February 6, 2007
In spite of economic sceptics, it is worth reducing climate risk
In the public at large, including sizeable sections of the business community, a new consensus on climate change has emerged: it is happening; it is important; and something needs to be done. The publication last week of the latest assessment from the intergovernmental panel on climate change and discussions at this year’s annual meeting of the World Economic Forum in Davos made the growing agreement on all these points plain. Yet there is one group among whom dissent reigns: economists. It was to them, above all, that Sir Nicholas Stern’s review on the Economics of Climate Change was addressed. It has failed to persuade. So much the worse for economists, the environmentally minded will declare. I disagree. Economists are trained to address the costs and benefits of alternative policies rigorously. Scientists are not. What then do economists object to in the arguments for early and forcible action to halt the increase in the stock of greenhouse gases? In essence, they make three arguments: first, the Stern review has exaggerated the economic costs of climate change; second, it has underestimated the costs of mitigating emissions; and, third, it has employed the wrong discount rate for relating near-term costs of mitigation to the costs of continuing on our present course. The remainder of Martin Wolf’s column can be read here (FT.com subscribers only). Discussion from our guest economists is free.
Jeffrey Frankel: I think Martin Wolf has it exactly right. (As does Martin Weitzman in his new review of the Stern Report.) Nicholas Stern may have been right for the wrong reason. It is as difficult to quantify the probability of the true climate catastrophe scenarios as to decide what exactly is the correct discount rate. But avoiding risk of global catastrophe is a more persuasive motivation than arguing that the proper rate of time preference is as low as 0.1.
As to the cost side of the cost-benefit calculation, it depends entirely on how we do it. If we allow trading in emission permits, both internationally and domestically (or substitute taxes on carbon and energy for other distortionary taxes), and if everyone participates, the true economic cost of successive Kyoto budget periods can be low. If we do it by clumsy command and control policies, the cost can be high. (Incidentally, if we do it by speeches about the miracles of technology, then it won’t happen at all.)
For once, I have to part company with the usual economists’ faith in cost-benefit analysis. The uncertainties around the discount rate, the catastrophe scenarios, and the costs are too high to make meaningful cost-benefit analysis work. But another difficulty is more serious still. Let us imagine we knew all these parameters, and could compute the optimal path in a Nordhaus manner. It would indeed probably start off with very small cuts in the near term, and ramp up deep cuts in the second half of this century, to give firms and technologists time to adjust to the new carbon-free future. This path would not be a useful guide to policy in the real world. Policymakers are not able credibly to bind their successors.
Posted by: FT Forum - Jeffrey Frankel | February 6th, 2007 at 9:51 pm | Report this commentAn announcement of deep cuts 50 years from now would not be believed, and would not much affect current behaviour. This point, more than any, is missed by most economists. The world (including in particular the United States) needs to take strong action in the near term, because it is the only way of convincing a rational observer that there will also be strong action in the long term. (http://ksghome.harvard.edu/~jfrankel/GEI.htm)
Lawrence Summers: Martin Wolf’s comments on the Stern Report are typical of a great deal of commentary that emphasises the insurance motive for aggressive anti-global warming policies and so endorses the conclusion of the Report and the views of those who commissioned it without defending the report’s analytics. This seems to be an entirely reasonable view.
I would be interested in his views and those of others in this online community on three questions that naturally arise:
Do you believe that in assessing major long lived policies like infrastructure that will last 100 years or major educational investments that discount rates like those used by Stern are appropriate? Should the World Bank and most governments change their practice of using relatively high discount rates of perhaps 8-10 per cent in project evaluation?
What is the best analytic case assuming normal discount rates for aggressive climate change policy? A corollary of Stern’s assumption that marginal utility declines only slowly with income is that risk aversion is not that great. Are there reasonable combinations of time preference, probability assessments on different global warming outcomes, and risk aversion for which urgent action is appropriate. I imagine so and Martin apparently believes that there are as do others who defend the report but it would seem helpful to spell these assumptions out.
Perhaps one way to get a handle on the issues here is to begin with some value judgements. What percent of GDP would readers be prepared to give up over the next decade in order to add annually a).01 per cent, b).05 per cent c) .1 per cent d).25 per cent to global GDP growth over the 100 year period 2020-2120?
