This paper analyzes the results of the climate-energy-economy model, Regionalized Model of Investment and Technological Development (REMIND), to assess the regional costs of climate-change mitigation for reaching the 2°C target with a medium to high likelihood. We assume that the global climate regime remains fragmented until 2020 after which a global mitigation target is adopted. We decompose the regional mitigation costs into (a) domestic and energy trade effects and (b) permit trade effects. Delaying cooperative action affects domestic costs by increasing the energy system's costs as a consequence of lock-in of carbon-intensive infrastructures. This is particularly true in developing countries with low near-term emissions reduction commitments. In a global cap-and-trade system, the effect of delayed action highly depends on whether or not the regions are over- or under-allocated with emissions allowances in the long term. Those with allowances exceeding their long-term emissions will likely benefit from the delay, while others suffer the consequences of higher long-term carbon prices.