A Response to Bjorn Lomborg’s Response to My Critique of his Energy Chapter
John P. Holdren
15 February 2002
Bjorn Lomborg has posted on his Web page a long response to the critiques that appeared in Scientific American of four of the chapters in his book, The Skeptical Environmentalist, including my critique of his chapter on energy. No part of my critique escapes rebuttal. Perhaps Lomborg felt obliged to use all of the submissions he received in response to the appeal for help he broadcast to a long e-mail list after the Scientific American critiques appeared. It is instructive that he apparently did not feel he could manage an adequate response by himself. (In this, at least, he was correct. But he could not manage it with help, either.) Just as the book itself betrays a seeming inability of its author to discriminate sensible arguments from nonsensical ones, so also does the posted response to my critique suggest that Lomborg just tossed in, uncritically, whatever replies popped into his head or into his e-mail “in” box.
As my review for Scientific American acknowledged, Lomborg’s energy chapter does contain a number of propositions that are correct (such as the observation that there is large potential in renewable energy sources and energy-efficiency improvements). The problem with the chapter -- and the rest of the book as well -- is that, as a famously brief review of a long paper submitted to a professional journal once put it, “What is right in this document is not new, and what is new is not right.” The book is in fact a seemingly random admixture of points that are right and relevant (but not new), points that are right but not relevant (and still not new), and points that are wrong. Notwithstanding that the author is said to have been trained in statistics, the book shows no sign of the use of appropriate statistical conventions and methods -- or any other systematic approach -- to distinguish what is right and relevant from what is not.
Lomborg’s rebuttal to my critique of his energy chapter is of even lower quality than the chapter itself, in that there is almost nothing in the rebuttal that is both right and relevant, beyond the two places where he concedes that the critique caught him in a mistake. Even there, he protests predictably that the mistakes are modest in number and unimportant in the context of the sweep of his argument. But, as will be shown here, the mistakes caught in my review are both more numerous and more important than he concedes. It must be added that the space allotted for reviews is always limited, as it was in this instance in Scientific American, making it impossible to mention never mind to explain every mistake that has been noticed. It should also be understood that, even if space were not limited, few reviewers would consider it their responsibility to explain every error that a deeply flawed work contains, once they have explained enough of them to establish beyond doubt that the author is not competent in the subjects he is addressing.
In what follows here, I identify some (again, not all!) of what specifically is wrong or irrelevant in Lomborg’s rebuttal to my critique of his energy chapter, under the following headings: misrepresenting what I wrote in my critique; obfuscating what he wrote in the book; persistent conceptual confusions; refusal to acknowledge offering vagueness where specificity was required; and refusal to acknowledge offering illusory precision where only approximations are possible.
Misrepresenting what I wrote
In the first paragraph of my critique, I note that Lomborg spends most of his energy chapter attacking a view that few if any environmentalists hold – namely, that the world is running out of energy. I then write
“What environmentalists mainly say on this topic is not that we are running out of energy but that we are running out of environment (that is, running out of the capacity of air, water, soil, and biota to absorb, without intolerable consequences for human well-being, the impacts of energy extraction, transport, transformation, and use). They also argue that we are running out of ability to manage other risks of energy supply, such as the political and economic dangers of over-dependence on Middle East oil and the danger that nuclear energy systems will leak nuclear-weapons materials and expertise into the hands of proliferation-prone nations or terrorists.”
Of this passage Lomborg writes “This is exactly the kind of exposition which I try to counter in my book – without any references Holdren manages to describe everything as going ever worse and even include into the environmental agenda concepts that are far removed from its core, such as nuclear proliferation, terrorism, and economic recession from oil price hikes.” [emphasis in original]
As is plain, my comment is a characterization of what environmentalists say the energy problem is about, for contrast with what the book contends they say it’s about. I do not describe “everything as going ever worse”, as Lomborg claims I do, either in the passage he is commenting on or elsewhere. (Other reviewers have also commented that Lomborg has the habit of mistaking what people actually wrote with what he apparently wishes they wrote; it is difficult to avoid the impression that he is either unable or unwilling to pay careful attention to the content of what he is reading.) He is even wrong about the absence of references: the continuation of my argument in the next paragraph presents four references for the proposition put forward in the quoted passage – namely, that environmentalists have long held a more nuanced view of the energy problem than Lomborg attributes to them (and quite evidently more nuanced than he holds himself).
