A sample text widget

Etiam pulvinar consectetur dolor sed malesuada. Ut convallis euismod dolor nec pretium. Nunc ut tristique massa.

Nam sodales mi vitae dolor ullamcorper et vulputate enim accumsan. Morbi orci magna, tincidunt vitae molestie nec, molestie at mi. Nulla nulla lorem, suscipit in posuere in, interdum non magna.

Rational alarmism on climate change

From an insurance viewpoint we should be much more concerned about climate change than most of us are. The focus of IPCC reports is on a likely global average temperature increase forecast of 3oC by 2100 in a range of perhaps 2oC to 4.5oC. The mid-point of this range would lead to dramatic changes in the way people live and work on planet earth. The high point of the range would lead to dramatic changes in our lives – the loss of key agricultural regions, water supply problems and so on.   

But what of the prospects of much higher climate changes? In a recent paper Weitzman (2009) examines the prospects for catastrophic climate change.  His estimates are based on 22 stochastic simulation models provided in IPCC (2007, p 721-722, 798-799).  Weitzman argues that if only gradually ramped up remedies are applied that in two centuries the probability distributions of temperature outcomes have decidedly ‘fat’- tails:

·        With probability 0.05 the increase in mean global surface temperatures will be greater than 10oC.

·        With probability 0.01 the increase in mean global surface temperatures will be greater than 20oC.

Such temperatures have not existed on earth for hundreds of millions of years and the rate of such increases has possibly not been experienced for billions of years.  With a 1-5% probability (not a negligible range of probabilities) there is the prospect of a worldwide catastrophe.

In fact when we insure our health or our homes we are much more concerned with extreme events than with moderate misfortunes that we can (with some inconvenience) handle ourselves with our own resources.  I am fearful of the remote chance of my house burning down more than having my TV set stolen.  I am much more fearful of contracting a serious though uncommon chronic disease than the common cold.

It seems to me exactly the same with climate change. Extreme events that occur with low – though not negligible probability – should be the focus of attention rather than average events.  In doing a cost benefit analysis (CBA) of the case for social insurance to offset the effects of such ‘fat’- tailed probability distributions Weitzman writes that ‘CBA is likely to be dominated by considerations and concepts related more to catastrophic insurance than to the consumption smoothing consequences of long term discounting – even at empirically plausible interest rates’.

Unless the consequences of rapid climate change cannot be offset by policy actions now this suggests urgency – a rational alarmism – with respect to climate change policy. Being alarmist in this sense is not irrational but accords with every day notions of insuring at relatively low cost against the prospects of extreme events.


IPCC, Climate Change 2007: The Physical Science Basis: Contribution of Working Group 1 to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge, UK, 2007.

M. Weitzman,  On Modeling and Interpreting the Economics of Catastrophic Climate Change, The Review of Economics and Statistics,  91, 2009, 1-19.

8 comments to Rational alarmism on climate change

  • hc

    Bek, That is delusional given the weight of scientific evidence. This is an issue that has been thoroughly dealt with.

  • John Mashey

    Have you read any good books on this by real climate scientists?
    If not, and you’re willing to spend a handful of hours, I can suggest a few written for a general audience by top-notch scientists.

  • Two points:

    1. Two centuries is a very long time–think about how much things have changed since 1809. Taking action now to deal with consequences in two centuries is almost certainly a mistake, since we don’t know nearly enough now about what will happen in the meanwhile to either estimate the probabilities that far out or our options for dealing with them.

    2. While considering low probability catastrophes, shouldn’t Weizman have considered one in the opposite direction? We are in an interglacial. Typical interglacials last ten thousand years or so. Even ignoring the fact that this one has already been running longer than that, that makes the chance that it will end in the next two centuries about 2%–comparable to the probabilities he looked at.

    And we know the consequences. A typical glaciation puts a mile of ice over the present locations of London and Chicago. Weizman is talking about temperatures that haven’t existed on earth for hundreds of millions of years–including times when the CO2 concentration was a great deal higher than it now is. I’m talking about conditions that have existed for much of the past million years.

    Which is more likely? If our concern is with the tails of the distribution, perhaps we should view warming as a feature, not a bug.

