Select Committee on Economic Affairs Minutes of Evidence

Memorandum by Professor Dennis Anderson, Centre for Energy Policy and Technology (ICEPT), Imperial College London

  I wish to thank the Committee for inviting me to submit evidence to its inquiry. My knowledge of the subject is not encyclopaedic, and my recent research has focused more on the engineering and economic aspects of energy technologies and policies than on climate change itself. I have observations on three topics:

    —  Estimates of economic damage of climate change—or the benefits of its mitigation.

    —  The economic costs of mitigation.

    —  Economic aspects of the work of the IPCC.


  Scientific opinion as to the likely magnitude of the environmental consequences of climate change ranges from deep scepticism to deep concern. Economic analysis of the benefits of abating the damages, it seems to me, is heavily tilted towards the sceptics. For example, a recent review[7] of the economists' estimates concludes that the marginal social costs of carbon, without equity weightings, is in the range £3-6/ton of carbon, £3-15/ton with equity weightings or £4-27/ton if a time varying discount rate is used.

  This is close to saying that the consequences of climate change are trivial from an economic perspective, worthy of a carbon tax (taking the average of this range to be £15/ton C) that would add less than 1.0p per litre to the price of petrol and 0.3p per kWh—about 4 per cent—to the average retail price of electricity. The economic damages arising from unleaded petrol and acid deposition from coal-fired power stations were each imputed to be appreciably greater than this, and to warrant greater expenditures on their abatement than would the prospect of climate change. Economists making such estimates are clearly on the side of the deep sceptics of climate change.

  This is not the place to dwell on the importance of scepticism in the search for scientific truth. By definition, however, sceptics (and the deep sceptics above all) must also ask themselves, what if we are wrong? It is in this spirit that I raise just three of several shortcomings of current economic analyses of the social costs of climate change:

    1.  The possibility of threshold effects is generally ignored. Yet threshold effects are a common concern in the scientific literature, for example in journals such as Nature and Science. Examples would be methane releases with a warming of the permafrost, the long-term implications for water supplies (in Asia in particular) arising from the retreat of glaciers, the switching off of the Gulf Stream, and the thawing (partial or complete) of the Greenland ice cap; but there are several others.

      The damage functions that appear in the models of the economists are simple functions of the global average temperature of one quantity or another, such as the upper layer of the ocean, which in turn is related to a simple function for estimating carbon accumulations. There is no positive feedback mechanism, for example, linking carbon releases from the permafrost to surface temperatures, no temperature related function affecting the behaviour of the Gulf Stream, nor do I see anything in the equations to relate damage (eg glacier retreat, or damage to the rainforests) to the history of temperatures: damage is related to temperatures at a particular level of carbon accumulations at a particular point in time. In short, the climate damage function commonly assumed is a gross oversimplification.

      Attempts to allow for the possibility of threshold effects indicate that the costs of damages would be several times higher—possibly an order of magnitude higher—than indicated above. [8]These estimates are, moreover, first cuts. Professor Nordhaus, who has made an attempt, candidly states "because the sources of abrupt climate change are poorly understood it is difficult at present to link economic policies to abrupt climate change . . . knowledge [of which] is extremely thin." [9]

    2.  The damage functions are falsely assumed to have the characteristics of a reversible process. If we were to gradually shift to a process by which carbon accumulations were stabilised, or even to begin on a process of carbon decumulation (a not infeasible possibility), the economists' damage functions imply that all the damage would eventually be undone. I raise this point not to argue for the desirability or otherwise of beginning a process of de-cumulating CO2 in the atmosphere, but to suggest that the damage function is scientifically mis-specified. The effects of climate change are not reversible processes.

    3.  The treatment of uncertainty. A recent review of published studies finds the mode of the estimates of the damage costs to be $5/ton of carbon, the mean $104/ton, and the 95 percentile $446, but then concludes that the marginal costs of damage are unlikely to exceed $50, roughly the upper limit of the range found in UK (of £4-27/ton C) studies. [10]Such estimates evidently dismiss the work of those who have come up with higher estimates than above.

