Technological change and the timing of mitigation measures

We use a coupled carbon-cycle and energy systems engineering model to analyze the future time path of carbon emissions under an illustrative CO2 concentration stabilization limit of 550 ppm. Our findings confirm the emission pattern as found by WRE: global emissions rise initially, pass through stabilization, in order to decline in the second half of the 21st century. We show that for a given CO2 concentration target, emission trajectories within an intertemporal optimization framework depend mainly on two factors: the discount rate, and the representation of technological change as either static or dynamic. We obtain a similar near-term emission time path as WRE when using a model with static technology and a discount rate of 7%. We obtain a trajectory with lower emissions in the near-term when using a lower discount rate and/or treating technology dynamics endogenously in the model. We briefly outline a model that endogenizes technological change through learning curves. We then compare differences in emission trajectories between alternative model formulations of technological change. They are sufficiently small as to be of secondary importance when compared to treating CO2 concentration stabilization as an inter-temporal optimization problem or not. Whereas our results confirm the computational results of WRE, we arrive nonetheless at different policy conclusions. If long-term emission reduction is the goal, we cannot follow 'business as usual' even in the short-term. Action needs to start now. Action does not necessarily mean aggressive short-term emission reductions but rather enhanced R & D and technology demonstration efforts that stimulate technological learning. These are the necessary preconditions that long-term reduction targets can be met with improved technology and at costs lower than today. We close by pointing out two further critical issues: uncertainty, and the possible mismatch between the world of economic models and that of climate policy. (C) 1998 Elsevier Science B.V. All rights reserved.