Posted by: Lawrence Summers | February 7th, 2007 at 12:15 pm | Report this commentPaul Seabright: I agree entirely with Martin that the way to think about policies to counteract climate change is as a form of insurance, not as just another investment decision. But the Stern report does not frame the choice in this way. On the contrary, its analytical centrepiece is a model of growth in which utility is the logarithm of consumption, and the pure rate of time discount is only 0.1 per cent.
The logarithmic form means that if our descendants are ten times as rich as we are, it will take an extra ten units of their consumption to justify the sacrifice of one unit of consumption today – which looks rather egalitarian at first sight. But there will be so many of these descendants, discounted by so little, that they overwhelm the egalitarian reasoning by sheer weight of numbers: transferring consumption one-for-one from us to them is justified so long as there are eleven or more generations to benefit. The risk of catastrophic outcomes if we do nothing is indeed discussed in the report, but its impact is bundled up into an overall estimate of the effect of “business as usual” on the average income of our on average much richer descendants.
This seems to me a mistake, both ethically and politically. Should we cut emissions today to reduce the risk of catastrophic outcomes tomorrow? Absolutely, if the risk is important enough and the outcomes bad enough. Should we cut emissions today to make some very rich people even richer, just because there will be a lot of them? Much more doubtful - and surely much harder to sell to the citizens of India, China and other developing countries, who will be among those making the sacrifices, and some of whom are now very poor indeed. Apart from anything else, we don’t even really know what the difference will be like between being nine times as rich and thirteen times as rich as we are today: current work in behavioural economics suggests that it could be something of an anti-climax. The question is surely not to be settled by a technical assumption about a parameter on a utility function chosen for its suitability for generating analytical solutions. It’s a question of shared ethical and political values that ought to be out in the open.
Unfortunately this is not the only part of the report where some very big issues are hidden in the details. Discussing “the challenge of stabilisation”, the report states that “stabilisation - at whatever level - requires that annual emissions be brought down to the level that balances the Earth’s natural capacity to remove greenhouse gases from the atmosphere.” Even on reasonably optimistic assumptions this is around 5 GtCO2e, less than 20 per cent of current emission levels. But thereafter the report discusses time paths to stabilisation in 2100 in terms of reductions of emissions to around 20 GtCO2e, which is not stabilisation at all but a level four times the rate compatible with stabilisation.
The date at which the “headline” cost of stabilisation is calculated is 2050, by which time annual emissions will, even on optimistic assumptions, have fallen to a level that is still six times that compatible with stabilisation. The calculated costs are therefore based on a partial policy which does not even deliver stabilisation or anything close to it. The real cost of a complete stabilisation is presumably substantially larger.
Bringing these technical assumptions out into the daylight need not lead to less urgent or less demanding policy recommendations. Perhaps the fact that complete stabilisation requires such massive cuts in emissions makes the case for acting now stronger rather than weaker. Perhaps taking the insurance motive seriously would imply larger rather than smaller reductions in the short term. It’s hard to tell on the basis of what the report gives us. The cautious reaction so far of economists to the report seems to me appropriate, even if the profession comes round in the end.
Posted by: FT Forum - Paul Seabright | February 7th, 2007 at 4:29 pm | Report this commentJohn Williamson: As one of the few economists who would happily spend way over 1% of GWP combating global warming, I want to explain why in terms of the economic logic set out so clearly by Martin Wolf in this week’s column.
The case in favour rests on the proposition mentioned by Martin Wolf that the costs are not all measurable in terms of GDP. The costs in terms of what is measured by GDP seem quite reasonably estimated, or conceivably they are somewhat over-estimated. The problem lies rather in terms of what is not included in GDP: the destruction of the natural environment that is in prospect, in terms such as the loss of species diversity. It would certainly be good to have better measures of just how much we stand to lose, but it is clearly a great deal even if the assertion that in the course of the next century we are in danger of losing one half of all the species inherited from the past proves to be an exaggeration. It is indefensible to refuse to take action until the case has been made reasonably certain. (Remember what Martin Wolf wrote about insurance: indeed, one does not insure against certainties.)
