The Global Warming Policy foundation today published a pamphlet WHAT IS WRONG WITH
STERN? The Failings of the Stern Review of the Economics of Climate Change by Peter Lilley MP.
Lilley criticises ‘Stern’s selective emphasis on alarmist interpretations and downplaying of uncertainties’ (p4). Richard Tol in his foreword claims that ‘Lilley also reveals that the errors are systematic and suggestive of an ideological bias’.
This surprised me, because in the years following the publication of the Stern review, researchers have been concerned that far from being alarmist, its impact estimates may well underestimate US and global damages.
I created PAGE2002, the Integrated Assessment Model used by Stern for its aggregate impact estimates, so I thought I would see what Lilley had to say about PAGE2002.
Lilley claims that for economic losses in PAGE2002, ‘a 2.5°C temperature rise is deemed to reduce GDP by between 1.5% and 4% – with a median 2% loss’ (p35).
I have no idea where he found those figures.
In the default PAGE2002 model, the actual reductions in GDP from economic losses for a 2.5°C temperature rise are between -0.1% and 1% of GDP with a median loss of 0.5%. That’s right. The default PAGE2002 model actually allows a small chance that the impacts of global warming will be positive, and Lilley overstates the median impacts in the model by a factor of four.
But there’s more. While Lilley acknowledges that in PAGE2002 ‘ a proportion of losses – lower in less-developed countries – is assumed to be prevented by adaptation’ (p35), he doesn’t say what that proportion is. Would you like to know? In OECD countries, it’s over 90%. The default PAGE2002 model assumes that 100 percent of the economic damages resulting from the first 2 degrees of warming, and 90 percent of economic damages above 2 degrees, are eliminated by adaptation.
Taking this into account, Lilley has overestimated the median economic damages in the PAGE2002 model by over a factor of 40.
I can’t help but think this is an error by Lilley that is, in Richard Tol’s words, ‘systematic and suggestive of an ideological bias’. If this is typical of Lilley’s rigour, then I know whose work would lead me to declare ‘it’s academic value is zero’, and it isn’t Stern’s.
Matt Simmons
Look back 100 (possibly even as little as ten) years at any prediction of future economy. The one thing they will have in common is that they are all entirely wrong. There is simple mathematical reason for this: to compute the fifty-sixth impact of balls on a billiards table, every single elementary particle in the universe needs to be present in your assumptions.
See http://www.anecdote.com.au/archives/2007/10/the_billiard_ba.html
I am sure all the intelligent people from Stern to Lilley to Hope all realise deep down that this is a futile exercise that will do the world no more good than rearranging the deckchairs did for the Titanic.
Yet they waste their time engaging in it anyway. THAT is what is wrong with this argument.
Richard Tol
How can one error, if that is what it is, imply a systematic bias?
We do not know exactly what the Stern Team did with PAGE. (We do know that arithmetic errors were discovered after publication but never corrected.)
Chapter 6 of the Stern Review is a whirlwind of numbers, but they never specify the impact function. I don’t know where Lilley got his numbers, but I cannot say for certain that he is wrong.
John Russell (Twitter@JohnRussell40)
Quote: “How can one error, if that is what it is, imply a systematic bias?”
All errors are not equal. I would have thought bias depends on the size of the error. Overestimating “…the median economic damages in the PAGE2002 model by over a factor of 40” …could be considered systematic bias when compared with (allegedly) a few minor arithmetical errors.
Peter Lilley
Dear Chris
Thank you for querying the figures I attribute to the PAGE2002 Impact Assessment Model.
In fact the figures I quote do come from your model – Figure 5 on page of your explanatory article The Marginal Impact of CO2 from PAGE2002: An Integrated Assessment Model Incorporating the IPCC’s Five Reasons for Concern. They are line 7 and refer to India so my paragraph should have read:
“The model is given a range of assumptions of impacts on the GDP of each geographic area for a 2.5°C rise in temperature. Thus, IN INDIA a 2.5°C temperature rise is deemed to reduce GDP by between 1.5% and 4% – with a median 2% loss. The loss is then set to increase as a power of temperature ranging between linear and cube – averaging 1.3.”
The words IN INDIA somehow got erased and I will reinstate them in future versions especially as it then makes more sense.
You also single out India in your excellent presentation to the Yale Symposium – page 48 where you point out that though “adaptation reduces impacts by 90% in OECD countries” it reduces it by only “50% in India”. That was the point I was referring to in the second quote to which you refer. Please correct me if I have misinterpreted your model, but as I understand it, even when India and other poor countries which constitute the bulk of the world reach current OECD levels of development they will still be deemed to adapt only by 50% not by 90%?
I am sorry if you thought I was trying to misrepresent your model. Far from it. The clarity and transparency with which you present all your assumptions and equations stood out as a model which I only wish others on all sides of this debate would emulate. So I regret all the more that a proofing error – mea culpa – led to that impression.
Best regards
Peter Lilley
cwhope
Dear Peter (if I may)
Thank you for responding. Unfortunately even your amended version is not correct.
The figures to which you refer are actually the multiplicative factors for India, not the absolute impacts. So they are the amount by which you have to multiply the % of GDP impacts in the EU to get the impacts for India. The most likely value is 2, so as a rough guide that can be used to multiply the median value for the EU which is 0.5% of GDP, to get 1% of GDP in India as a median value.
The interpretation of adaptation in India in the default model is correct. It stays at 50% even when India becomes rich.
See journals.sfu.ca/int_assess/index.php/iaj/article/download/227/190 for the source article.
It’a a shame that we didn’t have a quick chat before publication, when I could have explained this. Perhaps my writing is not always as clear as you have been kind enough to say it is.
Chris Hope
@cwhope
Peter Lilley
Further thanks. I will incorporate that explication of the figures into the revision of the electronic version of my report.
Aaron
You may all ready be familiar with this, but one of the criticisms of the Stern report which really struck home for me was that by Kevin Anderson and Alice Bows of the Tyndall centre for climate change research.
Their work on emissions pathways shows us that Stern and others such as the CCC use scenarios that are far too optimistic, having much too low growth rates in emissions and unrealistically early peak dates, as well as poor assumptions about the behaviour of non-annex 1 countries.
Their findings are summarised in this excellent and informative presentation :
http://www.slideshare.net/DFID/professor-kevin-anderson-climate-change-going-beyond-dangerous
(Particular criticism of Stern can be found at 22mins in)
Aaron
I found an even better set of slides by Kevin Anderson which specifically picks apart current integrated assesment models, highlighting the poor assumptions on which many (including the stern report) are based and makes it very clear why current policy means that we are very likely to miss the two degree C target for gloabl average tempreatures.
http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=6&cad=rja&ved=0CD4QFjAF&url=http%3A%2F%2Fwww.zerocarbonbritain.com%2Fresources%2Fconference-resources%3Fdownload%3D30%253Apost-carbon-worlds-and-transitions-kevin-anderson&ei=uS1LUK6PJKLB0gWxn4DwBg&usg=AFQjCNEhpNxkVLDgOHlNosTMyQBZ496RpA …