It is time to stop quivering in our boots in pointless fear of the future and just roll up our sleeves and build it.
- Ray Pierrehumbert

Thursday, May 26, 2011

McKitrick's Plan

I appreciate people sharing the broad outlines of how they think we ought to handle the carbon situation. I am not entirely serious with the proposal I have outlined; there is much that I like about it and I hope it can be salvaged, but so far nobody has pointed out what I think is its fatal flaw. We'll get back to that shortly.

Meanwhile, while we're on the subject, a couple of folks have raised McKitrick's plan, which bases a carbon tax on the instantaneous global mean surface temperature (or, according to one correspondent, the equatorial mean).

The equatorial mean, of course, is plainly silly: why pick the least temperature-sensitive spot to measure the anthropogenic damage, unless you were looking for a way to pretend it away.

There is some appeal to the idea, though; the idea is to find a measure of global change that is independent of the scientific theory, such that people who don't have any faith in the science could agree to it, and to use that measure to calibrate the response.

In practice McKitrick's idea seems either ignorant or disingenuous. I cannot take anybody who proposes it seriously, because they still remain firmly in the class of people who "don't get it".

Let's leave aside the usual confusion about "global warming", which is a symptom of anthropogenic climate change, not the disease. We can reduce global warming in a literal sense to zero easily enough with additional aerosol releases, but this will not avoid massive climate change.

Even if global temperature were the a complete measure of anthropogenic climate change, the problem is that it's a delayed measure. The ocean and sea ice take a while to respond; the land ice even longer; and the clathrates (we hope) still longer than that. (*) What this means is that the temperature we see now is the temperature we bought in the past; some of the response is delayed, and some of it is greatly delayed. This is part of what I would call a basic policy-level understanding of the science. If you miss that point, you don't know enough about what is going on in the climate system to venture a serious policy.

I like the idea that the policy should adjust to the evidence. But the idea that the instantaneous temperature is a measure of the sensitivity is a mark of denialist-influenced confusion. Unfortunately there simply is not a simple metric of how deeply in trouble we are. I think the basic idea is not unreasonable on its face, but the measure we use needs to account for the delays in the system.

We could apply something like this to the trillionth ton constraint. If the system is less sensitive than we think, we can loosen the constraint. The trouble is that is the system is more sensitive than we think, there is very little time to tighten the constraint. There is realistically no more time for dawdling, since even the trillion ton limit carries plenty of climate risk on present evidence. We should shoot for that now; shooting for two trillion and realizing the right goal is one trillion is no good if you find yourself already committed to a trillion and a half.

(*) These can be regarded as exacerbating feedbacks from the point of view of the atmosphere system, or slow modes from the point of view of the whole coupled climate system.


Grypo said...

The McKitrick Tax was proposed as a way to call-our-bluff, hilariously. And they all loved it. So, yes, they "don't get it".

Andrew Dessler said...

If you want to talk an approach like this, there are more sensible approaches: Hsu, Shi-Ling, A Prediction Market for Climate Outcomes (February 26, 2011). Available at SSRN:

I think this is quite an interesting idea, and one that could actually work (which means that the deniers/delayers won't ever agree to it)

gravityloss said...

Ok, it's hard to convince people to stop speeding. So how about if you apply brakes after you hit someone who was crossing the street?

Tom Fiddaman said...


An engineering analogy would be controlling the coolant flow in a nuclear power plant by measuring the amount of melted fuel at the bottom of the pressure vessel.

Not only is there a substantial delay, but the noisy measurements would impose needless volatility on the control.

Michael Tobis said...

Andrew, based on the abstract, Hsu's prediction market only looks three years into the future. So it would seem it doesn't escape the short horizon of standard practice in the marketplace at all. Did I miss something?

rustneversleeps said...

@ mt, Hsu actually speculates about the logistics of how (and how many) permits to issue for vintage years up to 40 years in the future... see. (An earlier, incomplete draft... so take with that caveat...)