Obfuscating what he wrote
Concerning pollution from coal, Lomborg wrote in The Skeptical Environmentalist that “in developed economies, switches to low-sulfur coal, scrubbers, and other air-pollution control devices have today removed the vast part of sulfur dioxide and nitrogen dioxide emissions”. I criticized this statement, noting that the actual emissions from U.S. coal-burning power plants declined only from 16.1 million tons to 12.4 million tons between 1980 and 1998 in the case of sulfur dioxide and from 6.1 million tons to 5.4 million tons between 1980 and 1998 in the case of nitrogen oxides (mostly emitted as NO, not NO2, but by convention measured as tons of NO2-equivalent). I noted that these reductions, while welcome, hardly amounted to “the vast part” of emissions. In his rebuttal, Lomborg now claims that he was referring to emissions per ton of coal burned, not to total emissions, and since coal burning increased between 1980 and 1998 the percentage reductions in emissions per ton of coal burned were larger than the percentage reductions in the total emissions. The fact is, however, that the passage in his book says nothing about emissions per ton. Its plain wording about reducing “the vast part” of emissions means, on any straightforward reading, that the total emissions are now vastly smaller than they were before. This is not only flat wrong but conveys a seriously misleading impression about an important point.
Lomborg carps further, on this issue, that I was unfair to him in choosing 1980 as the starting point for the before-and-after comparison, and he claims that the numbers would have been more consistent with his “removal of the vast part” formulation if I had started instead with 1970. In fact, however, total emissions of sulfur dioxide were slightly lower in 1970 (15.8 million tons) than in 1980, and total emissions of nitrogen oxides were much lower (3.9 million tons). The New Source Performance Standards, which tightened emissions limits on coal-burning power plants, were enacted in 1979. Thus it seemed to me that starting with 1980 would be the choice kindest to Lomborg’s position. Of course, in this I was assuming that he meant what he wrote – just emissions, not emissions per ton of coal burned. Note also that the nitrogen oxide numbers, which Lomborg somehow neglects to mention in his rebuttal, would not remotely justify the “vast part” formulation even if one accepted his after-the-fact suggestion that he really meant to refer to emissions per ton of coal and even if one started with 1970.
Persistent conceptual confusions
In my critique, I point out that despite the preoccupation of Lomborg’s energy chapter with the question of resource depletion he fails to explain the distinction between “reserves” and “resources”, which is absolutely essential to understanding the depletion issue. His response, in which he criticizes my statement that “it is likely that the most abundant potential replacements for conventional oil will be more expensive than oil has been”, makes one wonder whether he understands the distinction himself. For he seems to think he can refute my point about the costs of replacements for conventional oil with his claim, referenced to an Energy Information Administration report, that today it is “possible to produce about 550 billion barrels of oil from tar sands and shale oil at a price below $30, i.e. that it is possible to increase the present global oil reserves by 50 percent.” Whether such a quantity can be produced from tar sands and oil shale at a price near (never mind below) $30 per barrel is in fact highly uncertain, but more suggestive of Lomborg’s confusion in any case is the fact that the price he mentions is higher (according to his own Figure 65) than the price of oil has been for any prolonged period in the last 120 years except for 1979-86, in the aftermath of the second (1979) Arab-OPEC oil-price shock. This means resources of tar sands and oil shale that would only be economically exploitable at prices around $30 per barrel are in fact more expensive than oil has been for nearly all of the last century. They could only be considered “reserves” -- material that is exploitable with current technology at current prices -- in circumstances under which the price of conventional oil had risen to well above what has been usual for the past century, which was exactly my point.
