Understanding the Social Cost of Carbon

One of the most hotly debated issues in climate policy is the value of the social cost of carbon, which is an estimate of the damage that will come from releasing carbon dioxide into the atmosphere. The social cost of carbon is a useful measure to help us understand the price that should be placed on carbon today to limit carbon dioxide emissions, and minimize the climate-related damages that future generations will face.

Climate economist Gilbert Metcalf explains how the social cost of carbon is calculated, and looks at the factors that economists take into account in arriving at a value. He also discusses why the value of the social cost of carbon is so contentious, and why the cost estimates accepted by the Trump and Obama administrations diverge so widely.

Full Transcript

Andy Stone: Welcome to the Energy Policy Now podcast from the Kleinman Center for Energy Policy at the University of Pennsylvania. I’m Andy Stone. One of the most hotly debated issues in climate policy is the value of the social cost of carbon, which is an estimate of the damage that will come from releasing carbon dioxide into the atmosphere. The social cost of carbon is a useful measure to help us understand the price we should put on carbon today, such as through a carbon tax, to limit carbon dioxide emissions and the climate-related damages that future generations will face. The value of the social cost of carbon is debated in part because it will bring added costs to certain parts of the economy, particularly those reliant on fossil energy. In addition, projections of future climate damages are, by their very nature, uncertain.

On today’s podcast, a climate economist will talk about how the social cost of carbon is calculated and the many factors that economists take into account. My guest is Gilbert Metcalf, a Professor of Economics at Tufts University and Research Associate with the National Bureau of Economic Research. His work focuses on taxation, energy, and environmental economics. Gib, welcome back to the podcast.

Gilbert Metcalf: Good morning, Andy. Great to be here.

Stone: You know, I thought we could start out just with the basics. Could you tell us a little bit about the social cost of carbon — what it is, and why it’s such an important number?

Metcalf: Sure. As you said, the social cost of carbon is a measure of the societal damages from releasing one more ton of carbon dioxide into the atmosphere. And just for context, what is a ton of carbon dioxide? So that’s essentially burning half a ton of coal for electricity, or the amount of gasoline that a typical driver uses in four months of driving. And we care about that because the damages aren’t reflected in the price we pay for filling up our gas tank or heating our homes, and so policy is a way to — and particularly a carbon tax or cap and trade — is a way to include that cost of those damages in our purchases and use of fossil fuels.

Stone: Can you illustrate, in a broad sense, the types of future damages that we’re talking about here?

Metcalf: Well, we’re talking about things like increased mortality rates due to more frequent very hot days. We talk about the fact that the temperature increase might go up two or three degrees Celsius, which doesn’t seem like very much, but that does translate into more persistent extremely hot spells, and with it, increased mortality. There’s the loss of productivity. Think about farm workers and other people working outside — crop damage, damage to homes, businesses, and other infrastructure from more frequent and severe extreme weather — things like Hurricane Harvey in Texas a few years ago that dumped 48 inches of rain in Houston, Texas in 24 hours. And then there are things like the increased cost of coping with these hotter days and more extreme weathers — what we call “adaptations” — so storm-proofing, use of air conditioning, hardening our infrastructure. These are all costs that society has to bear that aren’t reflected in our use of fossil fuels.

Stone: And I imagine a lot of these damages we’re talking about — we’re assuming what these damages are, right? — because they’re in the future. We can’t know exactly what they are.

Metcalf: Well, we actually are getting a clearer sense because we are experiencing the impact of climate change right now. And so scientists and economists are getting much better at measuring what they are, rather than just sort of making model assumptions. But it’s true that what we don’t know is how much hotter it is going to get, in part because we’re learning every day more about the science of climate change, and we’re not quite sure what the policy response is going to be.

Stone: Now, the actual dollar value of the social cost of carbon has been very hotly debated. The Obama administration calculated the social cost of carbon at about fifty dollars. The Trump administration’s social cost of carbon estimates have been much, much lower than that. Why is that?

Metcalf: Right, they’re down to about a dollar. And there are really two big changes in the approach that they took that led to that dramatic drop. The first is that they focus only on domestic damages. The Obama administration, in their approach, said, “This is a global externality. Our emissions are impacting people around the world, just as their emissions are impacting us. Therefore, we ought to think about this from a global perspective.”

If you focus on domestic damages — and the reason the Trump administration does that is because OMB regulatory analysis guidelines say, “Focus on domestic damages for typical regulations.” Of course, climate change is an atypical kind of pollution problem, so it’s not clear you want to follow the guidelines in that respect. But be that as it may, if you do that, that’s going to cut the damages to about one-seventh. So that gets you from $50 down to about $7 or $8.