The cost of more vigorous policies to combat global warming will be felt in terms of diverting resources that could either be consumed now or could be invested to yield more output in the future. In either event there would be a loss of intertemporal welfare as it is usually measured by economists. For the poor, it is easy to agree that this would be a catastrophe. But it is possible to design policies to combat global warming in a way that would not inflict a net loss upon those beneath a specific poverty line. After all, one of the features of the Kyoto treaty was exemption of developing nations, on exactly this logic. Most economists would agree that the mechanisms involved were inefficient, and that it would be easy to do better through the use of taxation. But the relevant point is that for the non-poorit is dubious in the extreme to attribute any material loss in welfare to a reduction in either present consumption or its future growth rate. One rather firm conclusion from happiness analysis is that beyond a threshold of maybe $15,000 per year increases in measured GDP per head count for essentially nothing in terms of enhancing welfare. It is quite right to be prepared to compare costs and benefits rigorously, by the use of discount rates and distributional parameters, but where the benefits could be enormous and the costs can be heaped on those in a position to pay them, the exercise is superfluous.
The two points made above are logically distinct. Even those who resist the conclusions of happiness analysis should be prepared to admit the importance of costs not measured by GDP. It is the reluctance to do so that so distresses the public about economists, but it is not inherent in the discipline.
Posted by: John Williamson | February 9th, 2007 at 11:24 am | Report this commentAndrew Oswald: It may be helpful to think about the problem differently.
I believe we could and should work out a way for the next generation to pay. This is because it is those millions of unborn citizens who will benefit from a cooler world, and because doing so will solve the problem at the crux of the Stern-Nordhaus-Dasgupta debate, which is really the question: why should the current generation pay generously and pay quickly?
This means it is necessary somehow to design a way of bringing the future, with their cheque books, to the negotiating table
Here is a suggestion.
We print a government bond called a Global Warming Bond. These have stamped on them: “I pay out 1000 indexed pounds in every year - beginning in the year 2050 and going on forever”. These bonds would be given out, as a subsidy, to those people and organizations who reduce emissions today.
The bonds would have immediate value. A market in them would spring up. I shall assume that their status as government bonds would make risk of default negligible. One might object to this, but I shall leave it at that.
The attractive thing about these bonds is that (leaving aside technical issues about general equilibrium reallocations across asset classes) they would be funded essentially by future taxpaying citizens. Those earning and paying taxes in 2050 onwards would fund them. Our citizens, in 2007, would gain.
In this way, the unborn would subsidize us to cut carbon emissions.
I offer this as an idea. There may be a flaw in it; one would have to think through a formal model. In economics, we are used to tax neutrality results coming back to bite us when we hope they will not. But perhaps the GW Bond idea suggests an avenue of ideas on these problems.
Posted by: Andrew Oswald | February 9th, 2007 at 3:18 pm | Report this commentStephen Cecchetti: I am in complete agreement with the Martin Wolf’s views on the economics of global warming – that we need to do something now in order to reduce the probability of future catastrophe. In thinking about this, however, we face some important challenges beyond the technical issues on which many of my colleagues are now focused.
First, the vast majority of people really don’t understand risk. We put resources into making ourselves feel safe, not where we really are safe. Second, we are very bad at discounting the future. Most of us are extremely impatient and exhibit behavior that is inconsistent — they can’t keep themselves from doing things that they will surely regret. Saving for retirement is a great example. It’s a long way off, so you can always save for it later.
The inability to understand risk combined with our impatience make it extremely difficult to have a rational conversation about the consequences of global warming. Most of us can’t fathom the risk — we have no appropriate metrics to evaluate it. And most of us aren’t very good at discounting the long off future. So if you look at the debate, it’s about things like discount rates, growth rates, and difficult to estimate consequences that are far off into the future, but with costs that must be paid sooner rather than later.
So, how do we get our government to deal with all of this properly? My biggest fear is that we lack any sort of institutional apparatus for facing the problem. Not only are governments run by people with the same blind spots as the rest of us – the inability to evaluate risk and discount the future sensibly – but, as Jeffrey Frankel notes, today’s policymakers cannot commit the future’s policymakers to act.
