Essays in Industry Dynamics and Innovation: A Computational Approach

Abstract

This is a collection of essays related to dynamic economics on the boundary between Industrial Organization and microeconomics. Chapter one examines optimal patent design, and constructs a dynamic oligopoly model in which firms innovate along a quality ladder. Firms make decisions about both R&D investment as well as litigation spending, both of which drive profits. This is the first paper which studies patent policy in a dynamic setting that includes endogenous market structure and patent litigation. The model incorporates three observations about patent policy: (1) patent protection can have pro- and anti-competitive effects in a dynamic model; (2) private and social litigation costs are not second-order and should be modeled along with R&D; (3) policy mix matters and policy instruments cannot be studied independently. Industries characterized by higher variable R&D costs should have lower patent duration and stronger forward exclusion. Forward exclusion should be constant and patent duration should be decreasing in the market size for new innovations.

The second chapter develops new computational methods for the computation of the Abreu-Pearce-Stachetti equilibrium payoff set in games with imperfect monitoring. These games are of great interest, but are generally intractable computationally. The algorithm uses a grid search approach to plotting out the contours of the APS set. Theoretically, the proposed algorithm improves on the standard Judd-Yeltekin-Conklin algorithm by giving precise error bands on the difference between the computed approximate APS operator and the true APS operator, and by being more easily generalizable to more than two players. Testing show speed advantages for the grid algorithm in practical applications.

The final chapter returns to the patent question, and studies optimal patent duration in a dynamic oligopoly. The innovation process distinguishes between product quality innovations and process innovations, which have different implications for patent enforcement. I discuss how optimal patent duration, T* and industry concentration N(T*) vary by industry. Product lines with high entry/fixed costs should have lengthy patent durations and small firm concentrations. Those with rapid innovation rates or high probabilities of breakthroughs generally warrant shorter patent duration and lower market concentration.