Problems with the IPCC economic models.

The IPCC overestimates the amount of future economic activity and hence the emissions of greenhouse gases and hence the total future warming. It is asserted by the IPCC that by the year 2100, CO2 may quadruple in amount to about 1200 ppm to give an upper limit of the warming of 5.8 C, a number that is frequently cited in discussions of the topic. Only by burning all the fossil fuels without any economic, mining, or drilling constraints could a concentration of 1200 ppm be achieved in 2100. This upper limit scenario is ridiculous. Even the lower limit IPCC scenarios are ridiculous (Castles and Henderson, 2003), leading to such conclusions as the average citizen in North Korea will be richer than the average citizen in United States in 2100. Castles and Henderson’s work shows that even if the IPCC climate models are correct, if one uses realistic economic scenarios, the warming will be quite small, of the order of 1 C. Some discussions on this topic are reproduced below.

From The Economist print edition (Feb 13, 2003)

The Intergovernmental Panel on Climate Change had better check its calculations

AT THE beginning of 2001 the Intergovernmental Panel on Climate Change (IPCC) released, as the main result of its massive Third Assessment Review, a set of figures that have become the most-cited numbers in the field of environmental policy, and quite possibly the most-cited numbers in any field of public policy. The panel, whose task was to assess the extent to which emissions of greenhouse gases may warm the planet over the coming century, reported that “globally averaged mean surface temperature is projected to increase by 1.4 to 5.8̊C over the period 1990 to 2100.” This alarming conclusion has become the starting-point for popular and official discussion of global warming and the policies that might mitigate it. Bear in mind how expensive some approaches to the problem, such as the Kyoto Protocol, might be if governments actually succeeded in implementing them. Vast sums are at stake.

As a rule, the IPCC is careful to attach warnings to its projections. Journalists are impatient with that: they prefer “prediction” to “projection” (less vague) and like to talk of temperature rising by “as much as 5.8̊” rather than quoting the full range. This is all very misleading—but the panel cannot be blamed for the way its work is reported. What it can be blamed for is the seriously flawed methods it has followed in making its estimates.

In recent months, two distinguished commentators—Ian Castles of the National Centre for Development Studies at Australian National University, formerly the head of Australia's national office of statistics; and David Henderson of the Westminster Business School, formerly the chief economist of the OECD—have put together a critique of the panel's Special Report on Emissions Scenarios (SRES). The report claims to “provide the basis for future assessment of climate change”, but Mr Castles and Mr Henderson point to serious flaws in its analysis and results. Last year they began writing to the chairman of the panel. Following an invitation to a technical meeting convened by the IPCC last month, they have offered further comments. The critique which thus evolved is to be published next month*.

One key problem with the IPCC's report, sufficient by itself for Mr Castles and Mr Henderson to declare the document “technically unsound”, is the way the scenario-builders have based their projections of future output on national GDP estimates which have been converted to a common measure using market exchange rates. This procedure leads them to overstate the initial gaps in average incomes between rich and poor countries—because prices tend to be much lower in poor countries. Those gaps are in turn crucial for the IPCC's projections, because the method used in the scenarios assumes not only that the rich countries will continue to get richer but also, in most of the 40 scenarios considered, that the greater part of the (overstated) initial gaps between rich and poor will be closed by the end of the century.

The combination of overstated gaps and of built-in assumptions about the extent of convergence in the average incomes of rich and poor countries yields projections of GDP for developing regions which are improbably high. Even the scenarios which give the lowest figures for projected cumulative emissions in the course of the century assume that average incomes in the developing countries as a whole will increase at a much faster rate than has ever been achieved in the past.

Miracles and anomalies

The unreality of the assumptions about economic growth in developing countries is highlighted by disaggregated projections which were recently released on the SRES website. These projections imply that, even for the lowest emission scenarios, the average income of South Africans will have overtaken that of Americans by a very wide margin by the end of the century. In fact America's per capita income will then have been surpassed not only by South Africa's, but also by that of other emerging economic powerhouses, including Algeria, Argentina, Libya, Turkey and North Korea.

The SRES summary for policymakers tells anxious governments that the 40 scenarios “together encompass the current range of uncertainties of future emissions”. Plainly, this is incorrect. The panel's low-emissions scenarios make exceptionally optimistic assumptions about economic growth in the developing world. But it is impossible to say, without running the whole exercise afresh, what the properly calculated range of projections for temperature changes would be.

