associated with a higher rate of innovation in the supercomputer industry. Increased competition is also associated with increased welfare, but the marginal increase in welfare is decreasing in the number of competitors.