Another of Lomborg’s persistent conceptual confusions relates to his proposition that grid-connected wind power needs to be sized to meet the peak demand. My critique complained about the formulation to this effect in his book, and now he asserts that his formulation would be right for a system that consisted only of windmills and dams and that this is what he meant. The relevant passage in his book is not clear at all that what is meant is a grid containing windmills and dams but nothing else. But even in such an unusual power grid (which as far as I know does not exist anywhere) one would not size the windmills to be able to meet 100 percent of the grid peak demand. One would count on always being able to get some contribution from the dams, which after all can store up water (and thus the energy derivable from it) for whenever the peak load occurs. It would make no sense to try to scale the windmills to cover a case in which the dams could contribute nothing at all. In fairness, the details of how one calculates back-up capacity requirements for wind generators in power grids of various compositions are quite complicated. But Lomborg clearly hasn’t understood even the rudiments of the issue. As in so much of his energy chapter, he is confusing his readers because he is confused himself.
Vagueness where specificity was required
The least one ought to be able to expect in a book by a statistician is (a) clear specification of what is being depicted by the numbers that are presented and (b) appropriate indication of the magnitudes of uncertainties (as reflected, e.g., in the range of respectable estimates of a quantity of interest). Lomborg fails repeatedly to provide either one. As an example of his (frequent) failure to specify clearly what his numbers refer to, I cited in my critique his statement that “it is presumed that there is sufficient coal for well beyond the next 1,500 years”, without specifying the rate of coal use for which this figure would obtain. He now protests that it should be obvious that “the years of consumption are measured from the year discussed”. I invite the reader to return to the passage in the book to see how clear this is; but even more damning is that the formulation Lomborg offers now, following criticism that he muddled the point, is still not unambiguous: sometimes projected resource lifetimes are measured based on extrapolating a constant rate of growth of consumption into the future starting “from the year discussed”, rather than assuming a constant consumption rate, and sometimes they are measured “from the year discussed” by assuming that the consumption rate is following a bell-shaped curve (the famous “Hubbert’s pimple”).
Lomborg argues further that readers in doubt about what he meant should just consult his reference. But it seems a feeble defense to argue that it’s acceptable to be unclear in one’s book as long as the point is clear in a reference (which readers may or not be able to readily find and which, if they do, may or may not actually offer the promised clarity). Quite the contrary, it is the appropriate role of works of synthesis like Lomborg’s book to clarify and place in understandable context the quantitative details that are in the references, not the other way around.
His last resort is to protest that the exact number for the coal resource lifetime is not very important to his argument anyway, so I should not have bothered to criticize the lack of clarity about what it referred to. I agree the number is not important. My point is exactly that Lomborg’s energy chapter is a hodgepodge of such numbers – tossed together uncritically with little attention to clarity, accuracy, importance, or even relevance, never mind displaying or conveying understanding.
Illusory precision where only approximations are possible
Perhaps the most unexpected and inexplicable defect of all in a book by a statistician is The Skeptical Environmentalist’s almost pervasive inattention to the ranges of estimates that exist for quantities of interest – even inattention to the standard scientific practice of expressing quantities to a number of significant digits roughly commensurate with the precision to which the quantities are known. (I say almost pervasive because there are a few instances in which it apparently suits Lomborg’s agenda to emphasize ranges of opinion and uncertainty – most notably in his treatments of climate change and biodiversity – and then he does so.)
Among the examples of illusory precision in Lomborg’s energy chapter that I cited in my critique was his statement that “43 percent of American energy use is wasted”. In defending that statement, he writes in his rebuttal that “Of course, there are a lot of numbers that we do not know well, but the general idea in statistics is that if these numbers have been generated by a process described by evenly distributed errors, the more precise number is still the best predictor of the real number...” (It goes on.) This is gobbledegook -- a complete smokescreen. There is no set of numbers on the extent of waste in the U.S. energy system that any competent analyst would regard as data, never mind data that “have been generated by a process described by evenly distributed errors”, because there is no agreed definition of “waste” on which to base the generation of such data. Without a definition, there is no basis for saying whether 40 percent or 20 percent or 80 percent of U.S. energy is wasted, to say nothing of a two-significant-figure 43 percent.
Does “waste” mean the fraction of energy supplied that does not end up in the cooking pot or the lumen output of the lightbulb or the forward motion of the car? Or does it mean the difference between the performance actually achieved and the performance that is thermodynamically possible? Or does it mean the amount by which the energy being used to produce a given good or service could be cost-effectively reduced given current technologies and current energy prices? These definitions are very different, they give very different answers when applied to a given energy process, and it is an immensely complicated and difficult exercise to apply any of them to the energy-economic system of an entire country. Whole shelves of books, reports, and dissertations have been written about this. Nobody who had penetrated the slightest part of this literature would make Lomborg’s mistake of offering a precise number for the energy “waste” of a nation with no explanation of how this number was developed.
