29 Aug 2016

Cass Sunstein on Social Cost of Carbon

Climate Change, Shameless Self-Promotion 9 Comments

My latest at IER. How do you guys feel about my demonstration below? I think it’s pretty decisive, but then again, I’m presumably biased to like my own argument.

To avoid  confusion, I’m not going to use indenting etc. The italicized block quote below is from Cass Sunstein, and then the rest is my reaction.

==================================

First established by the Barack Obama administration in 2010, the central value for the social cost of carbon, last updated in 2015, is now $36. That figure is set within a range from $11 to $105, meant to acknowledge scientific and economic uncertainty. (Disclosure: As administrator of the White House Office of Information and Regulatory Affairs, I was involved in the process.) The $36 figure has international resonance; many nations are paying attention to it. It also plays a large role in discussions about the size of any possible carbon tax.

Although Sunstein doesn’t seem to realize it, the above statements contain a shocking admission. The “scientific and economic uncertainty” surrounding the concept mean that the Obama Administration was reporting a range for the “social cost of carbon” going from $11 to $105. When journalists or others cite the point estimate of $36, that is simply the “central value” in the broader range. As Sunstein himself admits later in his article: “Any particular number will of course be highly controversial.”

That range is enormous if the “social cost of carbon” (SCC) is going to be the foundation of the cost/benefit analyses of regulations involving greenhouse gas emissions. What would the world need to look like for Sunstein (or the circuit judges) to agree with the petitioners, that the SCC is an arbitrary concept?

Things that are actually settled science do not exhibit such controversy. For example, this NASA page reports that the (equatorial) radius of the moon is 1,738.1 kilometers. Now that’s just a rounded estimate, of course. If we polled various astronomers and physicists, they might disagree slightly on the best figure to report for “the radius of the moon.” However, no reputable scientists would say, “Any particular moon radius reported is controversial, but we are confident the range of the true figure is between 531 kilometers and 5,069 kilometers.” Would that make you confident that the scientists had a good idea of how big the moon was? Note that I used the same ratio of lower and upper bounds to the “central value” that Sunstein used for the social cost of carbon.

Or take human population. Nobody knows for certain how many people are on planet earth right now, but surely the proper experts have a pretty good idea. This website for example says the figure is some 7.4 billion people. To repeat, this is obviously just an estimate, and various scientists might disagree with each other and the best figure to report. But surely they wouldn’t say, “To account for the uncertainty in our analysis, we will report a range for the total global population between 2.3 billion and 21.6 billion. This is settled science—we are really confident the human population is somewhere in that range.”

If a government agency issued that kind of statement—saying the global population was somewhere between 2.3 billion and 21.6 billion—after studying the matter, and then issued regulations and taxes costing trillions of dollars on the basis of these calculations, would you feel good about the process?

Of course not. Such huge disparities in the bounds would indicate that scientistsdon’t really have a handle on these issues. And by the same token, when Cass Sunstein proudly tells us that the Obama Administration thinks the social cost of carbon is $36 per ton, but that it issued a range of $11 to $105 because of the “scientific and economic uncertainty” (his phrase, not mine), that should be Exhibit A in the petitioners’ case.

9 Responses to “Cass Sunstein on Social Cost of Carbon”

  1. Tel says:

    Officer, I swear I was doing somewhere in the range of 11 mph to 105 mph.

  2. Darien says:

    I think your example is pretty strong. In fact, as soon as I got to the part of Sunstein’s quote gleefully admitting the immense range, my eyes began to pop out of my face. It does indeed seem a bit of a massive, insurmountable flaw!

    Also, in point of fact, 87% of all arguments containing the word “resonance” are garbage. Or, well, 13% to 141%. Somewhere in that range.

  3. Silas Barta says:

    Wouldn’t your argument apply just the same to e.g. estimating damages in a civil tort? Those have all the same issues about estimating magnitude of the damages.

    (I’m sure you’ll bite the bullet there about the metaphysical issues surrounding “quantification of pain and suffering”, but most people aren’t on your side there.)

