An Amateur Appraisal
So far we have avoided an appraisal of the critique of Chapter 6 that was offered at the time by the Global Commons Institute and other outsiders. We have also avoided an appraisal of the treatment of this critique by the expert authors in the IPCC process. While something should be said of the latter, it is difficult to avoid in such a discussion an evaluation of the economic methodology in question. When faced with the ravings of a ‘crank‘, with (as one interviewee advised) ‘little understanding about economic systems,’ there is only so much polite listening that can be expected of the expert economists called in to do the Assessment. All the more so when political motivation is apparent. Can we dismiss the GCI critique as a silly campaign of misinformation and abuse? Or does it contain something solid that hits hard at the science behind the Assessment? Answers to these questions require economics expertise to which this blog can only aspire. What to do? With a view to reduce misinformed criticism (and notwithstanding many other concerns) this final post on the Price of Life Controversy restricts discussion of economic methodology to the key concerns raised in the controversy at the time, including many noted in the Assessment Report itself. And indeed, if this appraisal is cut down by an expert reader in the comments below, then this posting has not been in vain.
David Pearce: Co-ordinating lead author of IPCC Working Group III Chapter 6, which assesses estimates of expected climate change damages. His findings support his view that moderate mitigation measures are required. He rejects changes to the Summary for Policy Makers made under pressure from government delegations. And he is especially angered when, in response to concerns about the reliability of the data, the Chapter’s percentage figures for damages at doubling of CO2 (1.5 to 2%) are replaced with the vaguer ‘a few percent.’
Aubrey Meyer: Founding director of the Global Commons Institute (GCI) and instigator of the Price of Life Controversy. His criticism of the economic methodology and the uncertainty in the data behind the conclusions of Chapter 6 are taken up by poor country delegations at the Working Group Plenary and the climate treaty talks. He believes that damages will be much greater than given in Chapter 6, and his critique supports the GCI campaign for the wealthy developed countries (who are mostly doing the damage) to act immediately to stop the warming.
The first thing to say about this controversy is that the integrity of the IPCC process was saved. Chapter 6 was not changed after its final drafting by the expert authors.
As for the Policymaker’s Summary (and the ‘Synthesis’), it does indeed present a critical perspective on the methodology and conclusions of the Chapter. Due to the intervention of the government delegates, the conclusions of the Chapter are couched within strong warnings. The section opens with:
The literature on the subject of this section is controversial and mainly based on research done on developed countries, often extrapolated to developing countries…
This is true and in accord with the Chapter, as are the proceeding warnings about the unreliability of the data. The emphasis these warnings are given only reflect the intergovernmental reception of the Chapter—which is entirely legitimate.
After this opening, the summary of Chapter 6 continues: ‘There is no consensus about how to value statistical lives or how to aggregate statistical lives across countries.’ This is partially misleading. Valuation tables do vary and one cited report (see: this summary) does value all lives at OECD levels. But as for the aggregation of these values across countries through conversion to US dollars, this is only disputed outside the reviewed literature.
Other changes reach beyond the Chapter content and into the plenary debate so as to introduce curious and distracting artifacts like:
…the value of life had meaning beyond monetary considerations
…the Rio declaration and Agenda 21 call for human beings to remain at the centre of sustainable development.
Perhaps it does and perhaps they do. But it is unclear how these statements even add or change anything. If they are meant to contradict the damages findings then they fail.
Indeed, the Summary does present all the Chapter’s key findings. And, while sometimes, and sometime curiously, it does reach beyond these findings, in doing so it does not directly contradict them. Thus, overall, considering the extraordinary level of disagreement between the authors and a whole bloc of delegations, this is a remarkably successful outcome for the science-to-policy process that is the IPCC—a credit to those, including Jim Bruce, who managed to hold it all together.
The next thing to say is that the authors had good reason for their differential monetary valuation of life. Like it or not, it is in terms of a global currency that government and inter-governmental bodies need to assess damages in order to determine how best to invest their limited resources. The valuation is for assessing the damages of climate change. It does not itself prescribe policy for responding to it. It is descriptive and not prescriptive. But anyway, even if wrongly interpreted prescriptive, it is still not easy to come to Meyer’s dark interpretation—a rationale for the genocide of impoverished nations. Let me explain.