(I happen to agree with what I suspect Lomborg was trying to say, which is that the energy efficiency of the U.S. economy – defined as the value of the goods and services delivered divided by the amount of energy used – could be substantially increased in ways that would be cost-effective even at current energy prices. But that is not what he wrote. Tossing off a preposterously precise value for “waste” without specifying what he is talking about does not cut the mustard. With friends like Lomborg, improving energy efficiency will hardly need its many enemies.)
Besides his smokescreen about “evenly distributed errors” (a phrase in which he has inadvertently provided us with an apt characterization of his whole book), Lomborg’s further defense of his having presented an impossibly precise 43 percent value for a completely imprecise quantity is that he found it in somebody else’s book: “Moreover, the 43 percent is actually described right off one of the best-selling college environment books by professor Miller – is Holdren also claiming that he is wrong?” Lomborg’s citation is to the introductory-level text, Living in the Environment, by G. Tyler Miller, Jr. I have the 1975 edition of this book on my shelf, not the 1998 edition cited by Lomborg. In my edition, Miller does write that “over 50 percent of the energy use in this country is wasted”, and he presents a diagram showing a figure of 53 percent for 1973; but at least he then spends four pages explaining what he means by waste and conveying some sense of how a number in this range results from the definition he is using. If the treatment in the1998 edition does not reflect more of the range of serious thought and analysis on this issue that have taken place in the intervening 25 years, I would be surprised and, yes, critical. In any case, what is diagnostic in all this about Lomborg’s approach is that he appears to think that plucking one number uncritically from one reference (an elementary college textbook, at that) – and propagating it as established truth good to two significant figures – constitutes responsible scholarship. It doesn’t.
Another instance of the illusory precision in Lomborg’s energy chapter that I mentioned in my Scientific American critique – and which Lomborg seems to think is covered by his “evenly distributed errors” defense – is his statement that “the costs of carbon dioxide” are “probably…0.64 cents per kWh [kilowatt-hour]”. As with his energy-waste figure, Lomborg has a reference for this improbably precise number, in the form of a 1996 Resources for the Future report that reviews and summarizes a number of studies of the social (including environmental) costs of electricity generation. Consulting the reference is instructive. It does indeed contain the 0.64 cents/kWh figure cited by Lomborg, in a table labeled “Estimates and Best Guesses of Damages” with a footnote and accompanying text indicating that the figure is some sort of average of a considerable range of values presented in the report, which extend from at least 8 times smaller to at least 150 times larger. The text also contains the disclaimer that “Though the studies [reviewed here] devote considerable attention to global climate change from greenhouse gas (GHG) emissions, all conclude that damage estimates in the literature are too uncertain to be included with other estimates [of the social costs of electricity generation].” The reader may judge whether Lomborg has contributed to public understanding by suggesting, with this reference as his authority, that the cost to society from carbon dioxide emissions from coal fired power plants is “probably” 0.64 cents per kilowatt-hour.
Lomborg claims in his rebuttal that “Holdren could find little but a badly translated word and a necessary specification for nuclear energy production in this chapter”. Actually, as my original critique indicated to the extent practical in the space available, and as Lomborg’s rebuttal and this response make even plainer, his energy chapter is so permeated with misunderstandings, misreadings, misrepresentations, and blunders of other sorts that it cannot be considered a positive contribution to public or policy-maker understanding, notwithstanding its managing to get right a few (already well known) truths about the subject. Many of his mistakes are big ones: he bungles the issues involving reserves and resources that are critical to his core argument about oil remaining cheap; he drastically misleads his readers about the extent to which sulfur dioxide and nitrogen oxide emissions from coal-burning have been reduced; he trivializes the climate-change risks from coal’s carbon dioxide emissions by suggesting we know the impacts will only be worth 0.64 cents per kilowatt-hour. Other mistakes that are individually less important add up to a pattern of random incompetence. The sad fact is that Lomborg’s understanding of the energy issue is so superficial – and his reading of the literature he cites so uncomprehending and uncritical – that he is doing an actual disservice by trying to propagate what he imagines he has learned about it.
