Policy decisions should be based on a judgment concerning the maximum tolerable increase in temperature and/or carbon dioxide levels given the state of scientific understanding. The appropriate role for economists would then be to determine the least-cost global strategy to achieve that target. While this remains a demanding and complex problem, it is far more tractable and epistemically defensible than the cost-benefit comparisons attempted by most IAMs.I agree wholeheartedly.
Indeed, that is why we have a 2 C target. It’s not just that the economics of the problem is hopelessly intractable, though it is.
We have delayed long enough that economics does not meaningfully enter into target-setting. See this article by me and a more detailed analysis by Dana Nuccitelli at Skeptical Science.
The strategy may have been an interesting economic question once, but it is not so any longer. We have delayed so much that the optimum rate of decarbonization is simply "as quickly as is feasible", that is, we need to achieve the absolute minimum cumulative net CO2 emissions that we can without the decarbonization kicking off destabilizing damage to society in itself.
So the question we can put to economists - how fast can we put the brakes on without spinning out of control economically or politically - may actually fall within the range of the sorts of analysis that, at least purportedly, they can do. Since it’s on a short time scale and by assumption avoids tipping points, maybe their methodologies will help.
Of course, as with any policy/expertise interface, distinguishing the real experts from the charlatans is also a crucial issue.
I am not entirely happy with David's opening for the article, though, and it raises a cluster of ideas that I have been meaning to write about for a while. I hope to explain shortly.
4 comments:
Your link to Roberts needs fixing.
"Do IAMs adequately account for uncertainty? Do they clearly communicate uncertainty to policymakers?" seems like a silly thing to say. Compare: "Do AOGCMs adequately account for uncertainty? Do they clearly communicate uncertainty to policymakers?"
The answer to both questions is: it depends entirely on how the model results are presented.
A bit pedantic, if strictly valid of a criticism I suppose.
The problem is that it is easier to construct worthless models than worthwhile ones, and that in some fields it's easier to get away with than in others. Presumably David has tried to polish the edge off a rather blunt statement and fuzzed it up a bit too much.
Look under the hoods of these things. They are sloppy little spreadsheet exercises. Literally.
I asked Chris Hope what language his model was written in, and he said Excel. He initially seemed to be willing to be open about it, but he said he wouldn't send it until I had installed a proprietary Excel package costing some absurd amount of money for a blog project.
They aren't even prognostic, so there really isn't an algorthm at all. It's just a messy ad hoc formula in a spreadsheet with "underlying GDP growth rate" as one of the magic numbers. Yes, it's a constant, and it applies over the entire period of the projection, which it has to, because it all gets done in one step.
Phrase it however you want. For economists to be telling climatologists how to run their business and get away with it may be among the most stunning feats of hubris in human history.
OK, so its not what *you* want to talk about, but its a large part of what DR is talking about. I stopped when I got there and wrote my comment here; now I'm reading on and I'm finding yet more of the same. Down as far as "Most of the uncertainty in IAMs is parametric uncertainty" I'm reading criticisms that could be put at AOGCMs too, either directly or analogously. Now I've got to "many of their key variables are, to put it technically, pulled out of modelers' asses" which is the kind of thing you write for an audience that you expect to agree with you; not a skeptical one that you're hoping to persuade.
"Think about how insane it is to try to predict what's going to happen in 2100". Errm, indeed? Like, for example, CO2 emissions? Lets not try to predict them to 2100. Or to 2050? When shall we stop and declare that kind of stuff pulling things out of asses? In which case, how do we know there is a problem?
"certain inputs (e.g., the discount rate) are arbitrary" - no; that's bollox. What he means it, real-world discount rates produce results he doesn't like.
And the "fat tail" of CS is something JA has been arguing against for years.
However, while all of this is errors in DR's article, its all fairly irrelevant to (what I take to be) your central point, which is that damage from GW is well enough known -scientifically - that we should use economics not for a cost-benefit analysis of warming-vs-CO2, but merely to achieve a least-cost strategy to achieve, say, less than +2 oC.
I'm dubious that's a sensible argument, but at least it isn't obviously wrong.
Aside: I don't understand your "economists to be telling climatologists how to run their business". I don't see them doing that.
"I don't see them doing that.": I suppose it depends whether you consider Tol to be an economist.
"The Fat Tail": stay tuned. The sensitivity PDF is the wrong place to look for the fat tail. We haven't got rid of it.
"Discount rates": why bring this up? Surely you've heard the counter-arguments by now. You could write my response for me.
"Writing for an audience you expect to agree with you": Nothing wrong with that. This is exposition, not debate. Debate is harder, obviously.
" I'm reading criticisms that could be put at AOGCMs too, either directly or analogously." Yes, but as a matter of degree these problems are **immensely** worse for the economic models, and the empirical constraints are **enormously** less available and in any case mostly avoided, unlike in climate work. If they have anything like weather forecasting to keep them honest (arguably they do) it hasn't worked.
IAMs aren't even prognostic in a mathematical sense. They go from parameters to predictions in one big formulaic leap.
These things are astonishingly bad. Go look at them before you start complaining about David's distasteful metaphor.
Post a Comment