The difficulty of achieving a global agreement on climate change underlined in the previous sections depends on four main factors:

• The heterogeneity of countries with respect to the causes of climate change, the impacts, and the mitigation and adaptation costs. This factor mainly influences the profitability of the decision to sign a climate agreement. Some countries may lose when signing the agreement, even when environmental benefits are fully accounted for. As shown by Chander and Tulkens (1995, 1997), there always exists a system of economic and technological transfers that may make all countries gain. But this again raises the equity problem and the related burden-sharing issue. Equity may have a large impact on the existence and size of a climate coalition. As previously argued, and as argued by many policymakers and scientists, the way in which the burden of controlling emissions is shared across countries crucially affects a country’s decision to join a coalition. On the one hand, if the burden is not equitably shared, some countries may not find it profitable to sign the agreement. Profitability depends on two main factors: (1) the distribution of costs within the coalition and (2) the size of the coalition. It is possible that there exists a minimum size of the coalition above which it becomes profitable. And these two factors are strictly interdependent. On the other hand, equity also affects free-riding incentives. As in Section 10.2.5, in some cases it may be reasonable for some countries to transfer resources to other countries to induce them to join the coalition on which they would otherwise free-ride. In this case, the final outcome is not equitable–free-riders would gain more than countries in the starting coalition–but it may be environmentally and economically efficient.
  
• The strong incentives to free-ride on the global agreement and the lack of related sanctions. When all countries agree to control emissions, a defecting country achieves the whole benefit, because its incidence on global emission is marginal (with a few exceptions) and pays no cost. Hence, a defection with respect to a large coalition is the optimal strategy if there are no sanctions. However, credible sanctions are difficult to design (Barrett, 1994). Emissions themselves are hardly a credible sanction, because countries are unlikely to sustain self-damaging policies. Moreover, in this case, asymmetries play a double role: some countries may not gain from signing the environmental agreement, whereas some countries, even when gaining from environmental co-operation, may lose from carrying out the economic sanctions (Barrett, 1997c; Schmidt, 1997).
  
• The absence of environmental leadership. The process of achieving a global agreement can be a sequential one (Carraro and Siniscalco, 1993), in which case a group of countries take the leadership, start to reduce and/or control emissions and implement strategies such as to induce other countries to follow¹⁷. The presence of low-cost climate policies and equitable burden-sharing (Schmidt, 1997) are again important elements for the formation of an initial profitable coalition. As said, our definition of profitability accounts for the environmental benefits of emission control. Hence, benefits should be increased by increasing the number of countries that control emissions, but abatement costs should be minimized by exploiting all possible opportunities (including emissions trading). This is a prerequisite to achieving a strong leader coalition that can exert its leadership through the design of better negotiation rules, the implementation of transfer mechanisms, and the credibility of international-issue linkages. A preliminary model of the effects of leadership is given in Jacoby et al. (1998), who show how and when developing countries may join a leader coalition formed by Annex I countries.
  
• The focus on a single international climate agreement. As explained in Section 10.3.1, if countries may join different coalitions, which means that several agreements can be signed by groups of countries in the same way as countries form trade blocs, then the likelihood that all or almost all countries set emission reduction targets increases (Yi and Shin, 1994; Bloch, 1997; Carraro, 1997, 1998). The outcome of negotiations in which more agreements can be signed is usually a situation with several small environmental blocs (Carraro and Moriconi, 1998), but this can be considered a step in the right direction. If all or almost all countries set emission reduction targets within their own bloc (e.g., regional environmental agreements are signed), then, in a subsequent phase, negotiations among blocs may lead to more ambitious emission reductions.

Despite the warning that global agreements may be difficult to reach, many articles analyze the costs of agreements in which all countries participate, in one form or another (see, e.g., Capros, 1998, Ellerman et al., 1998; Manne and Richels, 1998; Shackleton, 1998; Bosello and Roson, 1999; Nordhaus and Boyer, 1999). The weakest form, discussed in Section 10.2.4, is that in which a few countries commit to emission reductions, but all accept trade emissions in a single international market. The strongest form is that in which a central planner is assumed to set optimal emissions levels for all world countries. This optimal solution is often proposed as a benchmark for actual negotiations and was often analyzed before Kyoto (see the collection of papers in Carraro (1999d)).

More interesting is the attempt made by Peck and Teisberg (1999) to model the negotiations between developed and developing countries to achieve a global agreement. This paper shows the potential for the achievement of co-operation to be achieved–the Pareto frontier is small, but not empty–but does not analyse the incentives to actually sign the agreement. However, the paper suggests a research direction that at least helps to identify the optimal emission reductions that are profitable for all negotiating countries.

The conclusions that can be derived from this type of empirical analyses are similar to those already mentioned for partial agreements. In the scenario in which baseline emissions are lower, it is easier to achieve a global agreement because lower emissions reductions are necessary (Barrett, 1997b) and consequently abatement costs are lower. Optimal emissions targets are such that they equalize marginal abatement costs. This optimal, cost-minimizing solution can also be achieved through an unconstrained emissions-trading system (Chander et al., 1999). Hence, either emissions targets are optimally set, or countries are allowed to trade emissions for any given set of targets through which a global consensus can be achieved. Of course, these two options have different impacts on equity. As shown by Bosello and Roson (1999), starting from the Kyoto targets, international unconstrained emissions trading among all countries achieves optimality, but reduces equity.

IPCC Homepage