All current, and likely near-term future, climate protection measures only cover a limited fraction of global emissions. A single value attached to (independent of the source that generates it), for market based instruments such as CO2 taxes or cap-and-trade systems, is insufficient to account for the complex economic interlinkages between specific emission-generating activities and CO2 emissions throughout the world. First, static partial and general equilibrium models illustrate how different types of emissions are subject to specific General Equilibrium Translation Factors and leakage effects, which define the optimal pattern of fuel-specific, unilateral carbon taxes. The leakage, which implies that regional emission avoidance may partly be offset in other regions and time periods, depends on the type of resources involved and the characteristics of the markets in which they are traded. Second, a dynamic model accounting for fuel exhaustibility shows that the time-dimension is crucial and that the relevant medium-term leakage may be much larger than suggested static rates. Sensible leakage rates depend on discount rates for future emissions and on uncertain future technological and political developments. The traditional leakage literature does not explicitly consider these, even though in their absence overall leakage would approach 100 %. Instead, literature has mainly focused on static fuel supply curves and rates of contemporaneous leakage.
The numerical simulations show that in a business-as-usual scenario the optimal unilateral OECD climate tax rate on CO2 emissions from oil may be only half of the tax rate on emissions from coal. This is reverted if the CO2 intensive coal-to-liquids conversion processes become an important additional source of liquid fuels in future: negative leakage occurs and the optimal current climate tax on oil emissions may be up to two times the genuine regional willingness to pay for global emission reductions, even if the substitution of crude oil by synthetic liquids starts only in the future.