An effective climate change treaty must promote the joint supply of two global public goods: climate change mitigation and knowledge of new technologies that can lower mitigation costs. R&D is especially needed to bring about substantial, long-term reductions in atmospheric concentrations of greenhouse gases, for this will require the development and diffusion of revolutionary, “breakthrough ” technologies (Martin I. Hoffert et al. 2002). In principle, such an outcome could be realized by the Kyoto Protocol approach, if that agreement were strengthened over time. However, that response may be inadequate (Kyoto makes no provision for R&D)—and, as I shall demonstrate, unlikely to succeed in any event. Can a treaty system relying directly on targeted R&D and the adoption of breakthrough technologies perform better in this same setting of anarchic international relations? I shall show that, as a general rule, the answer is no. Essentially, the same forces that undermine Kyoto also challenge the R&D and technology approach. There is one exception to this rule: R&D leading to breakthrough technologies exhibiting increasing returns can improve dramatically on the Kyoto approach, even when these technologies are otherwise inferior to the alternatives available. This suggests that our approach to treaty design should be strategic. 2I. The Kyoto Approach Begin by considering the abatement decisions of countries in the absence of a multilateral agreement. Let qi denote country i’s abatement and let Q denote aggregate abatement; with N countries, Q = qi i=1 N Â. Finally, let country i’s payoff be given by p i = bQ- c qi