I believe that facing this problem means stepping back from the technical arguments, and instead discussing how to design the right organizational structure to deliver the right policy. In the case of monetary policy, we solved the tendency of politicians to create inflation by making central banks independent. In the case of global warming, I believe that we need to be thinking along these same lines. What we need is a new institution, populated by technocrats, independent of political influence, charged with formulating an appropriate public policy
To paraphrase Inspector Harry Callahan, the San Francisco police officer played by Clint Eastwood in the movie Dirty Harry, we need to know our limitations. In this case, the limitations of our representative democratic structure will likely prevent us from taking the actions needed to reduce the growing threats of global warming.
Posted by: Stephen Cecchetti | February 9th, 2007 at 10:34 pm | Report this commentMartin Wolf: I want to thank the participants in the forum for their very interesting comments. I am feeling my way on this subject, so I find everything illuminating.
Let me start with the comments that seem to me entirely sensible (no doubt, because the authors express agreement with me). Jeffrey Frankel’s comments fall in this category. I agree, in particular, with his proposition that a commitment now to higher prices 50 years hence would lack much, if any, credibility.
Steve Cecchetti also agrees with me that we should act to reduce the chance of an irreversible process leading us into a world 5ºC or more warmer than today’s. He shares my view that we have no right to run such experiments with the only planet we have provided we can limit the risk at reasonable cost.
John Williamson emphasises an important issue that I touched only lightly in my column: what value should we put on things not included in GDP? I agree strongly with him that the destruction of flora and fauna matters even if that were not to come at the price of a lower GDP. I know that life will survive warming. But that does not mean humanity is entitled to act as the world’s grim reaper. Nor is there any convincing way of valuing “natural capital” against physical capital. You can ask people what they would pay for a species or look at what people pay to visit wild life. But if you tell non-economists that a species can be valued in terms of office blocks or cars, they will think you are crazy. They are right. This is an ethical decision, not a technical one.
Stepen Cecchetti and Andrew Oswald raise questions about institutions. Steve’s are rather fundamental. He asks whether democracy can deal with these sorts of problems. His answer is: no. He may well be right.
One can raise an even broader issue. Humanity has now come to dominate the entire planet and has created a global society of a (highly dysfunctional) kind in the process. As a result, it now has to provide a host of global public goods: security; economic development; maintenance of a liberal global economy; management of the global commons; etc. Do we have the institutional structure to manage such challenges? The answer is no. Just think about nuclear proliferation, nuclear terrorism or global warming.
So there are two institutional challenges: the first is how to make decisions at a global level - the collective action problem; and the second is how to make long-term decisions at all - the defective foresight problem. Steve focuses on the latter. I would suggest we should think about both.
I would also suggest that there are three distinct problems with democracy: one is the failure of telescopic vision; another is the inability to commit to policies that take effect in the far future; and the third is the lack of relevant knowledge in the body politic. Steve focuses just on the last.
The Chinese would certainly agree with Steve’s doubts about democracy. But I still hope that we can combine democracy with more effective decision-making. It will be hard, particularly at a global level.
Finally, in this round, let me comment on Andrew’s imaginative suggestion. Assume people cut their emissions now in return for these bonds. Then, in aggregate, people today will enjoy somewhat fewer goods and services. But they accept this voluntarily. The question then is this: will the future keep its promises? If its debt became big enough, it might default, instead.
Posted by: FT Forum - Martin Wolf | February 12th, 2007 at 6:17 pm | Report this commentMartin Wolf: In this second comment, I would like to respond as best I can to Paul Seabright and Larry Summers, who raise some fundamental issues about the Stern report.
Paul argues that the form of the utility function employed in the review determines the outcome. As he notes, under the form chosen, a doubling of consumption halves marginal utility. So if there are twice as many people in the next generation, enjoying twice the average consumption, there is a balance between the loss of one unit of consumption for everybody today and the gain of one unit of consumption for everybody alive then. Since the pure time discount factor in the Stern review is low (0.1 per cent a year) and the number of future generations is (we hope) enormous, there is a strong preference for taking costs today even if future generations are much richer than we are.
Furthermore, as I noted in my column, once one puts in the assumptions about growth in the review, even the most catastrophic outcomes it considers deliver very large increases in average incomes. But on its assumptions, future generations could easily compensate contemporary losers out of their own incomes. It is not obvious why the much poorer people alive today should act to reduce these losses, since they will not be compensated for doing so (except by a warm feeling of virtue).