The Skeptical Environmentalist was glowingly reviewed, not long after its appearance in English translation, in The Economist, The Wall Street Journal, and The Washington Post. The reviewers for those publications were evidently pleased with the book’s message – “environmental problems are not as bad as we’ve been told, and things are mostly getting better” – but they evidently lacked the background or the inclination to find the flaws beneath the surface of Lomborg’s glib and citation-strewn presentation. Subsequent reviews by natural scientists – which have appeared so far in Nature, Science, Scientific American, and American Scientist, as well as on a number of websites – have been blistering. They have called attention to more or less “evenly distributed errors” across Lomborg’s treatment of population, food, forests, air pollution, acid rain, climate change, and biodiversity loss, among other topics – errors including all of the types I identified in his energy chapter, and more – even while acknowledging, as I did, that in this potpourri Lomborg manages to get a few things right. Now an Economist editorialist wonders (2 February 2002, p 15) “What has inspired [this] fury?” and goes on to accuse Scientific American’s scientist reviewers, in particular, of being “strong on contempt and sneering, but weak on substance”.
Those who have read this far can judge for themselves whether I am weak on substance, but I would like to try to explain to The Economist, and to others who may be curious, where the anger and, yes, contempt come from. The practice of science, which includes the packaging of findings from science for use in the public-policy arena, is governed by an unwritten code of conduct that includes such elements as mastering the relevant fundamental concepts before venturing into print in the professional or public arena, learning and observing proper practices for presenting ranges of respectable opinion and uncertainty, avoiding the selection of data to fit pre-conceived conclusions, reading the references one cites and representing their content accurately and fairly, and acknowledging and correcting the errors that have crept into one’s work (some of which are, of course, inevitable) after they are discovered by oneself or by others.
Most scientists follow this code of conduct as best they can out of self respect and respect for the integrity of science itself. For those for whom these considerations might not be quite enough, there is little that can enforce the code other than concern with the cumulative harm to one’s reputation and standing that comes from one’s colleagues’ awareness of a pattern of infractions, or fear of the public denunciation by colleagues that may follow in the rarer instances of someone’s descending into more massive and willful disregard of accepted standards. Of course, for the deterrent effect of potential denunciation by one’s colleagues to work against such massive violations of the scientists’ code of conduct, it is important that the denunciation should actually happen in those instances when, occasionally, the deterrent fails. If the issue involves science for policy, moreover, a clear and forceful denunciation has the further purpose of avoiding an extreme and poorly founded interpretation of the relevant science being credited in the policy debate as lying within the range of respectable scientific opinion.
Now, it is apparent from reading just the first few pages of The Skeptical Environmentalist that Lomborg proposes to make the case that not just environmentalists, but a considerable part of the heretofore respectable environmental-science community, have been misunderstanding the relevant concepts, misrepresenting the relevant facts, understating the uncertainties, selecting data, and failing to acknowledge errors after these have been pointed out – in other words, that the scientist contributors to what he calls “the environmental litany” (namely, that environmental problems are serious and becoming, in many instances, more so) have been guilty of massively violating the scientists’ code of conduct. This would be interesting news indeed, if Lomborg could prove it. But reading further reveals that his attempt to do so is itself a richly populated pastiche of these very infractions. Every class of mistake of which he accuses environmentalists and environmental scientists who have contributed to the “litany” is in fact committed prolifically and indiscriminately in The Skeptical Environmentalist (except, of course, for refusing to acknowledge error – for this, one has to read his rebuttals).
That the responses of environmental scientists have conveyed anger as well as substantive content, then, ought to be understandable. Lomborg’s performance careens far across the line that divides respectable even if controversial science from thoroughgoing and unrepentant incompetence. He has failed thoroughly to master his subject. He has committed, with appalling frequency and brazen abandon, exactly the kinds of mistakes and misrepresentations of which he accuses his adversaries. He has needlessly muddled public understanding and wasted immense amounts of the time of capable people who have had to take on the task of rebutting him. And he has done so at the particular intersection of science with public policy – environment and the human condition – where public and policy-maker confusion about the realities is more dangerous for the future of society than on any other science-and-policy question excepting, possibly, the dangers from weapons of mass destruction. It is a lot to answer for.
