A theory of socially inefficient patent holdout

Abstract

This paper proposes a framework to analyze holdout in patent licensing negotiations. We show that when the validity of a patent is probabilistic, a potential downstream user has incentives to shun to pay the price offered by a patent holder to license the technology and risk being brought to court. These incentives are exacerbated when jurisdictions are local, and the downstream producer can approach courts sequentially. The informational spillovers across trials imply that this firm often finds optimal to go to court aiming to invalidate the patent in a jurisdiction due to the knock-on effect on future jurisdictions. This process results in excessive litigation compared to when the jurisdiction is global. The distortions from sequential litigation are likely to be aggravated when final competition is accounted for or when patent injunctions are not allowed.