    For a much more extensive discussion of the first point, see my recent Future Imperfect.

  • hc


    (i) 200 years is a ‘long time’ only from narrow perspectives. Depending on lags it may be foolish to wait to get better information about climate. A substantial burst of CO2 into the atmosphere as a result of burning fossil fuel has an atmospheric lifetime of perhaps tens of thousands of years.. Shorter-lived GGEs such as methane will disappear quicker but might be continuously released if the permafrost melts. With irreversibility of certain effects and uncertainty it does not always pay to wait indefinitely. The ‘optimal stopping’/’stochastic capital theory’ literature shows this.

    (ii) It could be that over 200 years favourable events such as ‘cooling’ come into play in which case there is the prospect of a favourable offset. But acting to deal with climate change will not impose disasterous prospects if such cooling comes into play whereas not acting will produce disasterous prospects if it doesn’t and the types of catastrophic scenarios envisaged by Weitzman do.

  • John Mashey

    David + hc:

    1) Ice-ages are normally driven by

    a) Milankovitch cycles (orbital, precession, wobble), which are very well predictable. In particular the pattern of summer solar insolation @ 65degN matters,and circularity of Earth’s orbit matters.

    b) “Trigger points” driven by Greenhouse Gases.

    2) Good analyses can be found in many places, of which my two favorites are written for general audiences by experts, and cost ~US$15-20 each. [These were the two I’d have suggested for Bek above.] The Amazon links have reviews:

    a) William Ruddiman, Plows, Plagues, and Petroleum, 2005, Chapters 3-5.

    and a bit newer one, with a slightly different viewpoint in:

    b)David Archer, The Long Thaw, 2008, Chapter 12 “Orbits, CO2, and the Next Ice Age”.

    3) Ice ages end fairly quickly (5-10K years), but the descent back takes much longer in any case. Short-term fluctuations have happened (Heinrich/Dansgaard-Oeschger Events), but those aren’t real ice ages. From the time when gaclers starting growing back from Baffin Island, to when you see them coming from Toronto skyscrapers is tens of thousands of years. Real ice ages don’t sneak up on us.

    We *might* have been heading for reglaciation in ~3000 years (near miss, now less near). Archer writes (p.156):

    “If mankind eventually burns 2000 Gton C (this is about the business-as usual forecast for the coming century), then it looks as though climate will avoid glaciation in 50 millennia as well, waiting for the next period of cool summers 130 millennia from now. If the entire coal reserves were used (that is 5000 Gton C), then glaciation could be delayed for some 500 millenia, half a million years.”

    4) But in any case, it doesn’t matter what Milankovitch cycles might have done. If there’s a technical civilization on the planet, it is not that hard to fend off an ice-page via:

    a) albedo adjustment (soot on ice is not bad, low-tech) and especially

    b) human-manufactured, very-long-lived (1000-50,000 years) GHGs, most likely Perfluorinated compounds (PFCs) like sulfur hexafluoride (SF6), PFC-14 (CF4), or PFC-116 (C2F6). All these are (per molecule) much more powerful GHGs than CO2, say ~20,000 stronger over a 500-year timeframe, which means you don’t really need much of it, i.e., far less than the CO2 or SO2 we emit.

    See IPCC WF I Chapter 2, Table 2.14 on p.212, and comments on top of p.145.

    5) Bottom line on new ice-age: you can worry about nuclear war, or the next killer asteroid, or a nearby supernova, but next real ice-age? No worries.

  • J

    Yes, David Friedman, 200 years is a long time but that is how long it took to get us into this mess, doesnt it make sense to use the next 200 to get us out?

  • Richard

    The shot at Steve Fielding was uncalled for. Take a look at the upper atmosphere the next time you fly. Pollution is now visible much further a field from cities than ever before. Take a look at sea life which is dying directly because of human pollution. Take a look at western Victoria which once was wheat belts and now is desert. Finally take a look at yourself. You may have grandchildren: do you want them to live in a polluted planet or a sustainable planet? Sure we live in a changing biosphere, but we must take responsibility to stop destroying and start managing. Is it the end of the world as we know it? I suspect not, but I’m not taking the risk to ignore the POTENTIAL or my role. As HC mentioned, its all about taking the risk.