Optimal timing of carbon capture policies under learning-by-doing

Using a standard Hotelling model of resource exploitation, we determine the optimal energy consumption paths from three options: dirty coal, which is non-renewable and carbon-emitting; clean coal, which is also non-renewable but carbon-free thanks to carbon capture and storage (CCS); and solar energy, which is renewable and carbon-free. We assume that the atmospheric carbon stock cannot exceed an exogenously given ceiling. Taking into account learning-by-doing in CCS technology, we show the following results: (i) clean coal exploitation cannot begin before the outset of the carbon constrained phase and must stop strictly before the end of this phase; (ii) the energy price path can evolve non-monotonically over time; and (iii) when the solar cost is low enough, an unusual energy consumption sequence along with solar energy is interrupted for some time and replacement by clean coal may exist.