      The treatment of uncertainty in economic studies is with few exceptions[11] shoddy. Outliers (even within the 95 per cent confidence interval) are ignored, scientific uncertainties on the nature and extent of climate change itself and of its possible impacts are dismissed where they lead to inconveniently high estimates of damage costs, there is an excessive focus on mean "certainty-equivalent" values, and irreversibilities are neglected—all this despite a now significant economic literature on decision making under uncertainty which finds that, when uncertainties are large, and when the effects of not investing would be irreversible, the best policy is to develop options, as discussed below.

      In sum, the benefit estimates so far are too unreliable to use, and rest too much on deterministic methods. We should turn instead to the economic principles that have been developed to cope with uncertainty and irreversibilities, namely those concerned with the development of options. We need to address the question:

    Given the uncertainties and the possibility of threshold effects, what sorts of policies are appropriate for developing the options for mitigating climate change?

      A carbon tax is among the array of policy instruments suggested by this question, since it is ideal for encouraging the development of carbon-neutral options. But its level needs to be decided not on the basis of scientifically unreliable analyses of the expected marginal benefits of mitigation, but in response to the question—will it lead to the development of options for responding to climate change if it turns out that the deep sceptics are wrong? [12]


Technological Options

  There is a broad range of carbon-neutral options available or emerging for addressing climate change. They have been reviewed in a number of studies by industry, academics and government committees: [13]

    (i)    Carbon capture and storage from fossil power plants. There are several possible routes, one being the pre-combustion production of hydrogen and CO2 (from natural gas or gasified coal), geological storage of the CO2 and the use of hydrogen for electricity generation. The CO2 could be used for enhanced oil or coal bed methane recovery on a low carbon cycle, an option that has been mooted for depleted oil fields in the North Sea.

    (ii)    Carbon capture and storage from hydrogen plants. A similar process to (i), with the hydrogen produced being used directly in buildings, industry and for transport fuels.

    (iii)    Carbon capture and storage from synfuels plants. The production of liquid fuels (synfuels) from coal on a large scale is a well known practice.

    (iv)    Nuclear energy.

    (v)    Substitution of natural gas for coal in power generation. There is still much scope for this, especially in Asia.

    (vi)    Renewable energy from wind, waves, tidal streams and photovoltaics—which is capable of producing energy on a very large scale. PVs are especially promising in the developing regions, where the energy yield is nearly three times higher than in the UK, and is better distributed throughout the year. The offshore resource is promising in the UK.

    (vii)    Use of biofuels for transport, and biomass wastes for combined heat and power.

    (viii)    Hydrogen from renewable energy to replace fossil fuels in transport—for use in high efficiency fuel cell cars.

    (ix)    The hybrid vehicle.

    (x)    Fuel cells and micro-generators for decentralised forms of combined heat and power. These are promising high-efficiency options.

    (xi)    A range of options (too numerous to list here) for improving energy efficiency in buildings, industry and transport.

    (x)    Storage technologies—to address the much-discussed "intermittency problem" associated with some forms of renewable energy

Two Dimensions to Policy, not One

  The case for a carbon tax (or the alternatives of tradable carbon permits) remains, and the point has been made many times by HoL and HoC Committees and many other parties that the UK's Climate Change Levy—which is really an energy tax unrelated to carbon emissions—should be transformed into a carbon tax, and applied to all hydrocarbon fuels based on their CO2 emission intensities.

  In addition, it will be necessary to develop policies toward innovation in carbon-neutral technologies, through RD&D programmes and incentives for the commercialisation of new technologies. This point has been accepted by all governments in OECD, including the US, which arguably is at the forefront, and several developing countries such as China, India and Brazil. Hence there is a good base to build on both to develop national policies and for international co-operation, a possibility that has been recognised by the Prime Minister himself, most recently in his current capacity as president of the G-8.


  The estimated costs of reducing carbon emissions by 40-60 per cent over the next half century range from 0.0 to 4.5 per cent of GDP, with a few outliers above and below this range; the average is 2½ per cent. The following table summarises the estimates of 40 studies undertaken over the past decade. 2½ per cent of GDP is of course appreciable in absolute terms, but would be small in relation to the growth of economic output, which is likely to be several 100 per cent higher in 50 years' time than it is today. A long-term transition to a low carbon economy would unlikely be disruptive to long-term economic growth.

