Wednesday, February 26, 2014

Determining the Social Cost of Carbon is Such a Gas

Climate change returns to the Supreme Court. On Monday, the United States Supreme Court heard opening arguments for the case of Utility Air Regulatory Group v. Environmental Protection Agency (EPA) to determine the extent to which the EPA can regulate greenhouse gasses (GHGs). I'll let the lawyers fight out the legal details in the Supreme Court, but it did make me think that if the EPA is going to regulate carbon, how is it going to determine the costs of carbon emissions?

In spite of the efforts by the Office of Management and Budget (2013) or the EPA's collaborative reportit's tricky to discern the social cost of carbon (SCC), which is the projected future marginal cost of emitting an extra metric ton of carbon. Why is that?  First, much like any other service the government provides, the government is incapable of adequately providing a value of its output, which is something I blogged about when criticizing the carbon tax about a year ago.

The second reason has to do with the nature of carbon. Carbon dioxide stays in the atmosphere for a century, which is a long time away. It is difficult to determine the effects because modeling the various impacts of climate change is imprecise. How is one supposed to accurately estimate environmental and economic impacts a century from now and tie them directly to current carbon output? Especially from a technological standpoint, but also politically and economically, so much can occur between now and then. Technology, especially technology emitting carbon to improve our lives, is the key to providing us with the ability to adapt to our environment, which is essentially impossible to predict. Also, how do you accurately blame mass migration, civil wars, extreme weather events, or economic cycles, all of which are already difficult to predict, on carbon emissions?!

Much of determining the value of the SCC is based on the computer modeling, also known as integrated assessment models (IAMs) of climate change scenarios, a process that makes me skeptical. Even economist Robert S. Pindyck, who is not a climate change skeptic by any means, points out how useless IAMs are (see here). IAM simulates economic activity in conjunction with our attempt to control carbon, which results in the extent to which we directly cause damage. What's more is that according to the IPCC's 5th Assessment Report (p. 16), its equilibrium climate sensitivity is likely between 1.5° and 4.5°C, which is lower than what the IPCC has predicted in the past. In spite of the lower estimated climate sensitivity (which is important because lower estimations means more time to find solutions to the problem), the SCC estimates just keep increasing. Go figure!

Let's not forget that the OMB is not reporting the global effects separately from the national ones (OMB Guidelines, 6.a.3), which has bearing in the sense that the United States should not go on the futile task of trying to unilaterally stop climate change. They call it "global warming" for a reason. Even with the United States' decrease in carbon emissions, in no small part due to increased natural gas production, there is not much the United States can do if China, the world's largest carbon emitter, does not cooperate.

One also has to contend with the issues behind the discount rate. The discount rate is a reflection of how much one is willing to pay now to avoid future damages. The higher the discount rate, the less significant the future costs. The discount rate captures three important assumptions: "humans prefer to receive benefits in the present rather than the future, that future generations will be richer and a dollar [will be] worth less less to them as a result, and the opportunity cost of capital (that there are a variety of investment options for any given sum)." According to federal guidelines (Section 8), cost-benefit analyses are supposed to include a discount rate of 7 percent, which the OMB forgot to do in the aforementioned 2013 report (the OMB only included 2.5, 3, and 5 percent [Figure 1]). It makes me wonder what the SCC would look like if the 7 percent calculations were included, although I could tell you they would either be negligible or possibly even negative. Given the assumptions made with the discount rate, we should not be so ready to dismiss the 7 percent discount rate as a possibility simply because its consideration has the plausibility of diminishing the SCC.

If the SCC were merely an argument over economic theory, I would consider this a nice, little academic exercise in polemics. However, the Obama administration is using the SCC to determine the extent to which it uses its discretion to regulate carbon emissions. Not only does production become more expensive, but the government is using the SCC to provide itself a carte blanche to regulate virtually any aspect of our lives. If we overprice carbon and it turns out the risk associated with increased carbon output is manageable, then the "War on Carbon" will be for naught.

This is not an argument about the veracity of ACC. It's about the importance of proper risk management and assessment.  Especially given the long timeframe and complexity of the variables in the modeling, we need to realize that we cannot prevent every low-probability, high-damage scenario, whether that is downright nuclear war, another terrorist attack, or that ACC will cause such turbulent weather patterns that it will wipe out humanity. Before coming up with solutions and scratching the itch to regulate everything that emits carbon, we should realize the complexities and limitations of IAMs and factor that into policy-making decisions before jumping into climate change policy that could do more harm than good.

3 comments:

  1. "First, much like any other service the government provides, the government is incapable of adequately providing a value of its output"
    You're double-counting problems here: In this case, the social cost of carbon combined with the size of reductions will yield the output. Of course it may be hard to calculate the reductions, but that's a data/statistics problem, not a government problem.

    "Also, how do you accurately blame mass migration, civil wars, extreme weather events, or economic cycles, all of which are already difficult to predict, on carbon emissions?!"
    How do we attribute anything to anything? Direct causal links are rare, but we can see larger patterns. It is, admittedly, imperfect, in all directions. We can just as easily underestimate the negative impacts as overestimate.