And then the other thing that they did was they used a much higher, what we call a “discount rate,” which is essentially the interest rate we use to convert future damages into today’s dollars. And they increased it from — the central estimate in Obama was 3% — to 7%. And that’s going to cut the damages to about one-seventh of what they are again. So that’s going to get you all the way down to about one dollar a ton.

Stone: So basically, the higher that discount rate, the less money needs to be invested today for future damage. Is that correct?

Metcalf: Well, the less that we would value those damages in the future. So here’s the way to think about it: If we think about a million dollars of damages in 250 years — because keep in mind, greenhouse gases persist for a very long time — how do we think about those damages in 250 years? A million dollars is about $600 if you discount it at 3%, but it’s only a nickel at 7%. So it’s a huge difference.

Stone: Okay, so how is the social cost of carbon calculated?

Metcalf: Well, the typical approach is to use what are called “integrated assessment models.” I actually have an article in the June issue of Scientific American where I sort of walk through how I use these models, how they’ve been used historically, and how they were used by the Obama administration to come up with the social cost of carbon.

The way the models work — these are models that combine a model of the carbon cycle, a model of the economy, as well as the damages from increased emissions. The most famous one is the dice model that Yale economist Bill Nordhaus developed, for which he got the Nobel Prize in Economics a few years ago. And these models — the way they use them is — they pulse the model. In other words, they add one more ton of greenhouse gases, say, in 2020. And then they compute the damages in each of the years going forward for 200 or 300 years, take the present discounted value, and add those up, and you get a dollar estimate.

Stone: So what is today’s accepted value for the social cost of carbon?

Metcalf: Well, it depends on who you ask. First off, it’s going to differ for different years, so let’s 2030 — a decade from now. If you look at the Obama cost of carbon estimates from their last estimates in 2016, or if you look at Bill Nordhaus’ estimates, you get about $52 a ton. Again, if you look at the Trump administration, you get a buck a ton. I don’t think anyone really takes the Trump number seriously who are seriously looking at climate change. And I think mainstream economists, by and large, would view the Obama estimates — the $52 a ton number, or even Nordhaus’ number — as the lower bound. And I say that because what we’ve seen, looking at these numbers over the years, is that the estimates are growing as we’re getting better at measuring damages and including in our damage estimation some factors that simply we weren’t able to incorporate in years past.

Stone: Now, in your new article in Scientific American, you spend quite a bit of time talking about three important inputs that are used in the calculations with the integrated assessment models. One of them, the discount rate we’ve already spoken about a little bit, but could you please give us an overview of these inputs?

Metcalf: Right. So as you stated, the discount rate is the key one. The second one is — What is the dollar measure of damages? And as I hinted at, we’re getting better and more inclusive at capturing all of the damages, or more of the damages. But one of the problems, and there’s a lot to be hopeful about there, because I think there’s an amazing amount of work being done by great economists around the world to measure damages. I think it’s really the new frontier in economics, in climate policy, in measuring damages.

Another issue is that there are certain scientific parameters that are hard to pin down that affect damages. The key one is something called “equilibrium climate sensitivity.” This is a number that tells you how much we expect the global temperature to go up if we double our greenhouse gas emissions. And I think scientists — if you ask them what that number is — the central estimate is going to be roughly 3 degrees Celsius, meaning a doubling of greenhouse gas concentrations would increase temperatures by 3 degrees. But there’s a wide range of estimates of that, and it could be lower, but it also could be quite a bit higher — as much as 6, or even more degrees higher — in which case the damages would be much, much greater for higher values of that parameter.

Stone: You know, one interesting thing that goes on here is if you look at estimates of the social cost of carbon, say, from a decade ago, those estimates are significantly lower than they are today. So it seems that the social cost of carbon keeps going up over time. Why is that?

Metcalf: Well, that’s true. And I think part of that is we have a better understanding and measurement of damages. But I think there are a couple of other factors. I should also mention that the third sort of great unknown is — how should we treat catastrophes, which are an extreme form of damage? These are events that are, by definition, extremely unlikely, so it’s hard to know how to model them in an integrated assessment model. You know, you could do what are called “Monte Carlo simulations,” where you do thousands of runs with different values or parameters drawn from a distribution. But the very nature of a catastrophe is it is so unlikely that you could do thousands of runs and may never get a catastrophe showing up in your realizations. So how did Nordhaus address that? Well, he just simply did this rule of thumb and increased his estimate of damages by 25%. There was no particularly good reason for doing that, other than he just knew that his number was too low.

So why is this number going up? Well, we’re getting better at measuring damages, but there are two other things that are driving the increasing number. One is that we have a better understanding of the carbon cycle. For example, the DICE model in its iterations over the years, as Nordhaus has improved his representation of the carbon cycle, that has driven up the social cost of carbon quite a bit. And the other thing is that the economy is simply growing more rapidly than has been predicted in years past. And with a larger economy, you have larger opportunities for damages.

