Carbooooooon!
(Supposed to be like “Khaaaaaaan!”)
My latest Mises CA post takes on the typical claim that “uncertainty” bolsters the case for government restrictions on carbon emissions. Plus, I work in a Captain Kirk reference. Some excerpts:
So what these researchers have formally shown, is that if you assume this shape of a damage function, but you are unsure of exactly where the curve is, then the mathematical expectation of “damage from a given amount of warming” is higher, the more uncertain we are about the exact position of the curve, other things equal. Or, supposing we know exactly what the damage function is, but we aren’t quite sure how much an additional ton of CO2 emissions will physically contribute to global warming, then the mathematical expectation of the damage (in dollars) from emitting that ton will be higher, if we have the same mean but more variance in our estimate.
…
Now that we understand the intellectual adventure, what can we conclude? It is simply building in the conclusion the authors wanted. Analysts could easily flip their arguments about uncertainty upside down to get the opposite answer. What if, for example, under “business as usual” the Earth would have significantly cooled, thus moving up the left-side of the stipulated “convex damage function”? In other words, as Earth approached another ice age, the damages from further cooling would be higher and higher. In such a scenario, human activities that trapped heat would be a blessing, deserving of government subsidies, not taxes (in the conventional Pigovian framework).
…
My question: Why couldn’t we apply the same reasoning to ANY THREAT AT ALL? There are killer asteroids, alien invasions, superflus, nuclear war, earthquakes, and on and on it goes. Why not, for example, have governments restrict the emission of radio waves, on the off-chance that too much “noise” will attract the attention of aliens who will conquer us? You might argue that we are really uncertain about the nature of such a threat, but hey–that just shows all the more why we need to cut radio wave emissions by 80% relative to 2005 levels.
Instead of writing “My question: Why couldn’t”, I think it would have been fair for you to have put “that’s why”. We are certainly lucky to live in the age of federally funded science monopolies and research cartels.
Bob, I made that point here:
http://econlog.econlib.org/archives/2013/06/uncertainty_can.html
Hey David,
Yeah a bunch of you guys were doing good stuff; I linked to Art’s post because he in turn linked to you and David Friedman on this stuff.
Sounds like you got interested in doing some research on Pascal’s Wager. Indeed, it could equally apply to any God (even a Green God), or as Rowan Atkinson put it, “The Jews were right.” Search the skit on You Tube if you haven’t seen it… A good gag. The man’s very clever.
Just off the record Bob, I get the impression that you are motivated because you personally like the idea of being Christian, and not because you live in fear of the consequence of failing to be being Christian. Maybe that seems a strange distinction to draw but perhaps it changes one’s perspective seeing the equation in terms of “more positive” as opposed to “less negative”.
Disclaimer: someone else shouted vodka tonight, I’m just as likely to pretend none of this ever happened 🙂
This has been a problem for a while. There is no perspective. With carbon capture and coal gasification research we could reduce foreign oil dependence add energy security cut energy costs and also actually dent global warming in a way chindia can copy.
At a cos comparable to the latest Obama Boondoggle.
I agree with what you say here except “cut energy costs – at least for medium term.” Any treatment of coal before or after combustion is bound to cost more than simply burning it and letting the combustion gasses go. This is why the market will not produce this outcome.
There I am referring specifically to the coal gasification part that could convert the huge coal reserves into more flexible fuels that can chase price markets.
Also, the market will certainly not choose to pursue such technologies by Obama effectively banning coal.
Attempting to be more doctrinaire about it, does the market tend to over or under-invest in research…in the economics theory view?
If the options that Krugman refers to (in the subsequent post) don’t actually exist and are the result of research, why is he so insistent on tacitly arguing against government’s role in basic research as a bona fide public good?
In the absence of catastrophic global warming (to which the solution would be to revert to previous levels of carbon production and their level of prosperity) the actual solution to carbon is technologies that don’t yet exist in rollout form, not simple preferences we could change just because Obama wants something new this week.
Obama got together with his political advisors and they said “people won’t let us go for nuclear” so they settled on screwing coal.
But someone needs to explain to Obama that want doesn’t get.
He needs to realize, as Salam points to Manzi pointing out, it is the new technology of Frakking that has made his opportunistic desires politically and practically possible by cutting natural gas prices.
http://blogs.reuters.com/reihan-salam/2014/06/06/technology-not-regulation-is-the-best-way-to-tackle-climate-change/
So if the damage function IS convex, you agree with the paper. Extra uncertainty raises the mathematical expectation of damage. Call this conclusion 1.
As you say, this apparently leads to the surprising conclusion that “Greater disagreement among experts increases the likelihood that the risk of exceeding a global temperature threshold is greater.” Call this conclusion 2. This conclusion on its own does lead to odd policy outcomes, such as the alien invasion one. How can we reconcile these apparently conflicting but correct conclusions?
