Title: Modelling Climate Impact on Economic Growth: More Harm, More Uncertainty or None of the Above? Numerical Experiments and Implications for Mitigation Policy.
Céline Guivarch is an economist at École des Ponts ParisTech, CIRED (International Research Center on Environment and Development). She develops and uses models, from analytical “toys” to complex Integrated Assessment Models, to explore the interactions between economic dynamics, energy systems evolutions and the environment. She is particularly interested in the role of uncertainty for climate policies decisions and for socio-economic scenarios for climate change research. She teaches Environmental and Resources Economics in an Advanced Master in Public Policies for Sustainable Development.
Before her current position, she worked a year in the Climate Change Unit at the International Energy Agency. And before her Ph.D., she was during two years the regional coordinator of the European project “Technical Assistance to Central Asian countries with respect to their climate change commitments”.
She holds an engineering degree (Physics and Earth System Mechanics) from École Polytechnique, a Master in Environmental Sciences and an Advanced Master in Public Policies for Sustainable Development from École des Ponts ParisTech and a PhD in Economics from Université Paris-Est.
Recent articles have investigated with Integrated Assessment Models (IAMs) the possibility that climate damage bears on productivity growth and not on production, the traditional route that follows Nordhaus' work. According to these articles, damage on productivity leads to a higher social cost of carbon (SCC) and thus promotes more stringent climate policies in the short run. This result is however difficult to reconcile with other pieces of evidence suggesting that the willingness to pay to avoid temperature increase is lower when temperatures affect growth rates rather than level of GDP. Here, we reconsider the evidence and show that two problems have been mixed in this recent strand of literature: the first is the total amount of damage; the second is the differential impact of damage depending on their bearing on production or productivity.
We build a simple climate-economy model and show that when total damage is the same, the ranking of SCC between a model with damages on production and a model with damages on productivity depends on parameters such as the utility discount rate or the elasticity of marginal social utility of consumption. Similarly, the uncertainty on the SCC depends on parameters considered. The difference in SCC comes both from where damages are located and from their total amount. Failing to disentangle the two effects, most of the recent findings are paramount to the tautology that higher damages lead to higher social costs of carbon. A way out of this state of confusion would be to agree on a common measure of damage across IAMs.
In the second part, we explore alternative measures to compare damages between models, and we investigate which differences are robust between a case where temperature affect growth rates and a case where temperature affect levels of production. We find that the following results are robust: (i) the welfare gain to mitigation is higher in cases where temperature affect growth, (ii) the increase of the SCC with delay to abatement is higher in cases where temperature affects growth. We also find that, if models are compared with a measure of damages that integrates damages over the short-term, the SCC is both higher and more uncertain when damages bear on productivity than when they bear on production. We discuss the implications for modelling choices and data and methodologies for models calibration. We conclude with implications for mitigation policies.