Stone: There is a great deal of uncertainty regarding the discount rate that should be used to calculate the present value of future climate-related damages. This discount rate is often framed in moral terms. Why is that?

Metcalf: So just to reiterate, we care about the discount rate because damages last far into the future. The lower that discount rate, the more we’re going to value those damages, and the more action we want to take today. So where do ethics come in? Where does morality come in?

Well, there are three key elements in the discount rate, and two of those really do reflect ethical considerations. One of them is something called “the pure rate of time preference.” That’s the economics jargon. It’s the rate at which we want to think about the well-being of people today, versus the well-being of people in the future, after we’ve controlled for the state of the economy and how wealthy they are.

And some argued — and Lord Nicholas Stern, the British economist who led the Stern Review, the exhaustive study of what we know about climate change, about a decade ago — Nick Stern would argue that we should use a pure rate of time preference equal to zero, which effectively means you should treat future generations precisely the same as current generations. And that’s completely an ethical argument. There’s nothing that economists can really opine to sort of give guidance on that. It’s something really more that philosophers can say something about. So using a pure rate of time preference of zero is going to help drive that discount rate down.

The second element is how do we think about valuing a dollar of damages to a wealthy person, relative to a poorer person. And there’s a parameter called “the elasticity of marginal utility of consumption,” which is a lot of jargon for, “How do we think about taking a dollar from a rich person and giving it to a poor person?” And if we think that that’s welfare-enhancing to a great deal, then that’s going to lead to a lower discount rate, as well, over time. So those two elements really reflect ethical considerations.

The third element is simply how rapidly do we think income is going to grow over time? That certainly comes into play for thinking about how we value spending money today to reduce damages that will affect richer people in the future. But the growth rate of income is really just a purely economic consideration. That doesn’t have an ethical dimension.

Stone: Some economists have said that the integrated assessment models that we’ve talked about, which are used to calculate the social cost of carbon — that these models are useless, and that’s because there are just so many uncertainties around the assumptions that they rely on. What’s your view on this?

Metcalf: Well, I think we want to be careful how we interpret what these economists have said. And Bob Pindyck at MIT is one of leading economists who has taken this view. And what Pindyck has argued is that the integrated assessment models are useless for the purpose of coming up with a number for the social cost of carbon. So he would not say that they’re useless in general, but they’re useless for that particular purpose. And the reason is that it provides this veneer of certainty and sort of scientific rigor where there is all this uncertainty.

And I think my response to that is, “Well, we’ve got to come up with a number somehow,” and Pindyck has not come up with a satisfactory alternative. And I think we can use integrated assessment models. They’re a useful tool as an input for the number we come up with. I totally agree with them. We shouldn’t take these estimates as gospel and as certainty, but we should recognize that they’re based on our imperfect understanding of the world. But having said that, they’re based on our understanding with the information we have today. We’ll update that information. We’ll update the models with new information. And we just need to recognize that our estimates today may be too high — or probably more likely — too low. But I think they are helpful as sort of a starting point for coming up with that number.

The other thing I’d say about integrated assessment models — and I think Pindyck would agree with me here — is that even if you take a strong position that they’re useless for the purpose for calculating a social cost of carbon, they can be extremely useful if you want to know what’s the right trajectory of tax rates that would get us to, say, net zero emissions by 2050, for example. So you can use them for determining how to set tax rates, if you have a goal — whether it’s an emissions reduction goal, or maybe a revenue target. So they’re helpful in that regard.

Stone: You know, it’s interesting you bring up that issue of tax rates. The next question I would like to ask you relates to that. How do we understand the correlation or the relationship between a certain carbon price or a carbon tax and the actual extent of carbon dioxide emission reductions?

Metcalf: Well, we start from the basic premise that demand curves slope down, and if you raise the price of something, we’re going to use less of it. So let’s just start with sort of basic Econ 101. But then the question is, well, how much do they go down? And here we’ve got sort of two approaches we can take. One is to do the kind of modeling that integrated assessment modelers have done, as well as general equilibrium modelers, like Marc Hafstead at Resources for the Future, working with Larry Goulder at Stanford. They’ve built very detailed models of the U.S. economy, which are calibrated to our understanding of parameters based on statistical analysis of data. You can use these models to determine what the impact of these various carbon tax proposals would be.

So for example, Marc Hafstead, working with other researchers at Resources for the Future suggests that one proposal from the Climate Leadership Council, which would levy a $40 per ton tax on our emissions — that would cut emissions immediately by about 18% and cut them in half by 2035.

Stone: So how realistic is a carbon tax set at the social cost of carbon, particularly given that we’ve never seen a carbon price anywhere near that high in the U.S.?

