Regional Climate Policy under Deep Uncertainty: Robust Control, Hot Spots and Learning

We study climate change policies using the novel pattern scaling approach of regional transient climate response, to develop a regional economy-climate model under conditions of deep uncertainty associated with: (i) temperature dynamics, (ii) regional climate change damages, and (iii) policy in the form of carbon taxes. We analyze cooperative and noncooperative outcomes. Under deep uncertainty, robust control policies are more conservative regarding emissions, the higher the aversion to ambiguity is, while damage uncertainty seems to produce more conservative behavior than climate dynamics uncertainty. As concerns about uncertainty increase, cooperative and noncooperative policies tend to move close together. Asymmetries in concerns about uncertainty tend to produce large deviations in regional emissions policy at the noncooperative solution. We calculate the cost of robustness in terms of welfare. If aversion to ambiguity is suciently high, optimal regulation might not be possible. The result is associated with the existence of regional hot spots and temperature spillovers across regions, a situation which emerges in the real world. In such cases, deep uncertainty about the impacts of climate change makes robust regulation infeasible. We show that, if resources are devoted to learning, which reduces uncertainty concerns, robust regulation is facilitated.