    "Even with the United States' decrease in carbon emissions, in no small part due to increased natural gas production, there is not much the United States can do if China, the world's largest carbon emitter, does not cooperate."
    Our own carbon-cutting actions make us look less blatantly hypocritical. They lower the competitive cost for China to take similar measures. They make US goods relatively more desirable for consumers who want to use consumer choice to help kill the world a little less quickly. The technology and methods developed can spread to the rest of the world, a sort of "trickle-down" emissions reduction. Even if we didn't care about carbon, saving energy is still something everyone wants to learn to do better.

    The primary assumption captured by the discount rate, when talking cross-generation, is our inability to comprehend, or care, that we could possibly be harming the next generation.

    "If we overprice carbon and it turns out the risk associated with increased carbon output is manageable, then the "War on Carbon" will be for naught."
    And we can just as easily underprice it, making our lack of action seem somewhere between irresponsible and suicidal.

    "we need to realize that we cannot prevent every low-probability, high-damage scenario"
    The only probability that is truly uncertain is the likelihood of action. If we change nothing and just keep on going there will be devastating effects. That's the difficulty of it all: there's a happy scenario where we fix the problem, a tricky one where we reduce emissions but maybe not by quite enough and then have to calculate all manner of problems and figure out if they were worth it, and then there's the runaway scenario where we tip over the point where it is too hot for the oceans to soak any more. At that point they start releasing it back out and if it gets too warm we run the risk of frozen methane being released.

    Think of it like blood. You wouldn't bother getting a transfusion from losing a few drops. At some point you'd lose enough that you might consider a transfusion. And then there's the point where losing too much blood means you die and all the cost-benefit analysis seems rather silly unless you assign very little value to your life.

    ReplyDelete
    Replies
    1. Andrew, as always, I appreciate your comments. Per the usual, I'll respond point-by-point:

      You're double-counting problems here: In this case, the social cost of carbon combined with the size of reductions will yield the output. Of course it may be hard to calculate the reductions, but that's a data/statistics problem, not a government problem.

      If only I were double-counting here. The inaccuracy of IAMs is separate from the notion that shadow pricing is a form of the government valuing its output. The fact that the OMB omitted the 7 percent discount rate is an example of that.

      How do we attribute anything to anything? Direct causal links are rare, but we can see larger patterns. It is, admittedly, imperfect, in all directions. We can just as easily underestimate the negative impacts as overestimate.

      If we can just as easily underestimate as we can overestimate, the certainty with which you predict Armageddon towards the end of your comments baffles me.

      The technology and methods developed can spread to the rest of the world, a sort of "trickle-down" emissions reduction. Even if we didn't care about carbon, saving energy is still something everyone wants to learn to do better.

      I need to better clarify that comment. I didn't mean for it to come off as "we should sit back and do nothing simply because China is going to increase its carbon footprint." Back when I blogged on the carbon tax, my skepticism was towards the carbon tax, as well as some other government-induced measures. I did and still do believe that R&D will be the best bet at reducing carbon emissions. That being said, I'm sure the next argument we can get into is the extent of government intervention in that endeavor.

      And we can just as easily underprice it, making our lack of action seem somewhere between irresponsible and suicidal.

      Theoretically possible. Objectively speaking, we're not going to know for about another century, by which time, odds are that we'll both be dead. Given that the OMB low-balled the discount rates a bit, I'm going to take an educated guess that it's overestimated.

      Delete
    2. The only probability that is truly uncertain is the likelihood of action. If we change nothing and just keep on going there will be devastating effects.

      This is where I become really confused by your earlier comment of "we can just as easily underestimate the negative impacts as overestimate." The confidence you have in the apocalyptic scenario astounds me. This is me getting into a disagreement about magnitude of the problem more than anything else. Estimated climate sensitivity is decreasing over time, and yet the situation is somehow getting worse. Take the economic cycles as an example of predictability. The CBO cannot even get it right with a 5-10 year timeframe. How do you expect climatologists to get it right with a century timeframe? This line of questioning can be used for predicting political instability or mass migration. I was unaware the government was so omnipotent and clairvoyant where it can predict these occurrences with such confidence.

      Think of it like blood. You wouldn't bother getting a transfusion from losing a few drops. At some point you'd lose enough that you might consider a transfusion. And then there's the point where losing too much blood means you die and all the cost-benefit analysis seems rather silly unless you assign very little value to your life.

      And I have to wonder why you or just about anybody else left-of-center hasn't applied this analogy to Social Security, Medicare, Medicaid, or the federal debt in general. I don't need a complex IAM system to tell you that those programs are more imminent of a threat to American longevity than climate change simply because they're set to hit us hard well before a century has come and gone. I would love to avoid a catastrophe, but without a proper risk assessment with all valid discount rates, predicting doom and gloom like that eerily comes off like the Jehovah's Witnesses.

      Delete