This dissertation considers the theoretical aspects of countries' incentives to cooperate on environmental good provision and resolve free-rider incentives, in particular with the formation of an International Environmental Agreement (IEA).

Following a short chapter reviewing the weakly renegotiation-proof equilibrium, the third chapter, "Sustaining Full Cooperation in an International Environmental Agreement through Learning and R&D", analyzes the effects of learning and R&D on an IEA by assuming that countries are uncertain regarding the benefits of pollution abatement. This paper shows that uncertainty improves the likelihood of obtaining a Pareto-optimal IEA, which is constructed as a weakly renegotiation-proof equilibrium in an infinitely repeated game, by allowing for a wider range of discount factors to sustain cooperation than in the no uncertainty case. Finally, this paper analyzes whether or not new knowledge gained through R&D is beneficial for sustaining cooperation and finds that achieving the Pareto-optimal IEA tends to be less likely if R&D reduces uncertainty. The mechanism driving these results is that uncertainty leads to a higher expected net loss from punishment to a defecting country, which implies that deviations are better deterred under uncertainty than with no uncertainty.

In the fourth chapter, titled "Renegotiation-Proof International Environmental Agreements with Social Preferences", I turn from purely self-interested agents and examine how social preferences, in particular, preferences for equity and efficiency, affect the likelihood of cooperation among countries to abate pollution when compared to the case of agents that are only self-interested. It is shown that an IEA with any level of cooperation, including full cooperation, exists as a weakly renegotiation-proof equilibrium for high enough discount factors. As social preferences grow stronger, the range of discount factors that can support cooperation increases, which implies that cooperation is more likely under social preferences. The key effect driving this result is that social preferences cause the net loss from the punishment of a defecting country to increase, which better deters deviations from cooperation.