At least with the IPCC, if not before, it was never going to be an all-or-nothing equation about whether we were going to stop global warming immediately in its tracks. Early intervention with ‘no regrets’ and ‘easy wins’ emissions reductions are more precisely identified due to this economic analysis. Moreover, with the twin finding of so many more poor lives under threat and their salvation so cheap, the economics of the Chapter suggests that spending money to save the poor is much more cost-effective than trying to saving the few among the rich. In all, it is hard not to be persuaded by a view common to those on the IPCC side in support of the authors. This is that the Global Commons Institute grabbed the opportunity to expose these ‘sensitive’ calculations, to interpret then crudely, and so to scandalise both the authors and their methodology in order to drum up opposition to the Chapter’s moderate conclusions.
While these two points need to be made in support of the process and the authors, they should not be used to veil some deeper problems with the Chapter, for they lie as though a thin cementing over a pile of sticks and straw. Probe a little deeper and the Chapter’s surface of scientific plausibility collapses into a jumble of chaotically aggregated quantitative data.
Consider first the great leap from the foggy and uncertain condition of the data field up into the confidence paraded in the aggregate damage figures. No climate change damages studies had been reported in agrarian economies nor in countries industrially underdeveloped—which is where most of the world’s population resides. The global assessment consists almost entirely of extrapolation from the exceptionally wealthy west. Even there, wildly different valuations are quoted for the same damages, including, no less, the base valuation for one statistical life (from 0.5 to 6 million). And wildly different sets of damages aggregate to suspiciously similar results. While critics were saying it is too early to make even the vaguest estimations, the result for the set of damages given by Nordhaus (1% US GDP), which exclude non-market factors, is well within the same ball park as those of Cline (1% US GDP) and Fankhauser (1.4% Global GDP) which include them.
Next, consider how the gaps in the data are treated. The Chapter 6 discussions of the estimations of damages in particular ‘categories’ (Agriculture, Sea level rise, Forests, Water supply….) explain how the quantitative results are given despite, and excluding, huge gaps in the data field. For example, the Health section (6.2.8) gives quantitative valuations that only include deaths due to ‘heat stress.’ Excluded are other indirect mortal consequences. But also excluded are all morbidity impacts that are not fatal. The unavailability of reliable data in what is assessed to be significant areas of damages might be a reason for withholding an aggregate result. But it is a distortion to give aggregated results with these unknowns areas of damage effectively valued at zero. (see: Meyer & Cooper, 1995 pdf )
The Chapter also presents confusion and contradiction in the assumptions underlying the various damages studies. This is not helped by the fact that Working Group I was in the process of downgrading its projections on warming and on sea level rise due to the recent introduction into the climate models of a dampener to simulate the cooling effect of sulphate aerosol emissions. The Chapter notes this problem but it can do little more than point out that the studies it is assessing are based on the 1st Assessment by Working Group I five years earlier [6.1.6]. The Chapter also notes that while sea level rise lags warming by centuries, damages are often assessed in ‘an arbitrary fashion’ for an immediate 1m rise upon the doubling of CO2—a figure that bears no particular relation to the estimated ranges of rise in either the 1st or 2nd Assessments. On top of all of this, in surveying the damages results for the most studied economy, the USA, the Chapter’s Table 6.4 [p203] clearly shows how the five studies base their work on three different assumptions of the equilibrium temperature rise for a doubling of CO2 (2.5oC, 3oC, & 4oC).
A Closed Loop
From the outset Meyer rejects the very idea of a differential US dollar valuations on human lives. As a methodology, he rejects global Cost/Benefit Analysis outright. If we don’t care to go beyond this rejection then there is not much to say about the GCI critique of the Chapter. However, they did critique the Chapter on various levels so that such concerns as those given above are elaborated in GCI pamphlets. Moreover, these concerns are summarized in at least one article for which Meyer was hoping to achieve peer review status. For whatever reason, it didn’t make it (1). But the more important point is that it could never have made it. There simply was no time. And this is where the incestuous ‘charmed circle‘ of IPCC lead authoring is most unsettling. Critics, amateur or professional, could be silenced or ignored for not peer review publishing their criticisms of papers cited by the draft report . Yet there was not enough time. And this was especially so when the cited articles had not yet been published.