* * * * *
JOHN P. HOLDREN is the Teresa and John Heinz Professor of Environmental Policy and Director of the Program on Science, Technology, and Public Policy in the Belfer Center for Science and International Affairs, John F. Kennedy School of Government, Harvard University. He is also Professor of Environmental Science and Public Policy in Harvard=s Department of Earth and Planetary Sciences and Visiting Distinguished Scientist at the Woods Hole Research Center. Trained in engineering and plasma physics at MIT and Stanford, he co-founded in 1973 and co-led until 1996 the interdisciplinary graduate-degree program in energy and resources at the University of California, Berkeley. He is a member of the American Academy of Arts and Sciences, the National Academy of Sciences (NAS), and the National Academy of Engineering (NAE), and he chairs the NAS Committee on International Security and Arms Control and the NAS/NAE Committee on US/India Cooperation on Energy. He was a member of President Clinton's Committee of Advisors on Science and Technology (PCAST) from 1994 to 2001 and chaired PCAST studies on nuclear materials protection, the US fusion R&D program, Federal energy R&D for the challenges of the 21st century, and international cooperation on energy innovation. He has been the recipient of a MacArthur Prize, the Volvo Environment Prize, the Tyler Prize for Environment, and the Heinz Prize for Public Policy, among others. In December 1995 he delivered the Nobel Peace Prize acceptance lecture in Oslo on behalf of the Pugwash Conferences on Science and World Affairs, which he served as Chair of the Executive Committee from 1987 to 1997.
 In a “Dear Sir or Madam” broadcast e-mail sent out by Lomborg on December 18, he wrote, inter alia, “Naturally, I plan to write a rebuttal to be put on my web-site. However, I would also love your input to the issues -- maybe you can contest some of the arguments in the SA pieces, alone or together with other academics. Perhaps you have good ideas to counter a specific argument. Perhaps you know of someone else that might be ideal to talk to or get to write a counter-piece.”
 It is possible that this formulation is apocryphal, but that does not make it less applicable here.
 The world oil price also hit $30 per barrel for a brief period in the oil-price spike of late 2000. At this writing, in February 2002, the price is about $21 per barrel (current dollars, equivalent to a bit over $20 per barrel in 2000 dollars).
 See, e.g., the classic M. King Hubbert, “Energy Resources”, Ch. 4 in Resources and Man, Report of the Committee on Resources and Man of the National Academy of Sciences / National Research Council, W. H. Freeman, 1969.
 Representative works are the classic Efficient Use of Energy: A Physics Perspective (American Physical Society / American Institute of Physics, 1975) and the recent, massive, Lawrence Berkeley Laboratory / International Agency collaboration, Indicators of Energy Use and Efficiency: Understanding the Link Between Energy and Human Activity (OECD/IEA 1997).
 Allan J. Krupnick and Dallas Burtraw, The Social Costs of Electricity: Do the Numbers Add Up?, Resources for the Future Discussion Paper 96-30, August 1996, http://www.rff.org/disc_papers/pdf_files/9630.pdf.
 It is particularly ironic that Lomborg would offer such a ridiculously precise estimate of the cost of the impacts of climate change from carbon dioxide emissions, inasmuch as the entire thrust of his book’s chapter on “global warming” is that practically nothing about the effects of greenhouse gases is known with certainty.
 This is his way of characterizing the only two errors he has been willing to admit, namely writing “catalyzing water” when “electrolyzing water” was meant and misstating by a factor of two the contribution of nuclear energy in countries relying on it.
 As suggested above, once this pattern has been established a work can be considered discredited. A critic has no responsibility to identify and explicate all of an author’s mistakes. People with the competence needed to do this have better things to do. To explain to nonspecialists all of the mistakes in Lomborg’s energy chapter would require replicating a substantial part of the introductory course on energy systems that I taught for 23 years at the University of California, Berkeley, and have now taught for 5 years at Harvard. As badly as Lomborg needs that course, I am not going to provide it for him here.