My general point is that global warming responses follows the same risk-averse model as normal insurance. If we were risk-neutral, we could go with the median forecast. Because we’re not, so the tails on the forecast matter.

I agree with Kimble that we not willing to spend everything to protect ourselves from miniscule threats (we don’t seem to be doing much about the threat of a large asteroid crashing into us for instance).

So really I see the grounds for disagreement about global warming response falling into categories of:
uncertainty about bias of global warming forecasts;
uncertainty about variance of global warming forecasts;
how risk-averse we are;
how costly CO2 abatement will be (and discount rates are another mine-field);
the costs and benefits of delaying (which depends in part on how quickly we will eliminate other forecast errors).

Combine any bunch of plausible parameters for all those values and you could quite reasonably support most positions without being ideological in any way (even, I suspect, if we start by stipulating that global warming forecasts are unbiased).

And then there are the realpolitik aspects – I suspect Lomberg’s argument is better if you think that there is a lump sum of global funds available for worthy causes, and that spending money on global warming detracts from spending money on third-world development, for example.

To CPW: I agree that carbon tax rates (in an alternative world where they were used) might ideally be adjusted over time. But I would expect that changes to these rates should reflect changes in our knowledge about the harm from emissions more than knowledge about external costs. The textbook level to set the tax is at the marginal level of *harm*.

To Kimble: I don’t doubt that there are some people crazy enough to say that we should do “anything and everything”. But that kind of extreme policy response is not on the cards. The best policy *should* reflect the range of possible outcomes – including very unlikely but catastrophic ones. Of course it is very difficult to reflect them properly in policy, because our estimates of the likelihood of these events are very unreliable. If we were going to conduct such an exercise, we would have to accept that there would be a lot of noise in our best guess. But we don’t get our best guess for *expected* harm by truncating the distribution of possible effects.

I have never given credence to the argument that X may be such a big problem that we must do anything and everything to address the possible problem, not matter how slim the chances of it occurring. People who say that have often given up on reason.

The critics of Lomberg have a point when they say that worst case global warming scenarios are so disastrous that their cost dwarfs any realistic estimate of the cost of reducing emissions, hence we need to immediately establish a cap on emissions.

But if we just go with a median global warming assumption, it’s quite possible that the optimal response from a cost-benefit perspective will allow some warming. There’s no way of telling until we have some idea of how costly carbon reduction will be, which depends mainly on the rate of technological advance of non-carbon energy sources. In this scenario, it would seem optimal to raise the level of carbon taxation over time as we observe costly reducing carbon emissions really is. (Caveat: that paper that says that slowly raising carbon taxes creates a perverse incentive to produce more now).

Waiting for more certainty on cost projections obviously has value, but delaying means the eventual response will be more costly. Have we crossed the threshold yet? We really need predictions about how rapidly climate modeling predictions will reduce their error bounds.

I think you will find that once we start hitting those points in time that they will be reset a little further out. I am pretty sure that most of those predictions, “if we dont do anything by March 2012, we are doomed no matter what!”, are just the predicters effort to give the rest of us a hurry up.

Waste of time listening to them in any case.