Surely what matters is the shape of the damage function. If the disagreement among experts does not affect the shape of the damage function, then conclusion 2 is true. However, if the disagreement is *about* the shape of the damage function then the uncertainty does not necessarily result in expectation of greater damage.
How does this fit with the alien invasion case? In the climate example you used “Damage” on the Y axis and “warming” on the X. For alien invasion we would have “Damage from invasion” on the Y axis, and “number of invasions” on the x axis. This is not likely to be a curve, but a flat line until “1”, then a jump to total destruction from the Emperor Zarg. If there is more uncertainty among experts about how much radio would push us over the threshold, then it IS more likely mathematically that we would reach this threshold. However, more likely does not mean that it is very likely. If the likelihood of total destruction doubles from 0.00001% to 0.00002% we need not be too concerned, and even though the risk is now double what we perceived it to be before, we will still conclude that cutting radio emissions would not be a correct response.
However, were some expert to say “No, I have evidence that the aliens will be helpful, not conquerors” – then that changes the shape of the damage curve. If this is correct, we have a flat line, then a large negative jump at “1”. This uncertainty surely reduces the expected damages, and this uncertainty suggests we should emit more radio waves.
The same would apply to killer asteroids. Increased uncertainty about how many big asteroids there were would probably lead us to spend more on detection.
So we must not get our uncertainties mixed up. Your example of the cooling world would be a different shaped damage curve, and would lead to benefits from emissions. The paper talks only of uncertainty in climate sensitivity given a convex damage curve.
So I think a broad claim that increased uncertainty must lead to greater expected damages is wrong, since it depends what the uncertainty is about. But who made that claim? The paper is talking about climate sensitivity, and stipulates that the conclusions apply if the curve is convex. It seems form our above discussion that everybody agrees with this narrow conclusion. If the curve is convex, then greater uncertainty in the climate sensitivity leads to greater expected damages.
The important question then becomes “is the curve convex?” If my analysis is correct, the more certainty we have about this *raises* the expected damages. If we have a 50% chance of a concave curve, then expected damages would cancel out. Is there any evidence of a concave curve? Yes, there is some, in the short term from some damage models. Is it a serious consideration that the curve continues concave? I don’t think so. We are so perilously close to the change form concave to convex of even the most optimistic estimates, that the chances of not getting to the point where the curve is convex is vanishingly small. To suggest this is a bit like saying the insurance models are all wrong because the payoff function is actually convex.
The main problem is that people are conflating the science with policy so they can dicate their desired policy outcome, by claiming that if you disagree with their preference you are a science denier and should be invited to neither debate.
The question is how much do we devote to rogue nukes or more directly people dying from the cold by drastically increasing increasing carbon costs.
Or, how much do we study the frakking risks because banning coal is de facto mandating dramatic increases in natural gas use.
Problems are relative.
“people dying from the cold by drastically increasing increasing carbon costs.”
How about moderately increasing carbon costs? No need for alarmism here!
No. People are going to die. I’m not alarmist whatsoever.
You can’t just shut down coal plants and increase electricity costs at all without people dying from the cold.
Why does this confuse?
I’m not criticizing you.
I’m seriously wondering why the fact that the hear and now marginal deaths from heating is somehow alarmism compared to a relatively far-off risk from sea levels or ocean acidification.
I think this irrationality is why people seem to be willing to ban coal power in order to mandate natural gas burning when we know this isn’t the actual solution to the problem.
It was the word “drastic” that I was labelling alarmism. It is a word open to interpretation, but when warmists use it they are accused of alarmism.
Well, if you want to add a sensible resource consumption tax, that, to me, by definition wouldn’t be drastic.
Obama, effectively banning coal power:
That isn’t “drastic”?
According to the estimates Bob recently talked about, it will cost $50 billion a year. The USA economy is 17 trillion, so it will cost at most 0.3% of the economy, possibly much less. Whether that is drastic is arguable. It is almost certainly more expensive that it needs to be to achieve the same carbon cut, as you say.
Harold wrote:
However, were some expert to say “No, I have evidence that the aliens will be helpful, not conquerors” – then that changes the shape of the damage curve…
There are experts in climate science who say that the modest warming they expect in the 21st century will actually confer net benefits. So now you agree that uncertainty among experts weakens the case for gov’t intervention, right?
So we must not get our uncertainties mixed up. Your example of the cooling world would be a different shaped damage curve, and would lead to benefits from emissions.