In a practice that preceded the 2nd Assessment and continues with the 5th Assessment, it is quite normal for important papers informing the overall conclusions of the Assessment to be published after the drafts are circulated for review. Sometimes important references do not appear until after the draft has been finalized. Over in Working Group I the two most important papers behind the controversial positive attribution claims of the published Report were written by the co-coordinating lead author of the ‘attribution’ chapter (Ben Santer). But these papers were published long after the review process was complete and the final draft circulated. There was no opportunity for a skeptical critique to influence the review. The skeptical critiques did not appear in peer review journals – and could not appear – until long after the publication of the Assessment (see here).
Similarly in Working Group III: the two most important references for the controversy damages assessment were written by lead authors of the damages chapter, namely Tol and Fankhauser. They were both published in 1995, months after the review process was completed at the end of 1994—after which Pearce, following the IPCC rules, refused to make any changes. Thus, there was a closed loop insulates the Chapter authors from any unwanted criticism.
And Pearce was in no mind to break it. He is repeatedly quoted citing ‘the lack of published literature‘ as a reason for rejecting alternative valuations of life. This is all well and good for a scientific assessment. But it should also apply for the assessing authors’ own work. And it is all the more important that it does apply when there is no apparent regard for—and certainly no IPCC procedure to deal with—academic ‘conflict of interest.’ The counter-challenge that should have been presented to Pearce is that the two works championed in his Assessment were, at the time of assessment and review, also not in the published literature. (2)
The GCI did their level best to break this circuit. Before any review draft was available they anticipated its content in a critique of a working paper by Fankhauser (pdf). And they were right: in many important ways this paper did indeed anticipate the tenor of the Chapter 6 damage assessment. Then, after the drafts were circulated—with important references to papers yet to appear in the journals—GCI attempted to get their tabled criticism into a legitimate form…but it was all too late.
This charmed circle is not just about authors shaking off agitating activists and skeptics. It also works against academic rivals, and it works the other way. The cited papers by Tol are critical of Nordhaus’s published work, but yet Nordhaus had no time to respond. At the other end of the spectrum, beyond Cline, is a report of catastrophic damages by German academics, Hohmeyer and Gartner (see this summary). It had been submitted to the European Commission and published already in 1992. Some of the very different assumptions underlying their alarming results are criticized and rejected in the Chapter 6 assessment. And so it may. But the picture starts to emerge of a group of authors gaining control of the Chapter so as to use their own papers, and those of their students, to overly emphasize their particular situation in a spectrum of expert opinion. The truth is that Pearce made an assessment which headlined work contained in a couple of PhD dissertation drafts (and one derivative working paper) while down-playing the published work of others—including studies by recognized senior experts in the field such as Cline and Nordhaus.
Nordhaus would have been a sobering read in the midst of early 1990s alarmism. In Managing the Global Commons, published in 1994, he offers a degree of circumspection which is informed by an historical perspective on the current state of climate change science:
A difficulty seen in many studies of climate change is that people look for problems and ignore opportunities; it is as if there exists an unconscious impulse to find costs and ignore benefits of climate change. A comparison of two sets of studies is instructive in this respect. Almost two decades ago, a series of studies was undertaken to investigate the impact of flights in the stratosphere on global cooling. Studies by d’Arge and others…found that global cooling of 1oC would impose costs in a number of areas. Of the nine areas of costs identified in the global cooling studies…only two were examined in the EPA study (1989) of global warming and none were calculated by the EPA to produce benefits. The largest estimated cost in the global cooling studies was the amenity effect of cooling, determined through regional wage differentials. This topic was completely ignored in EPA 1989. One is tempted to say that environmental impact studies can find the cloud behind every silver lining. (3)
Nordhaus’ historical perspective also gives insight into the limitations of our knowledge of the future over the extremely long periods involved in climate change damage assessments:
Simply put, humans live, move, and die faster than climatic impacts are likely to be noticed. This point can be seen by asking what the effect today would be if one’s grandparents or great-grandparents had contemplated a warmer globe. Many of them would have given the prospect a loud hurrah. Can one sensibly talk about health effects when we don’t even know what next century’s major health problems will be or what the population distribution will look like? [p59]
These comments concerning the potential for adaption at a slow rate of change were particular pertinent in 1994, as the expected rates of change of a number of indicators had only recently been revised substantially down. Estimates given in the late 1980s (e.g. at Villach ’85 and Toronto ‘88) that the atmospheric concentration of CO2 might double as early as 2030 were pushed out by the 1st Assessment to the more likely date of 2100. Then during the early 1990s, the dampening effect of sulphate aerosols was pushing down the temperature increase by 2100 by one third, and the expected rate of sea level rise had come down by much the same proportion (see Chapter 6, 6.1.6).