It's a useful sketch of an idea.

Andrew Dessler said...

it's been a while since I read his paper, but I think it operates on much longer time scales. it's actually quite a sophisticated proposal, so I'd recommend reading the entire thing (you can skip the discussion of climate science)

Rattus Norvegicus said...

...McKitrick's idea seems ignorant and disingenuous.

Fixed the typo...

The first time I saw this proposal, I picked up the trap(s) embedded in it. It really is the next worst thing to doing nothing.

Lou Grinzo said...

McKitrick's idea reminds me of the adage, "Almost every question has at least one answer that's simple, intuitively appealing, and wrong."

There is one way in which his proposal can serve a useful purpose, though. We can point to it and explain the awful timing situation we're in regarding climate change -- all that heating "in the pipeline", the additional melt from the poles and continental glaciers, the latencies involved in implementing any scheme large enough to make a meaningful difference, etc. In my experience, this is an aspect of CC that even a lot of very committed non-scientists don't understand. They think that if we reduced our CO2 emissions by 50% that the extra atmospheric CO2 would decline by the same proportion in a year or two. In fact, I think getting the message "timing is everything" across to lay people is one of the fundamental issues of climate communication.

EliRabett said...

nah, just use ocean pH

William T said...

The other problem is that the effect (warming) is due to the stock of CO2 in the atmosphere, not the flow. So it's not just this year's emmissions that should be charged, but emissions for previous years (decades) as well.

Perhaps if everyone has a carbon account accumulating their emissions over time, then each year they're charged a "global warming charge" on every ton they've accumulated to date, multiplied by the current temperature. anomaly...

The Evil Reductionist said...

Maybe bilateral deals would work better than a global deal. As more bilateral deals came into being, they would start to intermesh, and the effect would eventually be a global deal.

However, I think that most individuals do not get the idea that cutting carbon emissions is going to hurt. I think that politicians, and this includes Greens, do themselves no favours by trying to sell the message that coal miners will be able to move smoothly and easily into jobs in the renewables sector (for example) or that compensation will somehow be sufficient to cover for price spikes and other fluctuations when we run into energy transition problems.

We need to consistently tell people that the treatment is going to hurt a lot but that if we do not do it it will hurt a hell of a lot more. We need to make people proud to be the ones who are sacrificing so that their children and grandchildren do not have to. We need heros to exult and examples to follow. We need to decry waste and excess, as was done in WWII.

As an additional point, this means that the sellers of our message need to be as consistent on this issue as possible - no use of offsets, for example, which simply sends the message that if you are rich enough you do not have to change your lifestyle at all. This is something that jars a lot with people. Even if the maths of carbon offsets adds up, they whole thing has a bad feel about it.

adelady said...

Wot Eli sed.

Martin Vermeer said...

I would say that the fatal problem with this idea is that, overwhelmingly, the costs of climate change are incurred in the future, not the past or present. Any sensible policy will thus be based on those future events, or be plain wrong.

Yes, the future is hard to predict; that's a very fundamental property of the future. Models don't work very well. Predicting the present would be so much easier ;-)

The deceit in the McKitrick idea is that he is predicting the future too. He uses the present as a "proxy" for the future, but never clearly discusses what the proper coefficient should be. He has a "model" too, but never calls it that. He makes a "prediction" too -- that's where that coefficient comes from -- but never calls it that.

Keep your eye on that coefficient.

Layzej said...

Ok, McKitrick's idea is simple, intuitively appealing, and certainly not perfect, but is it really entirely wrong? I think it can be salvaged.

Your concerns are:
1) Equatorial mean temperature rise is not a proxy for warming due to CO2.

This was already anticipated and addressed in the original post. I suggested using arctic mean which has the advantage of being very sensitive to CO2.

2) Temperature lags CO2.

