In this paper, we investigate the extent to which the introduction of a negative externality influences the design of the optimal patent policy. First, we address the issue of structuring the value of the patent in the most socially efficient way. As we argue, the trade off between length and breadth should not only be driven by the impact that industrial protection has on the deadweight loss but also by its influence on the damage generated by the negative externality. Broader patents may indeed contain that damage more effectively by decreasing the equilibrium output or by making non-infringing imitations convey a smaller share of the external effect. Unlike Gilbert & Shapiro's findings, we show that narrow and infinitely-lived patents may not be optimal even though breadth is increasingly costly in terms of deadweight loss.
We then compute the optimal patent structure when the concept of breadth can include multiple dimensions. We alternatively consider breadth as the fraction of the optimal royalty fee that the patentee is allowed to charge by the public agency, as the cost of inventing around the patent and as the fraction of a product or process improvement that can not be captured by imitators without infringing the patents. Our conclusions suggest that the optimal patent structure is closely related to the way we interpret its breadth, to the size of the negative external effect and to the extent to which broader patents may make non-infringing imitations less harmful.
Finally, we turn to the question of how much to reward innovators whose R&D strategy determines the size of the negative external effect. We investigate two types of patent policy: a standard policy in which the value of the patent is the same regardless of the firm's R&D strategy and a modulated policy in which public agency may reward safe R&D strategies by giving stronger protection to the innovator. We establish that the latter type of policy is strictly more efficient than the former whenever the size of the externality is sufficiently large so the public agency actually takes advantage of the flexibility allowed by the modulated policy.