Table 1

% Global GDP losses
Total Mean
value of
Barker et al (2002)3 783 7282.6
Grubb et al (1993)3 161 0112.0
Totals6 8144 7392.4
US studies reviewed by Grubb et al (1993) 327 12142.5

  The main reason why the costs may be higher than the mean estimate of 2½ per cent of GDP is the problem of cost escalation: the costs of the carbon neutral technologies may turn out to be higher than projected, a not unusual occurrence, as may be the costs of improving energy efficiency. [15]There are, on the other hand, several reasons why the costs could be lower: the costs of the "marker" technologies such as natural gas and oil could rise with the growth of world demand (most studies assume the costs of oil to be ~$30/barrel, which already looks distinctly low); the rates of technological innovation may be higher than assumed; and there may be side benefits, such as the local and regional environmental benefits of the carbon neutral technologies and practices.

  There is an exception to the conclusion that the effects on the level and growth of GDP are likely to be small. It is that, if the transition to a low carbon economy were to be pushed to extremes over a very short period then costs would indeed be high. A study by the US EIA (cited by the Administration as a reason for withdrawing from Kyoto) estimated that a 7 per cent cut in CO2 emissions (relative to 1990) by 2010 implemented over the four years 2005 to 2008, would lead to a loss of 4.2 per cent of GDP by 2010, or around $400 billion. This result is not surprising since emissions in the US have already grown by 15 per cent since 1990, and meeting the target would now require the US to cut emissions by 25 per cent relative to what is projected for 2010; thus a rate of reduction of 6-7 per cent per year would be needed to meet the targets—five to 10 times faster than the UK's already ambitious targets for the period 1997-2010. It would entail the premature closing of coal mines and coal-fired stations and many other extra-ordinary changes in the energy system; and it would allow no time for innovation to reduce costs. Policies require a flexible and a phased response if costs are not to be extra-ordinary.


  I have not been involved with the work of the IPCC, other than making a marginal contribution to the discussions on emissions scenarios (SRES report) seven years ago. Overall, I find the reports balanced, well-referenced and informative, and that the work of the IPCC is conducted in an admirably open spirit. Lastly, the products of Working Groups I and II on the modelling of climate change and its impact seem to be well reflective of the possibilities ahead, while Working Group III has provided estimates of future energy demands and emissions that seem to me realistic and defensible, if perhaps on the low side.

  For example, the world's energy demand today is approximately 400 Exajoules (EJ); the SRES scenarios estimate that this could roughly double (the "B" group of scenarios) or triple (the "A" group) in the next half century, depending on assumptions about rates of economic growth, populations, income elasticities and various other parameters. These are reasonable estimates. Developing countries, for instance, currently account for 150 EJ or less than 40 per cent of world energy demand, and two-fifths of their populations are still unserved with modern energy forms. Yet within 50 years their populations are likely to be 7-10 times higher than in the rich countries; economic growth rates in populous countries such as India and China are exceedingly high; and their per capita income elasticities are in the range 1.0-1.5 (two to three times those of the rich countries, whose income elasticities are declining as markets mature). Putting such facts together does indeed indicate that world energy demands could rise 2-3 fold in any scenario of economic success in the next 50 years—and possibly by much more.

  I have two comments. First, there is a tendency, not only within the IPCC, to think of the "B" scenarios as being the "good" scenarios because they have lower energy demands and, by assumption, lower emission. I do think this is wrong. There are few pieces of economic news more encouraging for instance than those of China achieving a long-term economic growth rate of 10 per cent per year, and of India—having seemingly been stuck at barely 3 per cent per year for decades up to 1970—lifting its growth rate to 8 per cent per year. (Both countries, incidentally are graduating over 250,000 scientists and engineers each year, each 15 times the number graduated in the UK, and are likely to become sources of innovation in new energy technologies.) Secondly, emissions depend less on the level of energy demand than on the ways energy is produced and used. I argued, without success, that it is the higher growth scenarios, with, indeed, perhaps higher energy demands, that are the more interesting; being associated with higher rates of innovation in energy technologies and practices, they are also likely to be the more "environmentally friendly" ones.

  Second, the so-called "storylines" behind the scenarios are complicated and confusing, so much so that they have obscured their central message. This is that the difference between the scenarios of high and low emissions rests on technological development—and on all the policies that bear on it. It has been shown, I think convincingly, [16]that a low carbon future is technologically and economically feasible even—or perhaps especially—in the high growth scenarios. I hope the Fourth Assessment will revisit the scenarios of economic growth, especially in developing regions, and will be clearer as to the possibilities for the mitigation of climate change.