Paul concludes that this way of constructing a case for decisive action is weak. I am inclined to agree. He also notes that the costs of mitigation are biased downwards because they do not deliver stabilisation of the stock of greenhouse gases at 2050 or 2100. So he concludes that the costs of mitigation are underestimated, just as the benefits of doing so are exaggerated.
I find these points potent and return to the underlying argument in the last two paragraphs below.
Now let me turn to Larry’s three queries.
First, Larry asks what discount rate one should use for a project with a life of 100 years or more. I think the answer is that I do not know. How should one compare, say, a 40-year project to a 100-year project? The difficulty is that one does not know what the social rate of return will be between 40 and 100 years from now. If the reinvested profits from a shorter-term project end up earning lower returns than those invested in a longer-term one now, then the latter may be the right choice even if the returns on a short-term project are higher than on the longer-term one (over its shorter life).
The return on capital over the next two centuries is unknown. Economists assume that things will continue as now. That is possible. But it is also possible that the exhaustion of raw materials and, above all, of fossil fuels will drive the rate of return down to low levels over such a long period. To put the point another way, the assumption that real incomes per head will rise at 1-2 per cent a year for another two centuries may just be wrong. How many people believe that GDP can rise at the last century’s rates for, say, 10,000 years? When, then, will it stop? So even if one takes the opportunity cost of capital as the relevant discount rate, it is unclear what rate to use for very long-lived projects.
Second, are there any reasonable assumptions with “normal” discount rates that justify early action? I believe there are, though I (confidently) suspect Larry will disagree.
Assume that, on average, people two or three centuries from now will not be much richer than we are (this is conceivable, since we don’t now know what will power a far richer economy when fossil fuel runs out); assume a large value on loss of natural capital (i.e. mass extinction and loss of habitats) and assume that climate change may well generate huge local disasters, in particular massive shifts in rain-fall across the planet. Then we would probably decide we should act now. In this world, of course, rates of return and so conventional discount rates will be lower than we are used to. So Larry may well feel I am cheating. The crucial points here, I think, are two. First, how confident do we feel about prospects for growth in the very long term? Second, how happy are we to believe that aggregate GDP is a good measure of welfare?
Third, Larry asks us to decide how much we would give up now for somewhat higher GDP in future? Well, the risk-free real rate of interest is now about 2 per cent. So people are demonstrably willing to trade off income now against income in future at that low discount rate. In fact, one of the big problems we have, as Harvard’s Martin Weitzman shows in his forthcoming review of the Stern report for the Journal of Economic Literature, is that there is such a vast and inexplicable gap between the return on capital and the risk-free discount rate, as measured by government bonds. He concludes, if I understand him, that which rate to use is not obvious: it depends on the nature of the risks. All this ignores the ethical question of whether actual behaviour gives the “right” discount rates.
I would now like to return to what I see as the big point. This is about insurance under extreme uncertainty. We don’t know what will happen to the world economy over the next two hundred years. So we don’t know what the rate of return on capital will be over this extended period. We also don’t know how to value natural against physical and human capital. And we don’t know who will be most affected by climate change. But we do know that this “project” is very different from the normal investment project, since it affects the globe as a whole. We also know that the effects of our decisions may be irreversible.
If we make the wrong decision, we may bequeath a big mess to our descendants. In such a situation, we should surely err on the side of caution. If it turns out climate change was a scare, we can abandon the policy. We shall know fairly soon (in a matter of decades) whether it is indeed just a scare, as sceptics argue. If it turns out to be true, however, we shall be on roughly the right path. So, given the uncertainties, it is sensible to start acting now, provided we ensure that the costs fall mainly on the rich and are not excessive.
I have one final comment. The latest IPCC report seems to have halved the prospective rise in sea levels over the next century to a maximum of 0.59cm (24 inches). Under most of its scenarios it is much less. It may well be that, as the sceptics argue, the proposition that climate change is a potential disaster will wither away over the next decade or two. In that case, we will still have improved fuel efficiency, which should be useful: it will give us longer to develop good alternatives to fossil fuels.
Posted by: FT Forum - Martin Wolf | February 13th, 2007 at 9:29 am | Report this commentMartin Wolf: I have received the following comment from a correspondent on my last point above. It seems interesting.