No, you’re getting mixed up. In my example, I had temperature change on the x-axis and damages on the y-axis. With a *certain* convex damage function (of my form), and a *certain* knowledge of climate sensitivity, we can still say that gov’ts should subsidize (not tax) carbon dioxide emissions if we are uncertain about the baseline temperature change. E.g. maybe solar and other forces interact in such a way that, without massive greenhouse effect, global temperatures would drop 5 degrees C, and because of that pesky convex damage function (which is U-shaped, meaning it goes way up once you go left too much) that means we really don’t want to take chances, so we’d better emit some more Co2 just to be safe.
There are experts in climate science who say that the modest warming they expect in the 21st century will actually confer net benefits.
Who says that?
Josiah,
Depending on how liberal we are with the term “experts in climate science” I could easily come up with people, but for example I’m pretty sure Richard Lindzen would agree with my statement. I mean, we know for sure that Richard Tol agrees with it through about 2065, right? So are you saying you doubt there are two qualified people on planet Earth who think that window extends to 2100, rather than stopping at 2065?
So are you saying you doubt there are two qualified people on planet Earth who think that window extends to 2100, rather than stopping at 2065?
I was mainly curious if you had anyone particular in mind.
Josiah, don’t take it to the bank that Lindzen believes that. I am just saying, I would guess that he does.
http://joannenova.com.au/2010/04/co2-is-the-magic-gas-that-makes-plant-grow/
http://www.climatecentral.org/news/study-finds-plant-growth-surges-as-co2-levels-rise-16094
http://www.plantsneedco2.org/default.aspx?menuitemid=325
Presuming Bob isn’t sick of my link spam, this is another good one, but there’s heaps more out there:
http://www.stanford.edu/~moore/HistoryEcon.html
“There are experts in climate science who say that the modest warming they expect in the 21st century will actually confer net benefits. So now you agree that uncertainty among experts weakens the case for gov’t intervention, right?”
It depends on the uncertainty. The uncertainty here *is* the shape of the damage curve – it is concave in some models for modest rises. I said that this does reduce the case for action – the benefits here are what I meant when I said ” Is there any evidence of a concave curve? Yes, there is some, in the short term from some damage models.” I was agreeing that all uncertainty does not strengthen the case, only uncertainty in the x axis.
You are right – the cooling world is not a different damage curve, it is hypothesising that we may end up on the concave section. If the uncertainty were large enough for this to happen, then I agree it would weaken the case for action to curb emissions.
So we can conclude that some uncertainty (about the damage resulting from a particular temperature rise – the Y axis) will not necessarily strengthen the case for action.
Uncertainty in the climate sensitivity – effectively how far along the x axis we go with a certain CO2 rise will have different effects depending on the size of the uncertainty. If the uncertainty causes the likely temperature rise to extend along the convex part, then the case is stronger. If the uncertainty is large enough that it causes the temperature rise to extend “backwards” to the concave section, then this will weaken the case for action.
In practice, I don’t think anyone is saying that the uncertainty extends backwards in this way. In theory it could, but the uncertainty does not do so in practice. Nobody is seriously predicting a cooling world in the decadal timescale in the absence of AGW.
Harold wrote:
You are right – the cooling world is not a different damage curve, it is hypothesising that we may end up on the concave section.
At this point let’s stop arguing climate science and instead argue about geometry. (Abortion is next.) I’m saying we can have a curve that is everywhere convex; y=x^2 for example. So if x is negative, that means the world cools. And you could imagine the more it cools, the greater the incremental damage per additional degrees of cooling.
So to repeat, we could be utterly certain about the damage function, and the damage function could be everywhere convex, and yet “uncertainty about the climate” could definitely weaken the case for limiting CO2 emissions, depending on what the “uncertainty about the climate” is.
These researchers picked a very narrow set of things about which we are putatively “uncertain,” in order to get their result.
Apologies – I mis-used concave, I was mixing up downward sloping with concave . This does not change my argument. If the world cools sufficiently we get more damage. If the world warms sufficiently we get more damage. My point was that nobody predicts a cooling world over the next few decades. So whilst you are technically correct, that greater uncertainty does not have ot lead to greater case for action, in practice, the uncertainty we have *in climate sensitivity* does.
I don’t think the alien invasion analogy works.
We aren’t in fact very uncertain about whether radio waves will attract alien invaders leading to the end of humanity. We are certain that the chances of this happening are very small. If we thought that the chances of this happening might be high, then maybe we would restrict radio waves.
That’s not to say that the argument is right with respect to climate change, of course.
I think when alarmists (in popular media, not necessarily in scientific studies) refer to uncertainty, they are talking about the uncertainty of the RESULT, not necessarily the uncertainty of the cause/effect relationship.
Alarmists aren’t saying there is uncertainty as to whether human CO2 emissions will result in higher temperatures (that science is settled, remember). They are saying that there is uncertainty as to whether this warming will result in moderate inconveniences and a net loss of 0.05% of GDP, or whether it will result in the extinction of all life on Earth. Their point is that since we can’t *totally rule out* the extinction of all life on Earth, it’s better to be safe than sorry, and engage in draconian policies to reduce emissions.
