The motivation for this paper was to highlight the benefits of research cooperation as well as to depict the influence and drawbacks of changing spillover rates. Firstly, I introduced the well known linear-quadratic, 2-stage Cournot model to illustrate the effects of an increasing spillover rate on research output and size of the resulting coalition. Poyago-Theotoky (1995) assumes that coalition members can internalize the fruits of their research by entirely sharing their data and information in order to take better decisions and introduce better policies. In contrast, the spillover rate from the coalition members to the non-members varies between 0 and 1. Thus, the achieved cost reduction and R&D Output of coalition members will always exceed the output of non-members and also outweighs the Situation where no Cooperation takes place. A low spillover rate, Strategie Substitutes, to non-members decreases their competitiveness. In this case coalition members gain higher individual Profits than in the presence of high spillovers. Consequently, full cooperation is hardly possible because coalition members are unwilling to open up their coalition to a less productive agents in apprehension that the average payoff will decrease. When spillovers take the form of Strategie complements and are close to one, the competitiveness of outsiders does not lag behind significantly and a higher level of Cooperation and eventually even full cooperation is possible. From a social welfare point of view, this development is desirable since the socially optimal payoff can only be achieved by the grand coalition. The increasing spillover rate reflects the trade off between a high individual payoff and low social payoff on the one hand (when the spillover rate is low), and broad participation, high social payoff but relativley small individual payoff on the other hand (when the spillover rate is high). Full cooperation is only an equilibrium outcome when ß = 1. According to this reasoning Poyago-Theotoky (1995) concludes that for any spillover rate ß < 1 the size of the coalition settles at k </- n; research coalitions are of entry blocking type and refuse membership to less productive countries. To achieve a higher level of Cooperation a system of incentives has to be set up to compensate the initial coalition members for allowing more countries to join the coalition or for increasing the spillover rate to one or close to one. Subsidies could be an Option to compensate members of the equilibrium size coalition for the expansion of the coalition or the increasing spillover rate and to increase the private gains of coalition members. Equally it can be concluded that GEOSS needs some pioneering countries, which increase their spillover rate to a higher level to trigger the dynamics of coalition formation and eventually reach full cooperation. On the basis of the model of Carraro and Siniscalco (1995, 1997), linkage of negotiations can be proposed as compensation mechanism for the incurred costs of expanding the coalition. This additional agreement could take the form of a trade agreement, which allows the countries to generate extra benefits. Another proposition, which has been introduced in the second part of this discussion paper, was to change the rules of coalition formation and advocate the formation of multiple coalitions instead of only one partial coalition. After all, countries have different interests and needs, so why not proposing various coalitions for each field of interest. The results show that even by changing the rules of the game, full cooperation is hard to achieve. However, it might be possible to contract more countries to research Cooperation and have more coalitions in the equilibrium, which share data and generate better information to produce more environmentally friendly policies. These concerns have been investigated by Yi and Shin (2000). They transform the linear-quadratic Cournot club good model to a linear-quadratic public good model which displays a homogenous spillover rate for coalition members as well as outsiders. Whereas Poyago-Theotoky (1995) applies varying spillover rates and speaks of an entry blocking coalition, Yi and Shin (2000) assume that non-members refuse to join the coalition because they are better off as free-riders. Yi and Shin (2000) assume a high spillover rate, with an inter-coalition spillover rate equal or nearly equal to the intracoalition spillover rate and describe a Situation when Strategie complements are at work. Freely interpreted this implies that the individual payoffs are influenced by the sum of implemented policies and decisions and not by the composition of other countries research, which is reflected in the size and form of spillovers. Smaller coalitions implement fewer policies and incurr fewer costs but still internalise the gains of the higher research effort and policies of the larger coalitions. Yi and Shin (2000) introduce a set of conditions on the payoff function and introduce two games with either Exclusive or Open Membership rule. Neither of the games achieves a grand coalition for N >/- 5 due to high free-riding incentives. The Open Membership game, all agents can join whichever coalition they like, leads to a rather Symmetric coalition structure. Freely interpreted, this could imply that rather diversified coalitions with heterogenous agents emerge. In contrast, in the Exclusive Membership game coalitions are allowed to restrict membership, which allows them to select agents according to similar needs and ressources, for example agents who have to fight problems of desertification, or possess and manage rain forest areas. This practice allows them to concentrate their research and avoid the scattering of results by a less specialised country. Even though these results are instructive for the design of future research agreements, they have to be interpreted with caution due to the simplified framework. The results of this discussion paper suggest that the GEO secretariat could increase participation by (i) intervening in the coalition formation process and changing the rules of the formation process (e.g. by restricting membership to certain coalitions to a set of agents), (ii) offering compensations for the coalition members which contribute to the public good (e.g. offering an adequate cost recovery System), or (iii) linking GEOSS to an external agreement which allows members to generate additional benefits. These measures require that the GEO secretariat has enforcement or santctioning powers. This is not the case since participation to GEOSS is voluntary and non-legally binding, and the GEO secretariat serves as facilitating and coordinating instance to realise this 'system of Systems'. Consequently, full participtaion to GEOSS depends on the willingness of a set of pioneering countries and organisations who fully engage in the negotiations and release their data and information. These agents are aware of the fact that their effort and trust can contribute to greater social welfare.