“Dear Martin Wolf,
You write in today’s comment:
This is incorrect. The scientists at Real Climate, several of whom are IPCC authors, observe:
The upper bound of this effect is 20 cm, which gives 79cm in total — not a large change from the Third Assessment Report’s 88 cm. Also, this is not a “maximum”, in the sense that a zero probability is attached to higher values. There is nothing in the Fourth Assessment Report that precludes higher values but the uncertainty is very large.
Yours,
Per Klevnas”
Posted by: FT Forum - Martin Wolf | February 13th, 2007 at 12:28 pm | Report this commentBy Invitation - Claude Henry: In continental Europe as well, Nick Stern’s report is seriously discussed. Let me draw on some among these discussions, involving Italian and French economists.
1. Evaluation of mitigation costs and of damages from climate change.
They both appear to be underestimated. First some costs, that might tentatively be called transaction costs, are not included. Here is just one example: while the materials and techniques already available, at competitive prices, to renovate existing houses and flats make possible reductions of energy consumptions by a factor of three, most architects, builders and craftsmen right now are neither interested in making use of them nor yet able to do so. It will take time and resources to change this state of affairs.
Second, and more significant, there is a reason for underestimation that is deeply rooted in the nature of economics itself; it bears mostly on damages. Improving upon previous climate change studies, the Stern report includes in its estimations non-market effects that can nevertheless be evaluated in economic terms using proven indirect methods. But even these methods leave important, sometimes paramount, effects out of reach of economic evaluation. Think of the following one:
North Africa seems particularly vulnerable to climate change; recurrent droughts like those crippling southern Australia might easily change the narrow inhabited belt north of the Sahara desert into a mere extension of the latter. Where would the affected tens of millions people seek refuge? To the south is the Sahara, to the west the Atlantic Ocean, to the east the Libyan and Egyptian deserts. The only possible destinations are north, to Italy, France, Spain and beyond. Political correctness and moral sentiments tell that these refugees should be accommodated in the best possible ways. Realism suggests that they will be fought against with all available weapons, for Europe at the cost of democracy and even basic self-respect.
While mitigation costs seem to be somewhat underestimated (but not greatly underestimated; see the IEA 2006 World Energy Outlook) in the Stern report, damages from climate change seem to be vastly underestimated. In this the limits of the report merely reflect the limits of economics.
2. Discounting
In the important matter of discounting, the Stern report’s approach is not perfect. But it seems a better approximation to a really appropriate and consistent approach than may be found in other economic studies when significant environmental issues are at stake.
The consistency I have in mind has been advocated since a long time by Edmond Malinvaud. It can be (roughly) summarized in the following way.
Consider an economy where successive generations consume two different kinds of goods and services, H and N, produced respectively from artificial and natural assets. In the course of economic development, artificial assets are expanded and natural ones are eroded. The preferences and needs of the successive generations are such that you can expect (now, at instant o) that those twin movements of expansion and erosion will result in diverging price trajectories : prices for goods and services N will increase, possibly sharply, relatively to prices for goods and services H. Malinvaud’s point is that it is logically inconsistent and economically misleading to choose a discount rate (be it 1, 2, 3, 4, 5 per cent or more) – which is currently defined with respect to a numeraire built from H – without simultaneously considering a trajectory of relative prices for N. Both the numerical value chosen for the discount rate and the numerical values along the chosen trajectory of relative prices are open to discussion: but it is definitely wrong to use one without the other. In his great 1953 Econometrica paper, Capital Accumulation and Efficient Allocation of Resources, Malinvaud shows that efficient decentralized economic decisions cannot be reached from the mere displaying of intertemporal prices for one good; they require the displaying of a system of intertemporal prices on an appropriate set of goods (aggregations of goods are possible as long as their intertemporal prices are not too diverging).
On the basis of a two-sector H and N model (Calcul économique et développement durable, Revue Economique, 2004) and of corresponding preliminary calculations made in the context of the Stern report, Roger Guesnerie reaches conclusions similar to Stern’s as regards the balancing of interests between the successive generations affected by the perspectives of climate change. This is due in part to the way the discount rate is chosen and used in the report; indeed lowering the H discount rate partially corrects the lack of disentanglement between H and N.
