Because of the doubts about the effectiveness of the Kyoto Protocol, several scholars have asked whether other types of agreements can be designed to achieve large reductions of GHG emissions. Designing a profitable and stable international environmental agreement (IEA) that deals with the shortcomings of Kyoto-type agreement is the main motivation of this work. One idea would be to focus on technology improvements in order to reduce abatement costs, as this might increase a country's willingness to undertake significant emission reductions. For example, it could be beneficial to supplement a Kyoto-type agreement with technology elements if technological development depends not only on a country's own R&D investment but also on R&D by other countries through cross-country technology spillovers, see for instance, Carraro and Siniscalco (1997). Even with no explicit agreement on emissions, a technology agreement leading to increased R&D in clean technologies, and thus to lower abatement costs, might yield a reduction in emissions. This is the argument behind the proposals of a climate agreement on technology development, see for instance, Barrett (2006). The aim of the present work is to examine different types of international environmental agreements in order to determine what would be the dominant agreement (at the different levels of marginal damages) with respect to both the total costs of signatories countries and the level of cooperation. The model used in the analysis is a three-stage static model (explained in chapter 2), where the membership game is played in the first stage, the investment game is played in the second stage and finally the emission game is played in the third stage. Examining the international cooperation on technological development as a supplement to international cooperation on GHG emission reductions, is the main objective of the second chapter where four different types of emission agreements are analyzed. The four agreements share the main aspect that signatories countries act cooperatively in the third stage of the game (emission game). However, the second stage of the game (investment game) differs from one type to another depending on whether signatories are sharing R&D efforts and coordinating their R&D activities or not. In all agreements, it is assumed that effective investment in one country depends on the amount invested in R&D in that country as well as on the investment in R&D undertaken in all countries through technological spillovers. In the types of agreements that include information exchange (emission agreement with information exchange and emission and R&D agreement with information exchange), the technological spillovers is perfect among signatories, which means that signatories countries avoid the duplication of R&D efforts. However, the technological spillovers is not perfect among signatories in the other two types (emission agreement and emission and R&D agreement without information exchange). Examining the international cooperation on technological development as an alternative to international cooperation on GHG emission reductions, is the main objective of the third chapter where three different types of technological agreements are analyzed. The three different types of agreements (R&D agreement without information exchange, research joint venture agreement and R&D agreement with information exchange), share the main aspect that signatories countries act non-cooperatively in the third stage of the game (emission game). A comparison between all types of agreements (analyzed in second and third chapters), is also introduced in this chapter. An extension of the analysis which examines the robustness of the model for the dominant agreements at the high levels of marginal damages, is the main objective of the fourth chapter. The quadratic investment costs and the quadratic environmental damages are the two different assumptions that have been analyzed in this chapter. Main Conclusions According to the analysis introduced in this thesis, the main conclusions can be summarized as follows ∙ Cooperating on emissions even if it is accompanied with cooperating on investment without information exchange, is not enough to eliminate countries incentives to act as free-rider. ∙ Sharing R&D investments and avoid duplication of R&D activities is enough to stabilize the grand coalition at the high levels of marginal damages. This result is explained by the asymmetry in the spillovers parameter between signatories and non-signatories. As the agreement, which includes information exchange, implies that signatories fully internalize the spillover effects of their investments in R&D, signatories can eliminate emissions using less resources than non-signatories. The result is that the signatories' investment costs are lower than the non-signatories' investment costs and hence the total costs are also lower. Moreover, there are negative externalities for non-signatories stemming from cooperation, i.e. cooperation increases the total costs of non-signatories. In this framework, if one country abandons the grand coalition, its total costs increase because of the increase in investment costs, which makes the grand coalition stable. ∙ As far as signatories countries invest at the maximum level of R&D investment to eliminate completely the GHG emissions, cooperation in the third stage of the game (emissions) does not affect neither the level of cooperation nor the total costs. ∙ At high levels of marginal damages, it is found that the agreements which include both cooperation on R&D investment and information exchange are the dominant agreements. However, for low values of marginal damages, the dominant agreements are changing according to the different intervals of marginal damages as explained in details in the third chapter. ∙ By examining the robustness of the different assumptions of the model, it is found that both of the assumptions of constant returns to scale of the R&D investment (linear investment costs) and the assumption of linear environmental damages are not critical for achieving the result that grand coalition is stable and profitable, at the high levels of marginal damages.