05 Jan 2010

When Murphy Debates McKitrick, the Temperature Rises

All Posts No Comments

OK the post title was awful, I admit it. On my MasterResource post criticizing Ross McKitrick’s proposal to tie a carbon tax to observed temperature changes, McKitrick posted the following comment:

Hello Bob and others,

I’d like to respond to some of the points made here. The main concern I have in reading your posting is that you seem to have some ideal alternative in mind, in which perfect information for all time now and in the future is available to decision-makers, and there is no difficulty overcoming the obstacles to forming a coalition to implement it. So if you are going to criticize my proposal in comparison to an ideal alternative, I hope that in a follow-up post you will tell us what the ideal is. So far all the alternatives I have seen being seriously discussed are much worse, except for the strategy of doing nothing. However, my tax idea strikes me as a near approximation to doing nothing since the tax rate would start low and gradually taper off from there. That’s my expectation: others would expect the tax to start low and rise.

Your first objection is that we can’t say for sure what the present value of the marginal damages of a tonne of CO2 is. So? All we can ever do is work with estimates. Without some working estimate of the marginal damages you have no rationale for any emissions policy, whether price or cap-based. At least with my proposal, the direction of change over time has a higher chance of being correct than other instruments that do not build in learning mechanisms (i.e. everything else). Bear in mind that if you require perfect information then there is no possibility of setting any policy correctly.

In other words you are criticising my policy because in order to implement it requires picking an initial dollar value, and the choice is inherently uncertain. But that applies to any policy option. Given the uncertainty, my proposal lowers the expected costs of making a mistaken commitment, compared to any alternatives I know of.

To implement any policy, either an emissions tax, tradable permits (worse) or command regulations (worst), involves picking an explicit or implicit dollar value of the marginal damages. One of the worst aspects of regulations like the ethanol mandate and appliance efficiency rules is that they appear not to place a price on emissions control, when in fact they do. It just happens to be buried in the costs of regulation, and often comes out to thousands of times higher per tonne of actual abatement achieved than we would think is a reasonable emissions tax level.

You are also concerned about price volatility since the “Earth’s temperature” bounces around a lot. I have always suggested that the tax be smoothed out using a simple moving average, between 12 and 36 months, so that the change in the tax is primarily driven by a sustained trend not by weather variability. By applying a simple smoother it is not hard to ensure that the tax variations remain reasonable, especially in comparison to the cap and trade case where a vertical supply curve and a nearly-vertical demand curve interact. (That’s why Europe’s ETS system has such explosive price variability and always will, as will a cap and trade system implemented in any industrialized economy.) Having some variability is not a problem–we cope with energy price fluctuations all the time. In this case a certain amount of variabiltiy has a theoretical justification (marginal warming, if it is damaging at all, is more damaging at times when there has already been a spike in natural warming).

Your third objection is to suppose, what if the IPCC is correct, but natural forces mask warming for a while, then we delay the correct policy response. Once again you are imagining an alternative world in which we have perfect information. If the IPCC is correct, and we know them to be correct, then people will anticipate the future tax increase and plan accordingly. If the IPCC is correct but we do not know them to be correct (or believe them to be incorrect), then no one is in a position to make the right choice. At least under my proposal, policy will eventually update as new information becomes available. Every other policy proposal is impervious to new information. We have a choice between policies that assimilate new information as it becomes available, versus policies that never assimilated new information. We do not have the option of policies that assimilate information before it becomes available.

Note that in the above, I edited out a lot of his comment, since he was addressing other people and not just me. I replied:

Ross (if I may),

Obviously it’s hard to be precise in a blog post, but I think you are misunderstanding my problems. When you ask me for a rival policy that responds to new information and is superior to yours, I can glibly reply: Sure, let’s have a carbon tax calibrated to the best-estimate of the social cost of carbon. And then as we get new information (including new temperature readings), we update the SCC using Bayes Law and adjust the tax accordingly. That is literally the theoretically best policy, given our uncertainty.

Now why don’t we do that? It’s because we all can’t *agree* on what the SCC actually is. There are people who think it’s really high, like Joe Romm, and that we should be taxing the heck out of emissions right now. Then there are people like Richard Lindzen, who thinks the IPCC estimate of climate sensitivity is too high, and (I presume) the “optimal carbon tax” should be a lot lower than what most other experts would say.

So what I was doing in my post wasn’t so much saying, “Whoa, McKitrick’s plan backfires because the IPCC might be right!” Rather, what I was saying is that I don’t believe there is actually a wide overlap where both sides in the debate could agree to a McKitrick tax.

For example, to get Rob Bradley to sign off on the tax, the initial rate has to be low. So in principle, if tropospheric temperatures stayed flat for the next two years (or even fell), then we would have a very low carbon tax for 2 years in a row.

Knowing that possibility, someone like Joe Romm would insist that this low initial tax rate would have to go up by a factor of 20 or something at the first blip in global temperature. If we didn’t concede that to him, then Romm would never sign on to the policy, because there would be (in his mind) an unacceptably high probability of catastrophe.

Do you see where I’m going with this? Because of the natural variability in temperatures, I don’t see how you actually could get honest people who have wildly different estimates of the SCC to agree to a McKitrick-style tax. Thus, I don’t think you are necessarily “smoking the other side out” if they balk at your proposal, since you’re asking *them* to sacrifice their view on the front end.

Maybe that’s a good way to frame my point: We could just as well propose a Murphy temperature tax, in which we start out with a $100 / ton tax on carbon. But then for every 1% the tropospheric temperature falls behind the IPCC suite of models’ forecasts (given the observed emissions and updating the forecasts all along), we knock down the Murphy tax accordingly. So we can “smoke out” the skeptics and see if they actually think the IPCC models are overpredicting temperature rises, or if (on the contrary) the people criticizing Waxman-Markey etc. are really just philosophically opposed to government intervention.

Does that help clarify my point?

If I were famous, blogs across the geeconosphere would go nuts: “Murphy proposes $100 / ton carbon tax!!”

Comments are closed.