A correspondent points me at the work of Nordhaus and Boyer, attempting to estimate the externality from CO2 production and the appropriate response. I have not read the printed version, but there is a webbed version available. A number of things strike me:
1. Their conclusion is that
"current approaches, such as the Kyoto Protocol, are highly inefficient, with abatement costs approximately ten times their benefits in reduced damages."
2. In order to get even that result, they depend on including costs from a very uncertain estimate of the risk that global warming will result in low probability catastrophes of one sort or another:
"this approach is taken because of the finding of the first-generation studies that the impacts on market sectors are likely to be relatively limited."
Or in other words, without including the costs from such catastrophic risks, global warming doesn't seem to be a serious problem.
3. They estimated the cost from low probability catastrophes by asking a lot of experts how likely they thought it was that there would be a catastrophe, assuming a given rise in global temperature, that would reduce world GNP by at least 25%.
Even if one takes seriously the output of that sort of procedure, there is a striking asymmetry in their approach. They do not appear to have asked any experts what the chance is that preventing global warming would cause a catastrophe—or, to put it differently, that global warming will prevent one. Yet, as I keep pointing out, earth's climate was not designed for us, hence there is no a priori reason to assume that large negative results due to a few degrees of warming are more likely than large positive ones.
This is a striking illustration of my general critique of the "add up the externalities" approach to policy issues. Possible changes can have both positive and negative effects. If you want to conclude that something should be prevented, you focus on the negative effects, ignore or minimize the positive ones, and claim to have an objective argument to support your conclusion. That is precisely what Nordhaus and Boyer have done. The only surprising part is that, even after doing that, they still got only a weak version of the (presumably) desired conclusion.
This leaves me wondering what the pro-global warming view of Nordhaus and Boyer is. Their conclusion weakens the case for something like the Kyoto protocol--and that conclusion hinges on a procedure that cannot, I think, be defended. Correct that error and their work appears to provide no support for any substantial effort to prevent global warming.
P.S. After writing this, it occurred to me that there is one respect in which I am being too hard on Nordhaus and Boyer. Suppose we limit "catastrophes" to changes that are not only unlikely and very large, but also very fast. Then there really is some asymmetry to the situation. Humans are currently optimized against the current environment, making any change presumptively negative. I have argued in the past that that is not a strong argument with regard to changes that occur slowly—for example three degrees of warming in a century—since humans will in any case be changing what they do in many ways over so long a period. But that argument would not apply to a change that occurred over only a few years.
If, for instance, climate change makes Europe unbearably cold but adds an equal amount of acceptable land elsewhere, say in northern Canada, and does it in only a couple of years, there will be a very large net cost.
So they would have a legitimate case if they limited their category of catastrophes to rapid ones—but, so far as I can tell from what they wrote, they did not.