Mr Castles and Mr Henderson offer a variety of other criticisms of the SRES, and of the panel's treatment of economic issues more generally. They complain, for instance, that history is too much neglected in the consideration of future trends. They also point out that developments in the first ten years of the scenario period, 1990-2000, were pretty clear by the time the SRES was published in 2000, and that in some respects they diverged substantially from the scenarios' projections; yet the report pays them little or no heed. Mr Castles and Mr Henderson argue that the circle of those involved in the climate-change exercise has been too restricted. For the future, the panel should draw on a wider range of economic and statistical interests and expertise. In particular, where its member governments are concerned, there needs to be a greater involvement of economic ministries and statistical agencies, alongside environment ministries.

The full panel meets next week in Paris to review the preparation of its Fourth Assessment Review. It should take the opportunity to consider the Castles-Henderson critique and resolve to do something about it.

Hot potato revisited

From The Economist print edition (Nov 6, 2004)

A lack-of-progress report on the Intergovernmental Panel on Climate Change


YOU might think that a policy issue which puts at stake hundreds of billions of dollars' worth of global output would arouse at least the casual interest of the world's economics and finance ministries. You would be wrong. Global warming and the actions contemplated to mitigate it could well involve costs of that order. Assessing the possible scale of future greenhouse-gas emissions, and hence of man-made global warming, involves economic forecasts and economic calculations. Those forecasts and calculations will in turn provide the basis for policy on the issue. Yet governments have been content to leave these questions to a body—the Intergovernmental Panel on Climate Change (IPCC)—which appears to lack the necessary expertise. The result is all too likely to be bad policy, at potentially heavy cost to the world economy.

In our Economics focus of February 15th this year, we drew attention to (and posted on our website) telling criticisms of the IPCC's work made by two independent commentators, Ian Castles, a former head of Australia's Bureau of Statistics, and David Henderson, formerly the chief economist of the Organisation for Economic Co-operation and Development (OECD) and now visiting professor at Westminster Business School. Their criticisms of the IPCC were wide-ranging, but focused on the panel's forecasts of greenhouse-gas emissions. The method employed, the critics argued, had given an upward bias to the projections.

The IPCC's procedure relied, first, on measuring gaps between incomes in poor countries and incomes in rich countries, and, second, on supposing that those gaps would be substantially narrowed, or entirely closed, by the end of this century. Contrary to standard practice, the IPCC measured the initial gaps using market-based exchange rates rather than rates adjusted for differences in purchasing power. This error makes the initial income gaps seem far larger than they really are, so the subsequent catching-up is correspondingly faster. The developing-country growth rates yielded by this method are historically implausible, to put it mildly. The emissions forecasts based on those implausibly high growth rates are accordingly unsound.

The Castles-Henderson critique was subsequently published in the journal Energy and Environment (volume 14, number 2-3). A response by 15 authors associated with the IPCC purporting to defend the panel's projections was published in the same issue. It accused the two critics of bias, bad faith, peddling “deplorable misinformation” and neglecting what the 15 regard as proper procedure. Alas, it fails to answer the case Mr Castles and Mr Henderson had laid out—namely, that the IPCC's low-case scenarios are patently not low-case scenarios, and that the panel has therefore failed to give a true account of the range of possibilities. If anything, as the two critics argue in an article in the subsequent issue of Energy and Environment, the reply of the 15 authors gives new grounds for concern. This week the IPCC is preparing to embark on its next global-warming “assessment review”—and if the tone of its reply to the critics is any guide, it is intent on business as usual.

It is true, as the IPCC says in its defence, that the panel presents a range of scenarios. But, as we pointed out before, even the scenarios that give the lowest cumulative emissions assume that incomes in the developing countries will increase at a much faster rate over the course of the century than they have ever done before. Disaggregated projections published by the IPCC say that—even in the lowest-emission scenarios—growth in poor countries will be so fast that by the end of the century Americans will be poorer on average than South Africans, Algerians, Argentines, Libyans, Turks and North Koreans. Mr Castles and Mr Henderson can hardly be alone in finding that odd.

Tunnel vision

The fact that the IPCC mobilised as many as 15 authors to supply its response is interesting. The panel's watchword is strength in numbers (lacking though it may be in strength at numbers). The exercise criticised by Mr Castles and Mr Henderson involved 53 authors, plus 89 expert reviewers and many others besides. Can so many experts get it wrong? The experts themselves may doubt it, but the answer is yes. The problem is that this horde of authorities is drawn from a narrow professional milieu. Economic and statistical expertise is not among their strengths. Making matters worse, the panel's approach lays great emphasis on peer review of submissions. When the peers in question are drawn from a restricted professional domain—whereas the issues under consideration make demands upon a wide range of professional skills—peer review is not a way to assure the highest standards of work by exposing research to scepticism. It is just the opposite: a kind of intellectual restrictive practice, which allows flawed or downright shoddy work to acquire a standing it does not deserve.

