Bjørn Lomborg reviewed my book, Climate Shock(Princeton University Press, 2015), joint with Harvard's Martin L. Weitzman, for Barron's over the weekend. He started it by stating that "global warming is real."
So far, so good.
But the book is not about whether the climate is changing. It is.
The book is about whether we are getting the order of magnitude of its effects right. Weitzman and I argue forcefully — in prose in the text, supported by a significant amount of research going into the 100-page end notes — that it's what we don't know that really puts the "shock" into Climate Shock. Lomborg asks how we can know that, if apparently we don't.
The answer is simple, and it's a statistical point that can't possibly be lost on Lomborg, a former lecturer on statistics. The set of distributions that most directly represent climate uncertainty — the link between concentrations of carbon dioxide and eventual temperature outcomes — is inherently skewed. We know, and Lomborg agrees, that adding carbon dioxide increases temperatures. (Back to 19th century science.)
So we can very clearly cut off the distribution linking a doubling of pre-industrial concentrations to temperatures at zero. In fact, we can cut it off at least at around 1 degree Celsius (almost 2 degrees Fahrenheit). The world, after all, has already warmed by over 0.8 degrees Celsius (around 1.5 degrees Fahrenheit), and we haven't yet increased pre-industrial concentrations by even 50 percent.
Reprinted from Climate Shock, with permission from Princeton University Press.
That skewedness of the underlying distribution is real. It's important. The correct response, then, to those who are too sure about where the climate system will go isn't to say, "cool it." It's to take the uncertainties seriously. Those, sadly, are skewed in one direction.
Climate risk is not our friend. It ought to prompt us to rethink not just how we talk about climate change. It should also inform our response. The burden of proof clearly rests on those who argue against these statistical facts.
In fact, it’s not our quiz. Robert Socolow from Princeton has posed versions of these questions for a while. The result is usually the same: most people answer “Yes” to one or the other question, but not to both. You are either one or the other: an “environmentalist” or perhaps, a self-described “realist.”
Such answers are somewhat understandable, especially when looking at the polarized politics around global warming. They are also both wrong. Climate change is incredibly urgent and difficult to solve.
What we know is bad
Last time concentrations of carbon dioxide were as high as they are today — 400 parts per million — we had sea levels that were between 20 to at least 66 feet higher than today.
It doesn’t take much to imagine what another foot or two will do. And sea levels at least 20 feet above where they are today? That’s largely outside our imagination.
This won’t happen overnight. Sea levels will rise over decades, centuries and perhaps even millennia. That’s precisely what makes climate change such an immense challenge. It’s more long-term, more global, more irreversible and also more uncertain than most other problems facing us. The combination of all of these things make climate change uniquely problematic.
What we don’t know makes it potentially much worse
Climate change is beset with deep-seated uncertainties on top of deep-seated uncertainties on top of still more deep-seated uncertainties. And that’s just if you consider the links between carbon dioxide concentrations in the atmosphere, eventual temperature increases and economic damages.
Increasing concentrations of carbon dioxide are bound to lead to an increase in temperatures. That much is clear. The question is how much.
The parameter that gives us the answer to this all-important question is “climate sensitivity.” That describes what happens to eventual global average temperatures as concentrations of carbon dioxide in the atmosphere double. Nailing down that parameter has been an epic challenge.
Ever since the late 1970s, we’ve had estimates hovering at around 5.5 degrees Fahrenheit. In fact, the “likely” range is around 5.5 degrees plus-minus almost three degrees.
What’s worrisome here is that since the late 1970s that range hasn’t narrowed. In the past 35 years, we’ve seen dramatic improvements in many aspects of climate science, but the all-important link between concentrations and temperatures is still the same.
What’s more worrisome still is that we can’t be sure we won’t end up outside the range. The Intergovernmental Panel on Climate Change calls the range “likely.” So by definition, anything outside it is “unlikely.” But that doesn’t make it zero probability.
In fact, we have around a 10 percent chance that eventual global average temperature increases will exceed 11 degrees Fahrenheit, given where the world is heading in terms of carbon dioxide emissions. That’s huge, to put it mildly, both in probability and in temperature increases.
Climate Shock graph. There’s at least about a 10 percent chance of global average temperatures increasing 11 degrees Fahrenheit or more. Source: Climate Shock (Princeton 2015), reprinted with permission.
We take out car, fire and property insurances for much lower probabilities. Here we are talking about the whole planet, and we haven’t shown willingness to insure ourselves. Meanwhile, we can, in fact, look at 11 degrees Fahrenheit and liken it to the planet ‘burning’. Think of it as your body temperature: 98.6 degrees Fahrenheit is normal. Anything above 99.5 degrees Fahrenheit is a fever. Above 104 degrees Fahrenheit is life-threatening. Above 109.4 degrees Fahrenheit and you are dead or at least unconscious.
In planetary dimensions, warming of 3.6 degrees Fahrenheit is so bad as to have been enshrined as a political threshold not to be crossed. Going to 11 degrees Fahrenheit is so far outside the realm of anything imaginable, we can simply call it a planetary catastrophe. It would surely be a planet none of us would recognize. Go back to sea levels somewhere between 20 and at least 66 feet higher than today, at today’s concentrations of carbon dioxide. How much worse can it get?
Do we know for sure that we are facing a 1-in-10 chance unless the world changes its course? No, we don’t, and we can’t. One thing though is clear: because the extreme downside is so threatening, the burden of proof ought to be on those who argue that these extreme scenarios don’t matter and that any possible damages are low. So how then can we guide policy with all this talk about “not knowing”?