  1.  Economists' estimates of the marginal benefits of mitigating climate change are unreliable: they err on the side of the "deep sceptics", ignore possibilities for threshold effects, and ignore irreversibilities. A better basis for policy would be to address the question, which policies would be most effective for developing the technological and resource management options for mitigating climate change?

  2.  There is a rich and diverse array of technological options and practices for both the mitigation of—and, not discussed above, adaptation to—climate change. A carbon tax combined with an innovation policy would seem to be the best approach to developing the options. All OECD countries, most of all the US, and a significant number of developing countries, have accepted the case for an innovation policy, and there is an excellent opportunity for an international initiative in this area. In the UK it is also time that the so-called Climate Change Levy was transformed into a carbon tax, with its level being based on an assessment of its practical effects on the development of options.

  3. As to economic aspects of the work of the IPCC, I would suggest:

    (a)  The often tacit assumption that moderate economic growth rather than high economic growth is in some sense of the term more "sustainable" needs to be questioned, particularly in developing regions, where improved rates of growth would be far more favourable to development and the environment in the broadest sense.

    (b)  The confusing "storylines" behind the scenarios need to be simplified such that the central message of the scenarios becomes clear: this is that the difference between the high and low emissions scenarios turns fundamentally on technology development and use—and thus on all the policies that bear on these factors.

25 February 2005

7   David Pearce (2003) "The Social Cost of Carbon and its Policy Implications." Oxford Review of Economic Policy. Vol 19, No 3, 362-80. Back

8   Eg those of Nordhaus and Boyer, as reported in the review by RSJ Tol, S Frankhauser, RG Richels and JB Smith (2000) "How much damage will climate change do?". World Economics, Vol 1, No 4, October-December: 179-206. Back

9   WD Nordhaus (1999) The Economic Aspects of Abrupt Climate Change". Paper prepared for a meeting on Abrupt Climate Change: The Role of Oceans, Atmosphere, and the Polar Regions, US National Research Council, January 22, 1999. Back

10   RS Tol "The marginal costs of carbon dioxide emissions: an assessment of the uncertainties" Centre for Marine and Climate Research, Hamburg University. Working Paper FNU-19, April 10, 1993. Back

11   One exception is the treatment of uncertainties in discounting the future. David Pearce, Ben Groom, Cameroon Hepburn and Phoebe Koundouri (2003). "Valuing the Future: Recent advances in social discounting." World Economics. Vol 4, No 2. April-June. Back

12   More technically, it is the option value not the expected marginal benefits of a policy that is the appropriate basis for a policy when we don't know what the expected benefits are. Back

13   See eg Pacala S and Socolow R (2004). Stabilisation Wedges: Solving the Climate Problem for the Next 50 Years with Current Technologies. Science, Vol 305, Issue 5686, 968-972. President's Committee of Advisers on Science and Technology (PCAST,1997). Recommendations of the President's Committee of Advisors on Science and Technology, Panel on Energy R&D for the 21st Century, November. Anderson, D and Gross, R (2002). Assessment of Technological Options to Address Climate Change: A report for the Prime Minister's Strategy Unit. Imperial College Centre for Energy Policy and Technology. December. The well-known scenarios of the Royal Dutch Shell Group have reviewed the possibilities on several occasions. Back

14   Grubb M, Edmunds, J, Brink P and Morrison M (1993). The costs of limiting fossil-fuel emissions: a survey and analysis. Annual review of Energy and the Environment 20, 71-81; Barker T, Koehler J and Villena M (2002). The costs of greenhouse gas abatement: a meta-analysis of post- SRES mitigation scenarios, Environmental Economics and Policy Studies, Vol 5, 2002, pp 135-166. Nb: In most studies, the 50-70 per cent reduction occurs in 50 years' time, and is consistent with larger reductions and the stabilisation of accumulations beyond then. Back

15   The 2Ö per cent estimate, however, corresponds roughly to a doubling of final energy costs to the consumers, and is thus conservative on the scope for cost reductions through innovation. Back

16   See footnote 7. Back

previous page contents next page

House of Lords home page Parliament home page House of Commons home page search page enquiries index

© Parliamentary copyright 2005