Part of the remedy proposed by Mr Castles and Mr Henderson in their new article is to get officials from finance and economics ministries into the long-range emissions-forecasting business. The Australian Treasury is now starting to take an active interest in IPCC-related issues, and a letter to the British Treasury drawing attention to Castles-Henderson (evidently it failed to notice unassisted) has just received a positive, if long delayed, response. More must be done, and soon. Work on a question of this sort would sit well with Mr Henderson's former employer, the OECD. The organisation's economic policy committee—a panel of top economic officials from national ministries—will next week install Gregory Mankiw, head of America's Council of Economic Advisers, as its new chairman. If Mr Mankiw is asking himself what new work that body ought to take on under his leadership, he need look no further than the dangerous economic incompetence of the IPCC.

Recent letter from Ian Castles:


Dear Colleagues,

In August 2002, I wrote to Dr. R K Pachauri, Chairman of the IPCC, to raise concerns about the Panel's "Special Report on Emissions Scenarios" (SRES) and the economic and statistical content of other IPCC reports. I sent copies of my letter to many colleagues in the international statistical community, government agencies and universities who had been following my ongoing correspondence about the use and abuse of economic statistics in reports by UN agencies and the World Bank. Subsequently some climate change scientists and officials who are involved in climate change issues asked to be added to the list of those who receive my messages. Thus this message is being circulated to a somewhat larger group than my initial letter to Dr. Pachauri. I apologise to any recipients to whom it is of no interest.

My purposes in writing now are to (a) send the text of an article in this weekend's "Australian Financial Review", in which I give a brief account of developments in the debate that began with my letter to Dr. Pachauri 18 months ago (it is attached as "AFR 7 Feb 2004.rtf") ; (b) provide the link to the IPCC's recent press statement referred to in the article (it is available at ) ; (c) comment on aspects of the IPCC press statement (this is attached as "Comment IPCC 8 Feb 2004.rtf"); (d) offer some personal views on the relationship of the debate about the emissions scenarios to the wider debates about the science and the politics of climate change; and (e) provide links, for those who may be interested in the details of the controversy, to the other documents referred to in my AFR article (see under "References" below).


Two days after the press conference at which he released the IPCC press statement saying, in effect, that there was no conflict between policies to mitigate climate change and economic development (see "Comment IPCC 8 Feb 2004.rtf"), Dr. Pachauri said in his address to the UNFCCC conference in Milan:

“Overall, climate change is expected to negatively impact development, sustainability and equity. As it happens, the impacts of climate change will fall disproportionately on developing countries and the poor persons within all countries. This is likely to accentuate inequities in health status and access to adequate food, clean water and other resources. The increased exposures of populations in developing countries to relatively high risks of adverse impacts from climate change and their low capacity to adapt combine to make populations in developing countries generally more vulnerable than populations in developed countries.”

These claims may be supported by statements in IPCC reports, but they are misleading in that there is an implied assumption that the future capacity to adapt of developing countries will be less than the present capacity to adapt of rich countries. The contrary is likely to be true.

A "scenario", according to the SRES, is "A plausible description of how the future may develop, based on a coherent and internally consistent set of assumptions ... about key relationships and driving forces ..." (p. 594). In their response to Castles & Henderson, the SRES Team give an example to illustrate the need to maintain coherence and internal inconsistency:

"Mr Castles and Mr Henderson obviously consider scenarios of a long term closure of the North South-income gap highly unlikely, and many (including a larger part of the SRES authors) would agree with them. But does this make it illegitimate to explore in a "what if ... then" scenario exercise the implications in terms of GHG emissions if indeed such development were to take place? ... The answer is obviously that such scenarios need to be considered ... The special value of the criticized A1 and B1 scenarios resides precisely in the insight that such an income gap closure might not necessarily be associated with extremely high GHG emissions but could also unfold even in the absence of climate policies with comparatively low emissions" (vol. 14, nos. 2-3, p. 196).

Of course it was not illegitimate to explore in the SRES the implications of developments that "a larger part of the SRES authors" thought to be highly unlikely. But it IS illegitimate to use these unlikely scenarios to project increases in temperatures; then to treat these projections as FORECASTS of increases in temperatures; and then to examine the implications that would have followed from those improbable increases in temperatures if the huge increases in average incomes assumed in the projections had not occurred. This is precisely what Dr. Pachauri did in his speech at the UNFCCC, and what IPCC Working Groups II and III repeatedly did in their respective contributions to the Third Assessment Report (TAR).