Dialing up the Time Machine
For the uninitiated, one of the most incredible aspects of climate change damage estimations is the so-called inter-generational time-preference discounting. And it is hardly surprising that much of the debate over recent damages studies (Stern, Garnaut) has been over the level of discounting there applied.
The designated ‘discount rate’ is supposed to determine how much we value the estimated cost impact to future generations, whereby a low rate diminishes the value of future impacts more slowly. But there are no agreed parameters for its determination and the overall damage level is extremely sensitive to it.
The opening summary of Chapter 6 demonstrates how the choice of a discounting rate impacts on the variation in the estimated marginal cost of each extra tonne of CO2—which it gives as between $5 and $125. The reader is not mislead into thinking that this range might somehow reflect uncertainty in the data (we have already been told that the ‘confidence intervals’ of the quantitative results have not themselves been quantified). Instead, we are told that this range is mostly about the various selected rates of discounting. A difference of 3% in the discounting rate can compound into a difference in the marginal damages results of almost one order of magnitude (p183).
Within the circles of those advocating this method of inter-generational damages costing, the debate often reduces to the moral question of inter-generational equity, and about how much we do or should care for the welfare of future generations. However, those outside the debate often see an entirely different and overriding problem to which even Nordhaus alludes. This is that it is not so much an (immoral) disregard for future generations that would support high discounting, but more a recognition of our abysmal knowledge of the future and our dismal powers to influence it.
Imagine an Encyclopaedia salesman in the 1930s offering a purchasing-plan to my grandparents for the prosperity of their children. Now imagine that after his first success, the salesman pushes to sell the guarantee of a new edition for their prospective grand-children, and even a few sets (heavily discounted) for when their great-grand children arrive in the 1990s. Consider: What would be the over-riding consideration in their decision?
Even if grandma were sufficiently concerned for her descendants and she did recognize the value of encyclopedias, in hindsight we know it would be foolish for them to sacrifice some of their meagre wealth for my wealthy children. They could never have predicted Google and Wikipedia. But it is not only in hindsight that we all see the folly of such an investment. Just as they knew they don’t know the future, likewise today, we know that we don’t know whether some new form of energy will come to out-compete fossil fuels in our future.
Just as our great-great-grandparents in 1894 could not have predicted the conditions and concerns of energy production in 1994, why would we expect to know what they will be in 2094? Consider specifically the problem of urban transport pollution. It may well be that by 2094, without any helping hand, CO2 emissions will have diminished just as much as horse shit emissions diminished over the century up to 1994. And what if natural global cooling sets in again? With more certainty than anything else so far discussed, we know that the next ice age will eventually arrive.
Such skepticism of this inter-generational accounting should not be presumed to imply the disregarding of future generations in the decisions of the present. They are considered in all sorts of planning, all the time, only without such accounting and without such claims of extraordinary precision. In Australia we are reminded of such considerations by an exclusion zone larger than many European countries due to British nuclear testing 50 years ago. But the discount rates of radioactivity are much more certain than discount rates of the economic impact of emitted CO2. At every level the uncertainty of knowledge overwhelms inter-generational Cost/Benefit Analysis. Even at the level of the very history of the science. The history of Climate Change science since the late 1980s demonstrates the short half-life of our knowledge of something so uncertain as climate change and its physical impacts. The predicted rate of change and level of certainty are decreasing as the years go by…and this is before we even get to an assessment of the economic damages these (progressively moderating) impact estimations might cause. (4)
For those who take the high moral ground with concern for ‘our grandchildren,’ the hubris is double when they pretend to know the impact of our actions on the future while unable to establish the impact of our actions on the present or the past. The first question for these damage estimates should be empirical and historical: If we agree that there has been a warming of ½oC in the last century (anthropogenic or otherwise), then let us first identify the damage it has cause already before pretending to any idea of the future. In the Chapter (6.2) we are told this has not been studied. Why not? The reader needs to be reassured against the thought that even this empirical data is lost in the uncertainties and chaos of history—with its wars and depressions, with its revolutions, agricultural and technological, and with all those famously deadly winters.