This was also anticipated but perhaps not articulated clearly enough in the initial post. You can deal with the delay by pricing carbon well ahead of the temperature increase. That is, price the temperature today based on how far we have already moved from pre-industrial carbon levels to Hanson's point of no return carbon levels. Anticipate what the temperature will be when we hit the breaking point (considering business as usual) and make sure that your $/C prices carbon out of the market well before temperatures hit that point.

3) Temperature is not the sum total of the problem, it is merely a proxy.

Sure, but it is a proxy, and it is one that everybody is familiar with in the context of AGW.

The plan has the following advantages:
1) It was proposed by a skeptic.
2) It is simple, and intuitively appealing
3) The alternative is what we are doing now (nuttin'). The vocal minority has rejected science and will not be swayed by it.

Your reaction to this as a framework seems knee-jerk. Can you take another crack at explaining your apprehension to this ignorant or disingenuous person who just "don't get it" ;)

Steve Bloom said...

FYI, Layzej, Hansen thinks we are .2C from serious trouble, which implies that the need for a sharply increasing tax has already been established. Indeed it is probably too late to avoid some pretty nasty effects. The problem is that we're talking about this now rather than having already implemented it 20+ years ago.

Grypo said...


Your number 2) completely changes the McKitrick idea. 3) Temperature is not a good proxy to use, unless put together with other indicators. It's simplicity is irrelevant. Energy melts ice, energy invigorates storms, energy expands oceans, changes Hadley, etc etc. If we use indicators to set a price on carbon, it needs to be closer to the problem, either by grouping all the possible symptoms (could use known fingerprints of enhanced GH) or by using the energy imbalance directly. This would need less error on human aerosol contribution, tho. Any metric will have issues that will make for regulatory hell for countries, let alone globally. I do like the idea, but without proper measurement, it may be unworkable.

Layzej said...

So if we have already seen about 0.8C and we are cooked at 1C, then the current McKitrik tax would set the price of carbon 80% of the way to completely unaffordable. If the temp goes down (as skeptics assure us it will) then so does the cost of carbon.

Grypo said...

That improves the model a great deal (although not all .8C is attributable to man), but it changes the model in way as to lose all support we are hoping to gain from those skeptics supporting it. So I would choose another metric and tweak it to satisfy a less ideological crowd.

Layzej said...

Support will come from the fact that cost is tied to impact rather than input.

You need to use temperature because it is the metric that most people associate with AGW. It is easy to communicate and people will get it.

Seriously, how is this not the least unlikely policy?

Michael Tobis said...

Layzej, let's leave aside my somewhat idiosyncratic claim that the obsession with temperature is part of the problem.

Here's the problem simply in terms of temperature: we know that there are lags in the system. The temperature we see at an instant is the more or less instantaneous response of the atmosphere moderated by the delayed and complex response of the upper ocean and the sea ice and the still more delayed and poorly understood response of the ice sheets, the deep ocean, the soil, and the sea floor. Eventually, the equilibrium reached for a given atmospheric composition will propagate to all these subsystems until the energy flows equilibrate. So what we are seeing is, first of all, only a fraction of the response for the whole system.

To make matters even harder to think about, it's probably the case that the slowest components weren't even in equilibrium in the first place, since the last three million years have been a time of enormous climate shifts. Accordingly, it's not even fair to talk about equilibration on the longest time scales, but only about how we are changing the adjustment process.

But if you want to completely ignore that complication, you can simply focus on the intermediate time scale, that of the sea ice and the upper ocean. Here we have responses that are delayed by decades from the input. What's more, especially insofar as the ocean is concerned, the response is likely to contain an oscillatory component, such that instantaneous warming is masked by decadal swings.

On top of that delay, consider that on the policy side we are talking about an enormous amount of infrastructure which cannot be changed over night. Again we are looking at a delay on the order of decades. (For the purposes of this argument we can leave aside the delay in coming to terms on a policy, since a specific policy is being proposed.)