    • Bob Murphy says:

      Silas, maybe I’m being naive, but if a group of experts couldn’t pin my tort damages to less than a factor of 10, then I would agree it was not a “science” and was pretty arbitrary. I take your point to be, “But that doesn’t mean we’d throw out lawsuits right?” but that is still not the justification the public has gotten with climate change.

      • Silas Barta says:

        There is a variety of methods to calculate damages and the range can be very wide, so yeah, we’re already there.

        And what do you mean “that’s not the justification the public has gotten”? The justification is, carbon emission has negative externalities, they need to be assessed so that harm-spillers account for such costs when economizing.

        • Bob Murphy says:

          Silas, I”m saying the public has heard, “Pay no attention to those ‘deniers’ telling you that there are uncertainties. The science is settled. It’s tax time baby, so we can offset the damage we’ve been doing.”

          • Silas Barta says:

            If you jump off a 3 story building, you can be scientifically certain that you’ll hurt yourself, while still being uncertain whether that will mean broken bones or death.

            You can likewise be certain that temperature will rise 3-4 degrees but have much greater uncertainty about how much economic damage it will translate into.

            If scientists have been understating the size of the error bars, and/or those error bars include “CO2 is good on net”, then that would be a valid criticism. (And yes I know over certain time windows this is in fact the case.)

  4. Toby says:

    Bob, I don’t see how it follows that because the range of the social cost of carbon is large that it is not a science. For example, a chaotic system is also very sensitive to the initial conditions making it pretty much impossible to say what the state of the system is a few periods out. You, however, can still have perfect knowledge of how such a system evolves over time provided you get the initial conditions right. That seems pretty scientific to me. Evolution is another example, we can predict and describe the general pattern, but we can’t provide a precise “point estimate” there either. These are two counter examples to the claim that lack of precision means that something is not scientific.

    Furthermore, I am guessing here that the range is huge partly due to the sensitivity of the social cost of carbon to the social discount rate: i.e. how are we going to distribute the cost between the generations? This is a matter of taste. The cost of carbon, though, is probably a bit more precise. If we ignore the social discount rate for a moment, then how precise is the estimate?

    Looking at the figures here: https://www3.epa.gov/climatechange/EPAactivities/economics/scc.html we can say that going from a discount rate of 5% to 3% we go from $11 to $36 on average. If we then compare the average impact to what I presume is practically the worst case scenario at 3% we go from $36 to $105. Compare this to a fact such as smoking, there the estimate ranges that smoking increases your chances of lung cancer 15 to 30 times http://www.cdc.gov/cancer/lung/basic_info/risk_factors.htm. That’s a pretty wide range I’d say. This means that the range of the expected cost of smoking is pretty wide as well, right? And I don’t think you’d say something equivalent about the dangers of smoking.

    I think that your second point about the national vs. international cost of carbon is much stronger. Though, I wonder about two things there. First, the much lower benefits perhaps also imply that the damage to US citizens due to carbon emissions that are outside their control is much higher or am I mistaken here? Second, even if I am mistaken here does this mean that the US should price the social cost of carbon below the cost to the world as a whole? Imagine that second hand smoke is more dangerous to others than smoking is to you, then wouldn’t it at the very least not be very nice when you don’t take into account the damage suffered by others when you light one up?

    Finally, isn’t the proper response to a wide estimate for the social cost of carbon reason for alarm? For example, wouldn’t you want young children to wear a helmet when cycling even though most of the time nothing bad would happen to them? There too we have a wide estimate of what potentially could happen to a child when cycling, yet most parents, I hope, would still find it prudent to take some action.

  5. Harold says:

    Who has said that estimate of social cost of carbon is settled science? That is a straw man. The settled science is that carbon dioxide is causing warming. Nobody says that even the exact extent of that warming is settled and certainly nobody says that the costs of that warming are settled.

    So point out the large range, by all means, but to suggest that this contradicts the settled science idea is just wrong.

    Another example. We all know the science of guns. Yet Under certain reasonable assumptions, the average annual marginal social cost of household gun ownership is in the range $100 to $600. Things that are actually settled science do not exhibit such controversy.

    The Social Costs of Gun Ownership

    Phillip J. Cook, Jens Ludwig

    NBER Working Paper No. 10736
    Issued in September 2004

Leave a Reply to Silas Barta

Cancel Reply