Costing the Damage to Science
Without this reassurance about the scientific ground, it is hard not to follow Borg Lomborg and his Copenhagen Consensus and consider that there must be other areas where we have greater certainty about greater damages to human welfare and a real chance to avert them.
Without this reassurance and with the continuing disagreement over discount rates and social valuations, a skeptic would be concerned that the choice of parameter values might (consciously or unconsciously) serve to confirm pre-established conclusions. While the assessment of Chapter 6 serves moderate alarm, Hohmeyer & Gartner (and later Stern) give the sorts of results preferred by GCI. But how does one decide which is right? And this brings to mind a parallel controversy over social damage costings in the tobacco smoking debate.
In the 1980s, Cost/Benefit Analysis was used to justify taxation on tobacco to compensate especially for the cost of health care paid out by the community. However, in 1994 a new study claimed that the early death of smokers delivers a net social dividend. So, according to this CBA, smoking should be encouraged! Needless to say, the study was sponsored and promoted by the tobacco industry. But the question is not one of motives, virtuous or otherwise. Nor is it about who is right. Rather, it is about the lack of any solid, independent and widely agreed criteria of evidence upon which to ground the decision of who is right. In other words, there is no science. Until this situation improves, global inter-generational Cost/Benefit Analysis presents as no more than rhetoric, and worse, rhetoric to serve as propaganda.
Let’s not beat around the bush here. With the IPCC 2nd Assessment, scientists carried a solemn responsibility to serve monumental policy decisions. It was after the 1st Assessment, at Rio in 1992, that the IPCC was formally enlisted to inform the ensuing treaty talks. With the threat widely proclaimed catastrophic, with the ever-increasing demands for drastic global mitigation measures, this Assessment was perhaps the most important science-to-policy events up to that point in the entire history of science. And what happened? The science on ‘attribution’ was at best inconclusive and certainly controversial. And the same can be said for the damage costings. But instead of saying so, first Pearce, then Santer dragged their whole science into a political vortex, drunk with power and committed through a Faustian bargain struck by the weakest brother.
Santer’s Faustian bargain, undoubtedly more significant, tends to overshadow, and so unduly diminish, the tensions that had already surfacing elsewhere. We can see this through an editorial in Nature on 23 November 1995. This is at the height of the Price of Life Controversy but well prior to the commencement of Chapter 8 controversy. The headline ‘Global Warming Rows’ refers to two disputes scandalizing the Assessment months before Santer’s Chapter changes surface with the printing of the 2nd Assessment that following Spring.
After giving measured support to moderate warming alarm, the editorial repeats concerns arising from the Price of Life Controversy as well as from an embarrassing complaint to a US Congress committee by ‘bona fide scientists’ (i.e. Pat Michaels et al) over refusals to provide access to source data of key unpublished articles behind the Working Group I Assessment (more about this in another post). Nature concludes on one and then the other:
Good sense suggests that it is too soon, in what is certain to be a protracted argument, to begin putting numbers to the damages that may be done by global warming in such a literal fashion. This week’s rumpus [over access to source data] is more about the sense the technical critics of global warming have of their exclusion from the charmed circle of those with access to the most recent data from the climate models.
Nature’s final advise is that the IPCC…
…would do better if it abandoned its second and third working groups for the time being, and if it persuaded its first working group to follow a more judicial course. That, sadly, is not likely to happen.