So, for the alien analogy, the uncertainty wouldn’t be regarding whether radio waves attracted aliens. The uncertainty would be over the likelihood that said aliens would enslave us. The equivalent analogy would involve scientists saying “look, there’s a possibility that any aliens we attract with radio waves (or better yet, the SETI program, which is explicitly TRYING to attract aliens) would enslave all of humanity, therefore we need heavy-handed governments to eliminate all such programs and greatly restrict radio waves, after all, better safe than sorry, right?”
Their point is that since we can’t *totally rule out* the extinction of all life on Earth, it’s better to be safe than sorry, and engage in draconian policies to reduce emissions.
There are people who make that argument, but that’s not the argument that the folks Bob’s is critiquing here are making Bob’s summary of their argument is pretty good, in my opinion).
For the most part, I think the two sides are talking past each other. To the layman, “uncertain” is vaguely similar to “unlikely.” So if you say that greater uncertainty makes the case for action stronger. But when climate scientists talk about uncertainty they are (often) talking about confidence intervals.
I think we can rule out extinction of life on Earth. Even the very worst case will leave life on Earth. Earth will get on with it all quite OK even if we have 50ft sea level rise in 20 minutes.
I presume the aliens must have explained this to you.
Tel wrote:
I presume the aliens must have explained this to you.
If this were Facebook, you would have garnered at least 15 Likes for that one, Tel.
If this were Facebook, you would have garnered at least 15 Likes for that one, Tel.
Ah, Facebook, that infallible arbiter of human wisdom.
BTW, Bob and Tell, what probability do each of you assign to the possibility that radio waves from earth will lead to an alien invasion?
BTW, Bob and Tell, what probability do each of you assign to the possibility that radio waves from earth will lead to an alien invasion?
For me, it’s low, but I’m not certain that it’s low. And remember, we have through 2300, to be consistent with the climate change analyses. What’s the marginal cost of having an oldies’ radio station in the mix, if it makes the chance of invasion in the year 2285 go up by one-billionth of a percentage point? Discounting at 1.2% which is the ballpark of what the climate alarmists do?
I think it’s unknowable. We could presume that aliens are limited by the speed of light, same as we are, so the implications of radio launched today would lead to consequences possibly thousands of years in the future.
At any rate, the question you are asking is besides the point, the AGW alarmists want us to take action based on uncertainty, and they argue that greater levels of uncertainty imply a greater requirement for action. The exact probability is irrelevant, the real question is whether there’s something you can usefully do about it, based on close to zero real knowledge of what you are doing.
Let’s suppose some government committee did decide that electromagnetic radiation from Earth was dangerous, so they decided to shut down all microwaves between 1GHz and 2GHz… would this help? How would they come to this conclusion? Maybe just to be sure they decide to shut down more communication systems, but what about 50Hz/60Hz power systems? Maybe the aliens can detect that as well, better shut down the whole power grid. What about infra-red? What about visible light from our cities? Better shut down everything just to be one the safe side, because we are highly uncertain about this so we better take really drastic action.
Then again, the aliens could figure out that Earth was habitable just judging from the orbital distance from the Sun and looking at atomic spectra. They might invade anyway, regardless of our efforts.
We don’t know that shutting down industry will even prevent CO2 from rising. Given that during the GFC and around 2008/2009 the USA already had an extended period of consuming less fossil fuels, and the observation was this had no effect on the rate of CO2 rise in the atmosphere (not even a blip). If we did decide to shut down industry we would have no idea if it was helpful or not.
We don’t even know whether all that stuff about banning CFC’s will someday help reduce the ozone hole, because the ozone hole that was a big scare back in the 1980’s is still there today, it just doesn’t get reported on, because people are bored now.
http://earthobservatory.nasa.gov/Features/WorldOfChange/ozone.php
There it is, practically no change in size for several decades (but the shape changes often), so what effect did it have making our refrigerators less efficient? We don’t know. It was action based on uncertainty, and the outcome remains uncertain.
“We don’t know that shutting down industry will even prevent CO2 from rising.”
More alarmism. Nobody is talking about shutting down industry. Or at least no more people than are talking about 12ft sea level rises
Regarding the ozone hole, from your link: “the long-term trend in both characteristics is consistent: from 1980 through the early 1990s, the hole rapidly grew in size and depth. Since the mid-1990s, area and depth have roughly stabilized.”
So the action has successfully stopped the trend – but not yet resulted in reversal. If we had not taken action it is very likely that the trend would have continued, and the hole would now be much deeper and larger.
Perhaps surprislingly, Cass Sunstein’s 2005 volume, Laws of Fear, makes a similar point.