To build or not to build? Capital stocks and climate policy∗

We investigate (i) the impact of emission reduction policy on investment in polluting infrastructure, such as coal-fired power stations and (ii) optimal subsidies for “clean” alternatives with “learning” spillovers. We build a general theoretical model, and embed it in a fully calibrated integrated assessment model. Because emission reduction policy reduces investments in polluting assets, short-term emission reductions are enhanced—our “irreversibility effect”. Thus, “stranded assets” in this fuel-using sector have distinctive properties. We also provide a simple formula for how the optimal subsidy to deployment of a “clean” sector depends on its rate of “learning-by-doing” and on its socially-optimal growth. So, if the sector should grow faster for other reasons, its optimal subsidy is increased, showing that its optimal growth rate is faster still—our “acceleration effect”. Our calibrations show that, to limit global climate change to 2○C warming, investments in coal-fired power stations must end very soon. Considering second-best settings, we show that carbon taxes achieve stringent policy targets more efficiently, but subsidies to the “clean” sector deliver higher welfare, and are more efficient, when policy targets are more mild.