Summary/Abstract
This paper shows that commonly proposed emission taxes are not optimal for controlling climate change: they can achieve zero emissions but cannot induce negative emissions. The first-best policy charges firms period by period for leaving a stock of carbon in the atmosphere, not just for injecting carbon into the atmosphere. The author develops a version of this policy that requires emitters to post an upfront bond that finances a transferable asset (a “carbon share”). The regulator reduces this asset’s face value as damages accumulate and pays out the asset’s remaining face value once its holder removes the underlying unit of carbon from the atmosphere. The author shows that the optimal bond is equal to the maximum possible marginal damage from climate change, with the carbon share paying a dividend as long as the worst-case is not realized. Quantitatively, a bond that is double the optimal emission tax is sufficient to provide optimal carbon removal incentives in 95% of cases.