This paper studies policy instruments that correct insufficient learning-by-doing (LbD) and research and development (R&D) of renewable electricity technologies and insufficient investments in energy efficiency (EE) in the presence of carbon pricing. The theoretical model analysis shows how to re-adjust the first-best in second-best situations, in which one of the policy instruments is restricted. Calibrated to the European power sector, the first-best choice of all instruments reduces the climate policy cost by one third. Feed-in tariffs turn out to be good substitutes for LbD, but not for R&D or EE subsidies.