It is doubtful that Frank Ramsey (A Mathematical Theory of Saving, Economic Journal, 1928) was right when denouncing discounting as “ethically indefensible”. But as for “the weakness of the imagination” in the matter, he was doubtless right. Economists who are in the mood of criticizing Stern’s intertemporal approach should do more than merely recycling standard flawed arguments about discounting.
3. Risk and uncertainty
Aversion to risk is a factor amplifying the perception of looming damages. Hence it must be taken into account in the calculations of damages. This is done in a much more systematic and consistent way in the Stern report, than it has been in any previous assessment of climate change perspectives.
It is, however, the case that many among these perspectives are uncertain rather than risky, according to the distinction made in 1921 independently by Keynes and Knight Mathematicians and economists, in Italy, the USA and France, have recently generalized the von Neumann-Morgenstern model from risk to uncertainty (see for example P. Klibanoff, M. Marinacci and S. Mukerji, A Smooth Model of Decision Making under Ambiguity, Econometrica, 2005 ; for an introduction to decision-making under uncertainty see C. Henry, Decision Making under Scientific, Political and Economic Uncertainty, http://ceco.polytechnique.fr/CDD/PDF/DDX-06-12.pdf, to appear in a volume published by the Swedish Academy of Sciences). Their results, including the role of the so-called aversion to ambiguity (i.e aversion to uncertainty), have already been used to clarify some puzzles in insurance and finance.
The Stern report mentions these results, and stresses their relevance in the context of climate change. However it doesn’t make use of them ; it would indeed have been premature to apply them quantitatively in conditions as complex as those of climate change. Qualitatively it is nevertheless possible to say that, in general, aversion to highly uncertain and damaging perspectives is larger than aversion to similar perspectives when reliable probabilities are known. This constitutes yet another reason to consider that damages are underestimated, not overestimated, in the Stern report.
4. Science and technology
The range of scientific and technological tools – either already available or at a stage in development that makes it realistic to expect full availability within less than seven to eight years – is significantly broader than currently appreciated. This is particularly so as regards energy consumptions and carbon dioxide emissions in the housing and office building sectors. Many examples may also be mentioned from various industries; here is just one. To start and sustain many chemical reactions, you need either high temperatures (hence high energy consumptions), or appropriate catalytic substances; as quantum chemistry and nanoscience penetrate industrial chemistry, more and more reactions may be run on catalysts, allowing for significant energy savings. Transport also generates a lot of CO2 emissions; it would thus be most interesting to be able to manufacture alternative cars (hybrid, electric) with good motoring performances, really low emissions, and at competitive costs. A huge step in this direction might result from the development of nanobatteries that store electricity without requiring chemical transformations; such development is under way in several laboratories, in particular at MIT and Cambridge University. Progress along the same lines is under way on nanophoto voltaic cells at Berkely and Penn State in particular. It is interesting to note that most of the technical innovations I have been led to mention (or could mention) stem from quantum mechanics, that most abstract theory (so do IT, the laser, all nanotechnologies). They are clearly apt to reduce mitigation costs.
However, available efficient technologies that are not largely disseminated would be almost useless, and the obstacles against timely dissemination are formidable. Of course we have a wealth of economic, legal and political instruments, from market design and taxes to technical and behavioural regulations. They will all have to be deployed together, in consistent ways; just having a “cap and trade programme that starts with a relatively low backstop price” will not do. It remains true that changing behaviours and institutions (in North’s sense) is in general rather more difficult than devising new techniques; this very difficulty is more an invitation to act rather than to delay action.
There is however one technique that is essential but might be difficult, or even impossible, to develop at the required scale: carbon sequestration. I understand that MIT, aware of this unsatisfactory situation, is planning a comprehensive assessment of the technical and economic prospects of carbon sequestration. The best conclusion I suspect one can hope for is: technically possible but difficult, and costly. But also urgent, at the pace new coal-fired power plants are built in China and India, and planned in the US and Germany (the list is certainly not exhaustive). Depending on the conclusions reached in the assessment, one might conclude that a new Manhattan Project is needed. Robert Oppenheimer and his collaborators have organized and realized an incredible effort to win the race for the atomic bomb against the Nazis. Would the rapprochement sound emphatic? I am convinced – not on the basis of a rigorous cost-benefit analysis, I confess – that the climate threat on my grandson – not only on his grandchildren - is no less acute than the Nazi threat was in the 1930s on my father, who paid dearly for the lack of preparation in his country, and in most of Europe.
