Country-level social cost of carbon

Go to the profile of Katharine Ricke
Sep 26, 2018

The paper in Nature Climate Change can be found here:

We started work on this paper in early 2015, in what seemed like a very different geopolitical atmosphere than today.  The impetus for the research was purely novelty-based. The idea was to combine an approach to analyzing the climate effect of a marginal emission of carbon dioxide that Ken Caldeira and I had recently developed, with a climate damages model described in what was then a working paper by Marshall Burke and collaborators. My co-author Massimo Tavoni pointed out that by combining these two tools, we could produce the first comprehensive, country-level estimates of the social cost of carbon. Back then, the proposed analysis seemed to me mostly like an academic exercise because – for reasons explained below—country-level social cost of carbon is not conventionally considered a relevant metric for setting climate policy. 

Social cost of carbon is a measure of the economic damages associated with an additional present-day ton of carbon dioxide emissions. In theory, it’s more than just a policy-relevant metric—it’s practically policy-actionable. When presented with an opportunity to reduce CO2 emissions through some public policy intervention, SCC tells you how much you should be willing to pay to avoid emitting a ton of CO2. Country-level social cost of carbon is, by nature, a contentious concept. Because carbon dioxide is a global externality, a rational, coordinated human society (or what economists might call “the social planner”) would never want to set policy based on anything other than the global aggregate value.

After the Brexit vote and the 2016 presidential election here in the US, like many others, I was wondering about the implications of resurgent nationalism for international climate policy.  While the analysis and a draft of our manuscript were already complete, in late 2016 we revisited the story we were telling about the country-level SCC database we’d produced in order to focus on a more realpolitik framing for the paper.

In the United States, there sometimes seems to be a misconception that investments in climate change mitigation are charitable acts. Because climate change presents a more existential threat to developing countries, the argument is that having historically contributed the most CO2 to the atmosphere, the US should reduce its emissions in order to reduce the impacts of these past actions on others. On the other hand, our analysis demonstrates that the idea that the biggest beneficiaries of reductions in carbon dioxide emissions by the US would be other countries is a myth.  Our results suggest that based on pure self-interest, the US should be willing to pay around $40/ton to avoid an emission of CO2. What’s more, the US is very consistently in the top 3 of country-level social costs of carbon – by this metric one of the biggest losers from climate change.

In order to make an impact, to distill complex analyses for consumption by a broad audience, and to get through peer review, researchers have to develop a salient narrative about their work. When you do policy-relevant scientific research, even when following the scientific method perfectly, norms and values inevitably enter into the way you tell its story.  

Our paper took nearly two years to get through peer review; things got heated. The Burke-Hsiang-Miguel damages model applied in our reference case is highly controversial among some climate change economists, including a couple of the six referees we heard from. In addition, the peer review process required us to engage in a debate over whether it was even “right” to report anything other than a global number for SCC. We were subjected to quite a bit of substantive criticism that required us (led by Laurent Drouet) to re-up the sensitivity analysis again and again. I learned a lot through this process: about climate econometric models with fixed effects, about the entrenchment of the principle of utility maximization and plenty more. It was a lot of fun and I’m glad we can finally share the paper with everyone.

Go to the profile of Katharine Ricke

Katharine Ricke

Assistant Professor, University of California, San Diego


Go to the profile of Chris Hope
Chris Hope 23 days ago

Welcome to the Integrated Assessment Modelling community. This seems to be a serious attempt to grapple with the issue and produces mean SCCO2 values much more in line with recent PAGE09 analyses than other efforts.

I would challenge your assertion that 'Past estimates of the SCC have included limited uncertainty analysis focused mostly on a limited set of parameters such as the social discount rate'. This is clearly not universally true. 

See, for instance, 'Critical issues for the calculation of the social cost of CO2: why the estimates from PAGE09 are higher than those from PAGE2002', Climatic Change, April 2013, Volume 117, Issue 3, pp 531–543 particularly figure 3 and the subsequent text, which shows the eight major influences on the SCCO2 in the PAGE09 default model. Overall, the PAGE09 model uses probability distributions for over 100 of the most important inputs, and calculates SCCO2 values from 100,000 runs of the model.

Chris Hope