It is possible that scenarios will be used in a more consistent fashion in the AR4 than in the TAR, but the conclusions will still be flawed as a result of the IPCC's decision to persist with the unsound SRES scenarios in its next assessment report.

The world has been seriously misled as a result of the IPCC's temperature projections being treated as if they were forecasts, without regard to the assumptions underlying the emissions projections upon which the temperature ranges were based. For a particularly blatant example, see the summary of the WHO's publication (in collaboration with the IPCC's parents, the WMO and the IPCC) "Climate Change and Human Health: Risks and Responses" (2003), which is available at

Note, in particular, the label "IPCC (2001) FORECAST" (EMPHASIS added) in Figure 1.1; the description "21st century: very rapid rise" within the yellow block on the same Figure; the statement immediately under the Figure that "The IPCC (2001) has ESTIMATED that the global temperature WILL rise by several degrees centigrade this century" (EMPHASES added); and the statement in the text preceding the figure that "Climatologists FORECAST further warming ..." (EMPHASIS added).

Now contrast the 21st century segment of Figure 1.1 in the WHO publication with the corresponding segment of Figure 16 at the end of the World Meteorological Day Address 2003 by Dr. John Zillman, Australia's Director of Meteorology from 1978-2003 and President of the World Meteorological Organization from 1995-2003. This address is available at

In the text of his address, Dr. Zillman said that "In the IPCC community, 'climate change' means change on all time-scales, irrespective of the cause, and it thus includes both natural variability and any change that may result from human interference with the working of the climate system"; and that his chart "presents a number of possibilities for the twenty-first century ranging from substantial greenhouse warming on top of a naturally occurring warming trend to a future with only limited greenhouse warming offset by strong natural cooling which dominates the temperature trend through the second half of the century."

The range that the WHO Figure describes as an IPCC forecast is not a forecast at all. It is the IPCC's PROJECTION of the change in mean global temperature which would be attributable to anthropogenic causes if the modelled assumptions about climate forcing are valid and the projections of the driving factors influencing human-induced forcing are realised. To point this out is not to split hairs: the lower limit of the "band of uncertainty" that the WHO wrongly labels an IPCC forecast is well above the (negative) change at the lower end of the range of "possibilities" in Dr. Zillman's schematic chart.

Attention should also be drawn to the following sentence in section 1 of the WHO publication:

"During the twentieth century, world average temperature increased by approximately 0.6 deg. C, and approximately two-thirds of that warming occurred since 1975".

This is the IPCC's estimate from the instrumental record. But it is misleading to quote it without noting that, according to the same section of the WG I report, FIVE-SIXTHS of the increase of 0.6 deg. C in the twentieth century is estimated to have occurred between 1910 and 1945. The mean temperature is estimated to have DECREASED between 1945 and 1975. The WHO text gives the false impression that there was an exceptional surge in the last quarter of the century. With the additional information, it becomes clear that there is not a close match between the growth of GHG emissions and in estimated global mean temperatures.

Finally, note the last two sentences in section 5 of the WHO publication:

"Focusing attention on extreme events may also help countries to develop better means of dealing with the longer-term impacts of global climate change, although this capacity may itself DECLINE because of cumulative climate change. For example, increased food imports might prevent hunger and disease during occasional drought, but poor, food-insecure, countries may be UNABLE TO AFFORD such measures indefinitely in response to gradual year-by-year drying" (EMPHASES added).

These generalisations are grossly misleading, because the WHO focuses on the impact of changes in climate that (given other assumptions) would only occur in circumstances where average incomes and adaptive capacity would have greatly increased - but then asserts that this capacity may decline because of cumulative climate change. It is unscientific to focus on the implications of projected climate change and to ignore the implications of the factors that have been assumed to bring that change about.


The texts of letters that David Henderson and I sent to Dr. Pachauri in 2002, our presentations in Amsterdam in January 2003 and our other papers written at that time are available at (Click on the link in the footnote to the article). This swag of documents subsequently appeared in "Energy & Environment", vol. 14, nos. 2 & 3: 159-85. The response by 15 SRES authors in the same issue of "Energy & Environment" (pps. 187-214) is available at

A further article by Castles and Henderson in "Energy & Environment", vol. 14, no. 4: 415-35 is attached ("SRESRSPS.DOC.2.rtf"). The second article in "The Economist", supporting the Castles/Henderson critique and criticising the IPCC, is at A further response by 18 SRES authors will be published in "Energy & Environment, vol. 15, no. 1. (I do not have the text of this article in electronic form).

With best wishes

Ian Castles

Visiting Fellow
Asia Pacific School of Economics and Government
Australian National University


Summary: IPCC overestimates economic growth and warming by a factor of 2 to 4.