R&D competition and the persistence of market power: Theory and evidence from the chemical industry

This dissertation presents a theoretical and empirical framework for analyzing dynamic R&D competition and the evolution of market power in an oligopoly setting. A particular emphasis in the model is the conditions that induce a dominant firm to maintain or extend market power.; R&D is modeled as a two-step search process that improves the firm's technology level and captures intra- and interfirm learning under horizontal technology differentiation and research “dead-ends.” The research process encompasses past assumptions of the role of technology in current competition as special cases. Competition consists of a repeated two-stage R&D-then-market game with a unique Markov Perfect Nash Equilibrium (MPNE) in research strategies.; I find that the technology gap has three possible growth paths: shrink, display bounded growth or grow without bound. The bounded growth path is a new finding that demonstrates a limitation on the expected extent of dominance held by the leader. I also find a paradox between static market efficiency and the persistence of market dominance: as the market structure becomes more competitive, the larger firm has an increased incentive to remain dominant.; I then illustrate the theoretical findings by directly estimating the determinants of market evolution in the organic chemical industry. In contrast to previous studies, my estimation does not depend on unobservable research expenditure data. Instead, research expenditure is computed as the solution to the MPNE in research strategies. More importantly, the empirical model is constructed to allow for unobserved firm-specific pre-sample patenting and productivity differences. The model also allows for time-specific changes to the probability of research success from exogenous productivity growth and patenting by non-competitors. Estimation is an iterated process of estimating parameters and approximating optimal expenditure.; The results suggest that the organic chemical industry conforms to the bounded growth case where firms learn rapidly and/or conduct highly differentiated research. They also suggest the presence of unobservable research productivity differences that favor large firms and, a small technological barrier to entry. As an illustration of the versatility of the model, I provide a simulation of the proposed Dow-Union Carbide merger using the model as a framework for analysis.