Posted by: FT Forums | February 15th, 2007 at 4:26 pm | Report this commentThe author is professor of economics, Ecole polytechnique, and professor of sustainable development, Columbia University
By invitation - Terry Barker: I thought Martin’s article was spot on in identifying avoidance of risks as the critical basis of the economic analysis. Stern repeatedly emphasizes risks, but then when he comes to estimating the “costs” of climate change, he discounts them, most notably in the 0.1% risk of human extinction.
Such discounting is a problem in cost-benefit analysis involving risky outcomes, especially those in which there is a discernible risk of institutional catastrophe. In banking, for example, it would be foolish for a central bank to monetize the risk of a banking collapse and then act blindly to the risks of collapse: even if the risk is small, the damages from collapse are irreversible and asymmetric compared with the benefits of banking success. In the economics of climate change, risk analysis is the appropriate tool. Cost-benefit analysis is too limited and obscures the critical considerations of ethics and risks.
As an aside, I think that the argument put by Paul Seabright that the Stern Targets do not yield stabilisation is wrong. They do, within the likely ranges of the IPCC WG1 report, but in the far distant future and at temperatures more like 3 degrees above pre-industrial, rather than the 2 degrees of the EU target, or lower, to avoid even more risks.
On the costs of avoiding dangerous climate change, Paul is right to be cautious when he argues that “The real cost of a complete stabilisation is presumably substantially larger”. If we take his complete stabilisation to mean a target of 5Gt CO2e by 2100, this will yield an average long-term temperature rise probably below the 2 degree target, depending on the profile of the emissions 2007-2100 (the climate warming problem is one of stocks, not flows of GHGs). It is not a correct presumption that the real costs will be necessarily higher.
Higher costs are indeed the outcome of a few of the studies that have looked at more stringent targets, but they have serious flaws. For example, the draft US CCSP Report, about to be finalized, includes a projection, from the MIT EPPA model, that GDP costs by 2100 for the most stringent target in the report, will be some 16%GDP. On closer analysis the reason for this extreme result comes from the fact that even after nearly 100 years of escalating carbon prices (assumed to be known in advance by everyone, everywhere) reaching about $1600/tCO2e, road transportation is assumed not to be able to shift to the use of electric vehicles. In other words, the model is strictly limited in its representation of substitution possibilities.
The estimates of macroeconomic costs in the Stern Review come from an analysis of a large number of modelling studies, from which it is clear that the costs could just as well be negative. This seems obvious, if we consider the nature of technological progress and how it may be induced by climate policies.
It is also worth examining the assumptions of most of this modelling. Nordhaus’ model R&DICE is typical of the optimal growth models, assuming representative agents, full utilization of all resources (e.g. full employment throughout the developing world), without money as we know it, and most of the other assumptions of general equilibrium. The treatment of public finances and the global trade system is crude to the point of being absurd or non-existent. The many CGE models (as used in the US CCSP report) project 100 years into the future based on one year’s data - 1997 or 2001 depending on how up-to-date they are. This is not good applied economics, since there is a wealth of evidence on how the global system dynamically responds to energy price changes and shocks over the past 40 years.
In short, I think any conclusion that “the costs of mitigation are underestimated” ignores the modelling evidence, including the more recent work in multigas studies by the Energy Modelling Forum (EMF21) and the US CCSP studies. My view is that the costs of mitigation are uncertain, but less so than the costs of climate change; that there are large, near-term, and more reliable estimates of the environmental co-benefits of climate mitigation; and that the studies require careful assessment as to their assumptions and results before being treated as adequate.
Given the uncertainties, well-designed, pre-announced and anticipated policies for deep mitigation could in my view, lead to an increase in global GDP or to a decrease depending on timing and design. They are very likely to improve the health and welfare of millions of people living in grossly polluted cities in all parts of the world.
The author is Director, 4CMR, Cambridge Centre for Climate Change Mitigation Research, Department of Land Economy, University of Cambridge
Posted by: FT Forums | February 23rd, 2007 at 1:39 pm | Report this comment