The concept of ＇joint implementation＇ is the subject of ongoing negotiations under the Framework Convention on Climate Change. Proponents argue that allowing joint implementation of the objectives of the Convention will increase the cost-effectiveness with which those objectives are achieved. Opponents argue that allowing trade in carbon emission credits will reduce the incentive for domestic greenhouse gas emissions reductions, and may undermine the commitments of the developed nations to take the lead in achieving climate stabilization. This paper examines the argument for joint implementation from the standpoint of cost-effectiveness. It addresses the determination of incremental costs for abatement options, and the assessment of cost-effectiveness under complex systemic conditions. Using national greenhouse gas abatement costing studies from four different nations - two developed nations, on economy in transition and one developing nation - the paper examines both the general claim for cost-effectiveness of joint implementation and the implicit assumption that emissions reductions will be easier and cheaper in developing nations and economies in transition than they are in developed countries. The results indicate firstly that benefits from joint implementation may be highly dependent on the level of associated transaction costs and secondly that greenhouse gas emission reductions may be considerably cheaper for some developed countries than they are for developing countries. The paper stresses the need for well developed methodological guidelines under which both cost and cost-effectiveness can be assessed, and points to the dangers inherent in allowing ad hoc trading in emissions credits by a heterogeneous community of private investors.