Bayesian Learning and Regulation of Greenhouse Gas Emissions

Abstract

We study the importance of anticipated learning - about both environmental damages and abatement costs - in determining the level and the method of controlling greenhouse gas emissions. We also compare active learning, passive learning, and parameter uncertainty without learning. Current beliefs about damages and abatement costs have an important effect on the optimal level of emissions. However, the optimal level of emissions is not sensitive either to the possibility of learning about damages, or to the type of learning (active or passive). Taxes dominate quotas, but by a small margin.