Climate change “fat tails”: my paper with Gerard Roe

Update March 19, 2013: The paper is now in print! Climatic Change, April 2013. Published version here, ungated version here or below.

I’ve had the very good fortune of working with UW climate scientist Gerard Roe on a new paper, Climate sensitivity: should the fat tail wag the policy dog?

The paper is currently under peer review. Here’s the abstract:

The small but stubbornly unyielding possibility of a very large long-term response of global temperature to increases in atmospheric carbon dioxide can be termed the fat tail of high climate sensitivity. Recent economic analyses suggest that the fat tail should dominate a rational policy strategy if the damages associated with such high temperatures are large enough. The conclusions of such analyses, however, depend on how economic growth, temperature changes, and climate damages unfold and interact over time. In this paper we focus on the role of two robust physical properties of the climate system: the enormous thermal inertia of the ocean, and the long timescales associated with high climate sensitivity. Economic models that include a climate component, and particularly those that focus on the tails of the probability distributions, should properly represent the physics of this slow response to high climate sensitivity, including the correlated uncertainty between present forcing and climate sensitivity, and the global energetics of the present climate state. If climate sensitivity in fact proves to be high, these considerations prevent the high temperatures in the fat tail from being reached for many centuries. A failure to include these factors risks distorting the resulting economic analyses. For example, we conclude that fat-tail considerations will not strongly influence economic analyses when these analyses follow the common—albeit controversial—practices of assigning large damages only to outcomes with very high temperature changes and of assuming a significant baseline level of economic growth.

10 responses to “Climate change “fat tails”: my paper with Gerard Roe”

  1. Please notify me if when this paper is published. Also, I came across your (Bauman and Ladyka) review (in Climate Progress) of economics textbooks with respect to climate issues and am wondering if you think textbook buyers are climate savvy enough to respond accordingly if your review is widely distributed to economics professors? Are there climate based (biased) reviewers in other subjects too: IR, PolySci, all the communication arts, etc? If not, this would seem like a must do for a more broadly based climate savvy academia.

  2. I’m struggling here. You say: “If climate sensitivity in fact proves to be high, these considerations prevent the high temperatures in the fat tail from being reached for many centuries. ”
    Climate sensitivity is the eventual temperature response from a doubling of CO2. You claim that the higher is this parameter, the later temperature impacts are likely to be realised. This contradicts the message I have taken from what climate scientists say. So where are the climatological references supporting that please?

    YB: As I understand it (and keep in mind that I’m the economist on this paper, not the climate scientist) a temperature increase of (say) 1C will be reached sooner if climate sensitivity is 4C than if it’s 2C. But the amount of time it takes to reach 4C if climate sensitivity is 4C is greater than the amount of time it takes to reach 2C if climate sensitivity is 2C. Does that make sense? See the references in the paper for more, and I think you can also see it from the asymmetry in the graphs in our paper.

  3. Thanks for your reply, I hope I get time to check out the paper. However, even 2C is commonly regarded outside of economics as pretty bad, because it implies a multiple of that increase at the poles and makes currently marginal agricultural areas sub-marginal. To assume continued global economic growth in such (and worse) circumstances seems to me to be a delusional fantasy, if, as you say, a common one in economics.
    But my main concern is this: why do you only consider temperature rise, and say nothing about ocean acidification or sea-level rise? These are regarded as certainties under business as usual, and accelerating sea level rise is already a fact according to e.g. NASA. I haven’t seen anything credible contradicting this. To put this in perspective, development tends to be concentrated towards coastal regions, a trend which is only increasing. Relatedly, ice melt already exceeds the worst projections of the IPCC reports. Your draft paper does not even mention these phenomena!

    YB: The question of why there are such differing views about the impact of 2C (with many scientists terrified, and many economists blasé) is important and fascinating, but I don’t currently have an answer, sorry. (It is, however, at the top of what passes as my research agenda.) As for why we focus on temperature rise: this is in keeping with the “fat tail” literature. In some ways temperature rise is simply a proxy for “all the bad stuff that could happen”. Admittedly this is quite a simplification, but again the focus of our paper is the “fat tail” idea that we could see 20C temperature increase this century. As we note in the paper, “The most pressing questions are not about the likelihood of 10 or 20 C temperature increases, but about the impacts of 3-6 C temperature increases and about the ability of economic growth to cushion climate impacts in the decades and centuries ahead.” Ocean acidification and sea level rise are both topics for concern from this perspective.

  4. YB

    Have you used market rates for your discount rate? FRED shows 20-year TIPS with a close to zero real return at present. The latest Credit Suisse annual Global Investment Returns Source book shows this long maturity collapse in real rates is not isolated to the U.S. Are you adjusting your discounting assumptions to take into account real market data? If not, why not?



  5. Well, 400 years is a very long time. We can quite easily get into a fat tail of 4 degrees of warming by end century – I am sure you know the science as well as me. So why shouldn’t we be using market rates to discount damages in such a tail. There seems to be a very strange contradiction here. When Lord Stern suggested a very low discount rate it was decried as not reflecting people’s revealed preferences with respect to consumption through time and the market’s expectation over growth.

    Now the market is suggesting that Stern’s discount rate is about right – or perhaps too high. Well you can’t have your cake and eat it! What is it to be: are markets giving correct signals or not? If we suddenly decide we don’t like what the market is saying and therefore shall ignore it, where is the logical consistency within any general equilibrium theory.

  6. Yoram

    Thanks, have seen your qualifications. And yes I did actually find you via ‘The Economist” quote, which annoyed me intensely. I agree with most of what you say above. My current gripes are 1) the damage function within DICE and so on appears ridiculous given how the science is progressing and 2) Nordhaus’ caves and mansions analogy only works if you assume robust growth.

    I am not saying that market interest rates are the only determinants of the discount rate, but I do find it perplexing that they were referenced a lot during the attacks on Stern (in effect saying he was being too normative) but now are being quietly ignored (since they have collapsed to Stern-like levels). It also appears that ‘growth’ economics (or should I say ‘lack of growth economics”) is undergoing a bit of revival as witnessed by all the fuss over Robert Gordon’s paper. So some of this should be taken into account each time we think of popping 2-3% GDP growth into the models. FYI actuaries are now starting to put the new normal into their models so why not climate economists?

    All in all, the neoclassical treatment of climate economics (most famously with DICE) was constructed during the Great Moderation. The empirical side now looks highly dubious, and don’t get me started on the theoretical side. (And did I mention that the damage function work stinks.)

    I aim to do a long post on the discount rate post the credit crisis, but it is complicated! Still trying to get my head around it :)

  7. BTW, TIPS aren’t the only market rates that have collapsed; this phenomenon is everywhere.

Join My Email List