To this day Pat Michaels continues to be lampooned as a Merchant of Doubt. Yet his attempts to investigate the evidence hidden behind the conclusions of a charmed circle in Working Group I were vindicated once the papers were finally published (see here). Indeed, his story stands as an unheeded omen of a larger controversy to come.
As for Aubrey Meyer’s intervention in Working Group III, the Chapter authors took our interloping fiddler for a fool. But, foolish or not, it took Aubrey Meyer to expose the folly, the monumental hubris of those pretending to put ‘numbers to damages that may be done by global warming.’ As far as he did this, and as far as he succeeded in influenced the IPCC into qualifying the finds of the expert authors of Chapter 6, he deserves our thanks.
As for Working Group III, for reasons that remain unclear (Someone please help me out here!), never again was a full IPCC Working Group committed to an economic assessment. From the 3rd Assessment onwards the IPCC reverted to the division of labor for the working groups as was originally devised for the 1st Assessment—physical science, impacts and mitigation. Economists continue to contribute to IPCC Assessments, but never again were they invited to enter upon such a broad stage.
Many folks have generously assisted with the research on the Price of Life Controversy. The author is especially grateful to the following: Richard Tol who answered many questions, providing a copy of the Montreal Plenary minutes and providing draft corrections; Sam Fankhauser, William Nordhaus and Erik Haites who responded to many queries; and Aubrey Meyer who responded to many queries and assisted in my navigation around the GCI website and archive. Others, including the IPCC’s Jonathan Lynn, assisted in the futile search for Plenary drafts and comments. John Zillman offered the first introduction to the controversy.
Any errors of fact in the posts are by the author. Notification of corrections are encourage and they will be applied (using
strike through) and acknowledged as soon as they are verified.
1. This paper is often cited as Meyer and Cooper, The Ecologist, September 1995. However, I cannot find it in any issue of The Ecologist around this time and Meyer has not been able to clarify the publication status. Sometimes it is given the status of a ‘working paper’ and thereby associated with The Ecologist. Note also an editorial on page 2 and 3 of the Jan/Feb 1996 issue of The Ecologist: ‘Cost-Benefit Analysis: the problem not the solution’ (in this archive). This editorial covers the controversy in support of the GCI.
- Sam Fankhauser’s thesis published as the book Valuing Climate Change
- Richard Tol’s damage assessment article: ‘The damage costs of climate change: Towards more comprehensive calculations‘ Environmental and Resource Economics, 5, 353-374, 1995.
It is often claimed that IPCC reviewers are provided with access to unpublished papers upon request, and this is indeed a requirement (although I am told that requests were rare and I am reminded that they were harder to honor in the early 1990s than they are today by email attachment). If this system worked then it would be fine so long as these papers were considered the status quo in that field. But it is generally because they are making new and unconfirmed claims that they are the subject of controversy in the first place. Moreover, until published, they remain in a limbo immune from formal criticism. The authors generally make a commitment to the accepting journal that their work has not previously been published and that it will not be published elsewhere after acceptance. They can usually present to peers at conferences but they are often asked not to speak to journalists, nor provide access to their data. The latter prohibition is precisely what was used to deny Pat Michaels access and so prevent a critical analysis of the science behind the novel conclusions of Working Group I prior to the publication of SAR (more in another post).
3. Quote from p59. An example of such a ‘cloud’ might be in the high valuation of VoSL due to ‘heat stress’ deaths. I recently asked one of the surviving Chapter 6 authors whether VoSL tends to measure the value of young and health lives, and if so whether this might inflate the valuation of damages due to ‘heat stress’ when it is likely to kill those that are already sick, already incapacitated, very old, or dying. He agreed that it would, as VoSL does not differentiate on ‘disability,’ and noted that now more sophisticated tools are available including one called ‘DALY.’ This and other instruments now measure the marginal value of an additional year of life.
4. This point is easier to make now, but it was already made by the early Australia skeptic Brian O’Brien back in 1990 and again, prior to SAR, in 1995. O’Brien noted the moderation over time (especially from Toronto ’88, to FAR ’90) in the predicted rate of man-made climate change. What concerned him was the reluctance to adjust accordingly the assessment of impact and so re-assess the implications for mitigation policy decisions. He makes the point again in 2000 after SAR and against Kyoto.