15 March 2010

Bias in IPCC AR4 WG III? A Guest Post by Richard Tol and Chris Green

This post continues Richard Tol 's series looking at Working Group III of the 2007 IPCC report. Here he is joined by Chris Green, professor of economics at McGill University.

In this Part VI Tol and Green conclude that,
Table SPM.1 is wrong and misleading. It is wrong to suggest that emission reduction costs are independent of the assumed future developments. It is misleading to conclude that emissions can be substantially reduced at negative economic cost.
Please have a look at their full discussion below. If you have questions or criticisms of their analysis please submit them in the comments, I am sure that Richard and Chris will be happy to engage.
Double Counting and Negative Costs

The first 5 parts of this series one of us (Tol) looked at Chapter 11 (Parts 1, 2, 3) and Chapter 3 (Part 4, 5) of the Fourth Assessment Report (AR4) of Working Group III (WGIII) of the Intergovernmental Panel on Climate Change (IPCC). Parts 4 and 5 looked at Chapter 3.

Here, we return to Chapter 11. The first and second order draft of the chapter and the review comments can be found here.

Chapters 4-10 assess the literature on emission reduction by sector (energy, transport, buildings, industry, agriculture, forestry, waste). Chapter 11 synthesizes that material, and is thus a key step towards the Summary for Policy Makers.

According to Table SPM.1, emissions can be reduced in 2030 by 5-7 GtCO2eq/yr at zero marginal costs (and thus a total benefit). This represents 7-10% of emissions in the A1B scenario (68 GtCO2eq/yr), and 10-14% of emissions in the B2 scenario (49 GtCO2eq/yr). That is, the ability to reduce emissions and make money at the same is independent of the no-policy scenario.

You can look at emission reduction in two different ways. First, you can take the present as your starting point. You would find that there are many opportunities to save energy. This is because state-of-the-art technologies are superior to the technologies that are actually in use (which were state-of-the-art when they were bought years or decades ago). Therefore, you would conclude that there are great opportunities to reduce emissions at low or negative costs.

You can also look at this from a different perspective. Energy efficiency has improved greatly in the past – without the support of climate policy – and this is likely to continue into the future. New and improved gadgets that will make people better off will be sold anyway. This is part of any reasonable projection of future emissions without climate policy. However, you would conclude that climate policy would push energy efficiency harder – and that this would cost money.

The two perspectives are equally valid, but they should be kept separate. The fuel efficiency of car engines does not improve by itself – engineers will need to make an effort. But they will make that effort because the buyer demands a bigger car with the same mileage. This does not constitute a low- or negative-cost option, let alone a success for climate policy – because it would also happen without climate policy.

Table SPM.1 mixes the two perspectives. The SRES scenarios A1B and B2 rightly assume substantial technological progress in the absence of climate policy. The results can therefore not be compared to the negative cost emission reduction potential. Table SPM.1 makes that comparison nonetheless.

Table SPM.1 even suggests that the negative cost potential is independent of such things as energy prices. The 2030 price of energy is higher in A1B than in B2. Therefore, more energy saving equipment would break even in A1B than in B2 – and the negative cost potential should be larger in A1B than in B2. It is not in Table SPM.1.

WG3 has no excuse for making a mistake like this. This issue has been known for at least 20 years, and was recently repeated in the “dangerous assumptions” paper by Pielke, Wigley and Green in Nature which rallied against the confusion between emission reduction by climate policy and by other forces.

How could this happen?

The First Order Draft of the Summary for Policy Makers does not have Table SPM.1 or anything like it. The Second Order Draft has a table which does show negative emission reduction costs – but only against a single scenario. Table SPM.1 is based on Table 11.3 in Chapter 11. Here is a comment on the Second Order Draft:
Table 11.3 is sensitive to energy price development
Response:
Table 11.3 provides alternative estimates of potentials for different carbon prices, but for specific baselines. Clearly the baselines are affected by the assumed energy prices, but it was not possible in the time available to undertake the comparison exercise using different baselines.
Here’s a comment on Table SPM.1:
We have serious reservations about the validity and comparability of the underlying estimates from Chapters 4 to 10 […]because there is no demonstration that the estimated mitigation potentials have been evaluated properly with respect to the two baselines used for this analysis. The mitigation estimates in the SOD apparently do not take account of the technological changes already embedded in the SRES B2 and WEO 2004 reference scenarios against which the mitigation estimates were made ...
Response:
a fully consistent modelling framework is not available; that is why a scenario analysis was applied
So, the authors of both the SPM and Chapter 11 were well aware that the results are problematic.

There was also this comment:
The estimates of achievable mitigation (by 2030) presented in tables SPM-[1], TS-19, and 11-3, may not in fact contribute to mitigation from baseline. The estimated achievable emission reductions may be absorbed by the energy-intensity reductions and decarbonization embedded in the SRES B2/WEO (2004) baselines. If in fact estimated mitigation possibilities can truly contribute beyond that which is already embedded in the baseline scenarios, then that should be demonstrated in detail, not simply assumed.
Response:
Rej[ect]: mitigation potential is additional to what is included in baseline
Another comment (on Table 11.3):
None of the discussion of mitigation potential […] belongs in an IPCC Assessment Report. […] Calculations of mitigation potential are only possible through application of a model that […] develops an explicit baseline […] and makes consistent assumptions about costs and performance of present and future technologies. There is an implicit model behind the original calculations presented in the chapter, but it is so simplistic and leaves out so many critical factors in estimating a marginal abatement cost curve that the results would not be accepted for publication in any reputable journal. I would be embarrassed as a chapter author to take responsibility for such flawed and na├»ve analysis in a report designed to be an assessment of the best research available. […] It likewise appears to assume that the same amount of mitigation be achieved at a given cost no matter what baseline emissions pathway is chosen – which clearly cannot be the case because the technologies assumed as mitigation measures may already be adopted in the baseline. Moreover, there is no demonstration that the estimated mitigation potentials have been evaluated properly with respect to either of the two baselines used for this analysis. The mitigation estimates in the draft apparently do not take account of the technological changes already embedded in the SRES B2 and WEO 2004 reference scenarios against which the mitigation estimates were made.
Response:
REJ[ect]: The dependence of the estimates on the baseline is recognized and allowed for.
So, when the comment was gently worded, the IPCC authors admitted that the results are problematic – but no action was taken. When the same comment was raised in a less gentle voice, the IPCC authors simply denied the problem – and no action was taken.

Table SPM.1 is wrong and misleading. It is wrong to suggest that emission reduction costs are independent of the assumed future developments. It is misleading to conclude that emissions can be substantially reduced at negative economic cost.