Metcalf: That’s a good question, and that’s really a question for Capital Hill watchers and political scientists, but I will just note that the $50 price is pretty darned close, and in fact is, essentially, the starting tax rate in seven carbon tax bills that have been filed in the U.S. Congress this session.

I think what we’re seeing is that the Obama-era social cost of carbon is a focal point for legislators. And just to give some context for that, since people don’t think in terms of dollars-per-ton of carbon dioxide, a $40 per ton tax rate translates into 38 cents per gallon of gasoline and something on the order of 2 cents per kilowatt hour of electricity. So it’s not a huge increase, but of course none of these tax bills would just leave the price at $40 or $50 a ton. They would have it go up over time.

So I think the price is realistic. I think the greater challenge is getting a tax enacted, period. That’s the hard sell. And that would be hard to do, whether you’re talking a tax rate of $10 a ton or $100 a ton. Just to get over that hurdle is going to be really huge. I think the actual initial price in whatever legislation eventually passes Congress is less the issue than the very fact of having a carbon tax.

Stone: It’s interesting, as you note in the new article, in Europe some countries have managed to implement some fairly aggressive carbon price numbers. Is that right?

Metcalf: Well, that’s right. I detail these in my Scientific American article, and what you see is that Europe’s carbon taxes can range as high as $120 per ton of carbon dioxide. The median rate of the 15 countries with carbon taxes is $19 a ton. So we’ve seen carbon taxes there. We also have the example closer to home, which is the province of British Columbia as a province-wide carbon tax, which today is at $40 per ton — $40 in Canadian dollars.

Stone: So let me ask you this, and this is one of the key questions that people often think about when they talk about anything that has to do with a price on carbon. What are the economic and employment impacts that you would expect from a carbon tax that truly reflects the social cost of carbon?

Metcalf: Well, we have enough data now that we can actually do empirical analyses as opposed to rely on models. And so if you look at my analysis of British Columbia’s province-wide carbon tax, and if you look at analyses of the carbon tax in Europe that I’ve done with Harvard economist Jim Stock, what we find is that there is a negligible impact on GDP or aggregate employment. In fact, many of these models suggest that you could actually get a modest positive impact on employment, given the fact that you’re taking the revenue from a carbon tax and recycling it through the economy.

So I think the answer is, “No big deal” at the aggregative level. We will expect, and we have seen — and there’s a great study looking at jobs in British Columbia by an economist, Yamazaki, that indicates we are definitely going to get shifts in jobs away from the fossil fuel industries and from carbon-intensive industries, towards the non carbon-intensive sectors. So we’re going to see big shifts in employment. We know that we have tens of thousands of jobs in wind in the United States, and hundreds of thousands of jobs in solar energy, as documented by the Department of Energy. So I think part of the job of policy-makers is not only to set a price on emissions, but to help with that transition away from a carbon-intensive to a carbon-free economy.

Stone: All right, so Gib, where does public opinion come into play in terms of implementing a carbon tax?

Metcalf: What’s striking is that polling that has been done by Yale University researchers, along with colleagues at James Madison University — they find that a majority of Americans support the Climate Leadership Council’s fee and dividend approach. And strikingly, an even stronger majority favor a carbon tax, with the revenues used to cut taxes.

But to be clear, there is an ideological split here. Independents and Democrats strongly favor this approach, but less than a majority of Republicans support it. But I think that’s quite remarkable — that you’re still getting somewhere between 40 and 50% of Republicans who are saying, “We should be doing these sorts of policies, too.”

Stone: So a final question for you, then. How do you actually go about implementing a carbon price?

Metcalf: There’s a lot of good work that has been done here, and I should start with the Willie Sutton maxim — Willie Sutton, of course, being the bank robber who was asked, “Why do you rob banks?” And he said, “Well, that’s because that’s where the money is.” He actually didn’t say it. A journalist made it up. But be that as it may, the Willie Sutton maxim says, “Look, most of our greenhouse gas emissions come from burning fossil fuels, creating carbon dioxide.” So that’s where we start. And it’s easy to do. You can piggy-back on existing federal excise taxes so the administrative overhead is quite low to do this. So that’s the way to do it.

Stone: Gib, thanks very much for talking.

Metcalf: You’re welcome. It has been a pleasure.

Stone: Today’s guest has been Gilbert Metcalf, Professor of Economics at Tufts University and a Research Associate with the National Bureau of Economic Research. Keep up to date with research and news from the Kleinman Center for Energy Policy by subscribing to our monthly email newsletter on our homepage. And you can get immediate updates by subscribing to our Twitter feed @KleinmanEnergy. Thanks for listening to Energy Policy Now, and have a great day.