Climate modelling with endogenous technical change: Stochastic learning and optimal greenhouse gas abatement in the PAGE2002 model

This paper looks at the impact of ETC on the costs and benefits of different abatement strategies using a modified version of the PAGE2002 model. It was found that for most standard abatement paths there would be an initial "learning investment" required that would substantially reduce the unit costs of CO2 abatement as compared to a business as usual scenario. Furthermore, optimising an abatement program where ETC has been included leads to an increase in cost uncertainty during the period of widespread CO2 abatements due to our lack of knowledge of the learning investments involved. Finally, the inclusion of ETC leads to a slightly deferred optimised abatement path followed by a rapid abatement program. Together, the results draw attention to the possibilities of 'uncovering uncertainty' through proactive abatements. 'Learning about learning' could become an important consideration for any plan to optimise future abatements. (c) 2006 Elsevier Ltd. All rights reserved.