On top of the delay in the physics and the delay in the economics, and the oscillatory components of the physics, we have a fourth issue, and an especially crucial one. The system does not respond to carbon emissions, it responds to concentrations, which essentially amount to cumulative emissions. The forcing is a ratchet. We can only go forward more slowly; going in reverse is physiclally conceivable, but is so expensive in energy terms or in dollar terms that it completely wipes out any prior benefits of going forward!

So if we have a period when it is warming less than anticipated, even if we accept that "warming" is the problem and not just a symptom, we cannot be sure that the slower warming (or even cooling) is not just an oscillation on top of the underlying trend. More to the point, if we respond to this by loosening the screws in emissions, we will not see the results of our loosening for a long time. Partly, because the resulting response will be delayed, but even more, because the resulting forcing will take a long time to appear above the cumulative forcing of the past.

The net result is that the idea is conceptually sound only on very long time scales. If we look at the temperature response over a century, we can calibrate our behavior over the next century. But there are a dozen streams of evidence to show that this is the wrong time scale.

We are in trouble now; if the apparently abject failure of Copenhagen really sets us back a decade or two as appears to be the case, we are already in crisis unless the dozen streams of evidence are somehow leading us drastically wrong.

In short, the instantaneous temperature is less than the equilibrium temperature and the scaling is underconstrained and noisy on economic time scales.

Layzej said...

Yes the instantaneous temperature is less than the equilibrium temperature. Why does this matter? Both are going up. Why not peg carbon to the one that we can easily measure?

And sure the temperature is going to fluctuate. Use a 10 year mean to smooth out any lumps or bumps. A slower warming trend would not loosen the screws, it would just tighten them slower. A prolonged drop in temperatures on a 10 year mean doesn't seem conceivable at this point, but even so: If we are at 0.6C now and temperature drops to 0.5C then carbon would be at 50% of completely unaffordable instead of 60%.

I agree with your concerns regarding infrastructure. This is something that needs to be dealt with quickly regardless of the policy. You would need to set the initial cost of carbon appropriately so that the system has time to adjust while also accurately reflecting how deep in the hole we already are.

Michael Tobis said...

Short version: because the year over year rate of change of annual temperature contains very little information about the foreseeable trajectory.

This is in fact the sleight of mind that the denialists are forever doing. Embodying it in policy is to write perpetual confusion into law.

Steve Bloom said...

I, and I suspect most everyone here, would be perfectly happyto see a carbon tax scaled per Hansen. The difficulty is that McK. et al. would find that entirely unacceptable. The wheel comes full circle.

Martin Vermeer said...

> So if [...] we are cooked at 1C


you just admitted to basing yourself on a model. How would you know that we are cooked at 1C (or any other numerical value)? How would you convey that knowledge, and establish its credibility, to somebody questioning it? That's the problem.

Layzej said...


Every policy that could deal with this is going to infringe on individual liberty. There is going to be push back from the "I can do what I wanna" crowd no matter what. Regardless of the policy we will need to base the boundaries on the best conservative science. Michael's 65 Tons/person solution is also based on science. I see this as a selling point not an obstacle.

The trick is to find a policy that appeals to the moderates. This policy will appeal because it is tied to the impacts.

Regarding Michael's point that yearly fluctuations are independent of the trajectory:

The ten year mean of the temperature doesn't really fluctuate all that much. It has followed a predictable path straight up for 40 years. (

I don't buy that this is going to change in the future. We have 0.6C still to come just from the CO2 that we have already spent. According to Hanson there is no natural cycle or volcanic event that could put a dent in our current trajectory. Michael thinks that the trajectory of the 10 year mean temperatures may change. I can't presume that I know more about the science than him but it seems to contradict Hanson.

Martin Vermeer said...

> This policy will appeal because it is tied to the impacts.

Eh, not in a meaningful way. Remember what I said about that coefficient? The moment you establish your 'tie' with realistic numbers climate science wise, your proposal is politically dead.

James Annan said...

Late as usual, but that "prediction market" sounds rather similar to what I was talking about back in about 2005 (JAMSTEC site borked hence using wayback).