Ambiguity Reduction by Objective Model Selection, with an Application to the Costs of the EU 2030 Climate Targets

Richard Tol

doi:10.3390/en7116886

Abstract

I estimate the cost of meeting the EU 2030 targets for greenhouse gas emission reduction, using statistical emulators of ten alternative models. Assuming a first-best policy implementation, I find that total and marginal costs are modest. The statistical emulators allow me to compute the risk premiums, which are small, because the EU is rich and the policy impact is small. The ensemble of ten models allows me to compute the ambiguity premium, which is small for the same reason. I construct a counterfactual estimate of recent emissions without the climate policy and use that to test the predictive skill of the ten models. The models that show the lowest cost of emission reduction also have the lowest skill for Europe in recent times.

Highlights

Uncertainty abounds in climate policy and ambiguity, too
Further research and better observations should improve our understanding of the climate system and how it responds to climate policy
I assume that the estimated emission reduction was induced by a carbon price of $13/tCO2, the average European Union (EU) ETS price over the period

Summary

Introduction

Uncertainty abounds in climate policy and ambiguity, too. Uncertainty about climate is Energies 2014, 7 partly irreducible; climate change is a problem of the future, after all. Too; the climate system is too complicated to construct a model from first principles; so different modellers will make different choices, and the models will disagree. Further research and better observations should improve our understanding of the climate system and how it responds to climate policy. We should be able to learn which models to believe and which to discount or even discard. This paper contributes to that, assessing which models of the costs of climate policy are the more credible ones and applying that insight to the 2030 target of the European Union (EU)

Methods

Results

Conclusion