Optimal Rules for Patent Races

This paper analyzes optimal patent policy in a dynamic multi-stage innovation model where firms compete to receive a patent and its associated prize. Firms control their independent R&D project and invest to reach the stage at which a patent is awarded. At this stage, the laggard firms are forced to leave the race and the winner continues to invest in R&D until the innovation process has achieved its goal. Firms are assumed to possess perfect information about each others' innovation state and cost structures. A planner, who cannot distinguish between the firms, chooses the stage at which the patent is awarded and the magnitude of the prize to the winner to maximize either a ''social'' or a ''consumer'' surplus. We study the evolution of competition along the race path, explore the socially optimal patent policy and the sensitivity of R&D investment and competition to the prize level and the degree of heterogeneity between the firms.