What’s your number?
We can begin to insure ourselves from climate change by pricing emissions. How? By charging at least $40 per ton of carbon dioxide pollution. That’s the U.S. government’s current value and central estimate of the costs caused by one ton of carbon dioxide pollution emitted today.
We know that $40 per ton is an imperfect number. We are pretty sure it’s an underestimate; we are confident it’s not an overestimate. But it’s also all we have. (And it’s a lot higher than the prevailing price in most places that do have a carbon price right now—from California to the European Union. The sole exception is Sweden, where the price is upward of $130. And even there, key sectors are exempt.)
How then do we decide on the proper climate policy? The answer is more complex than our rough cost-benefit analysis suggests. Pricing carbon at $40 a ton is a start, but it’s only that. Any cost-benefit analysis relies on a number of assumptions — perhaps too many — to come up with one single dollar estimate based on one representative model. And with something as large and uncertain as climate change, such assumptions are intrinsically flawed.
Since we know that the extreme possibilities can dominate the final outcome, the decision criterion ought to focus on avoiding these kinds of catastrophic damages in the first place. Some call this a “precautionary principle”— better to be safe than sorry. Others call it a variant of “Pascal’s Wager” — why should we risk it if the punishment is eternal damnation? We call it a “Dismal Dilemma.” While extremes can dominate the analysis, how can we know the relevant probabilities of rare extreme scenarios that we have not previously observed and whose dynamics we only crudely understand at best? The true numbers are largely unknown and may simply be unknowable.
Planetary risk management
In the end, this is all about risk management—existential risk management. Precaution is a prudent stance when uncertainties about catastrophic risks are as dominant as they are here. Cost-benefit analysis is important, but it alone may be inadequate, simply because of the fuzziness involved with analyzing high-temperature impacts.
Climate change belongs to a rare category of situations where it’s extraordinarily difficult to put meaningful boundaries on the extent of possible planetary damages. Focusing on getting precise estimates of the damages associated with eventual global average warming of 7, 9 or 11 degrees Fahrenheit misses the point.
The appropriate price on carbon dioxide is one that will make us comfortable that the world will never heat up another 11 degrees and that we won’t see its accompanying catastrophes. Never, of course, is a strong word, since even today’s atmospheric concentrations have a small chance of causing eventual extreme temperature rise.
One thing we know for sure is that a greater than 10 percent chance of the earth’s eventual warming of 11 degrees Fahrenheit or more — the end of the human adventure on this planet as we now know it — is too high. And that’s the path the planet is on at the moment. With the immense longevity of atmospheric carbon dioxide, continuing to “wait and see” would amount to nothing else than willful blindness.
As greenhouse gases accumulate and global temperatures slowly rise, what can we do to insure against the catastrophes of climate change? Economics correspondent Paul Solman talks to the authors of Climate Shock.
Each ton of carbon dioxide emitted into the atmosphere today causes about $40 worth of damages. So at least says standard economic thinking.
A lot goes into calculating that number. You might call it the mother of all benefit-cost analyses. It's bean-counting on a global scale, extending out decades and centuries. And it's a process that requires assumptions every step along the way.
The resulting $40 figure should be taken for what it is: the central case presented by the U.S. Government Interagency Working Group on Social Cost of Carbon when using its preferred 3% discount rate for all future climate damages. But it is by no means the full story.
Choose a different discount rate, get a different number. Yale economist Bill Nordhaus uses a discount rate of slightly above 4%. His resulting price is closer to $20 per ton of carbon dioxide. The Stern Review on the Economics of Climate Change uses 1.4%. The resulting price per ton is over $80.
And the discount rate is not the only assumption that makes this kind of a difference. In Climate Shock, we present the latest thinking on why and how we should worry about the right price for each ton of carbon dioxide, and other greenhouse gases, emitted into the atmosphere. There are so many uncertainties at every step—from economic projections to emissions, from emissions to concentrations, from concentrations to temperatures, and back to economics in form of climate damages—that pointing to one single, final number is false precision, misleading, or worse.
Of course, that does not mean that we shouldn't attempt to make this calculation in the first place. The alternative to calculating the cost of carbon is to use a big fat zero in government benefit-cost calculations. That's clearly wrong.
Most everything we know about what goes into calculating the $40 figure leads us to believe that $40 is the lower bound for sensible policy action. Most everything we know that is left out would push the number higher still, perhaps much higher.
It's not over 'til the fat tail zings
As just one example, zero in on the link between carbon concentrations in the atmosphere and eventual temperature outcomes. We know that increasing concentrations will not decrease global temperatures. Thank you, high school chemistry and physics. The lower bound for the temperature impact when carbon concentrations in the atmosphere double can be cut off at zero.
In fact, we are pretty sure it can be cut off at 1°C or above. Global average temperatures have already warmed by over 0.8°C, and we haven't even doubled carbon concentrations from preindustrial levels. Moreover, the temperature increases in this calculation should happen 'eventually'—over decades and centuries. Not now.
What's even more worrying is the upper tail of that temperature distribution. There's no similarly definitive cut-off for the worst-case scenario. In fact, our own calculations (based on an International Energy Agency (IEA) scenario that greenhouse gas concentrations will end up around 700 parts per million) suggest a greater-than-10% chance of eventual global average warming of 6°C or above.
Focus on the bottom row in this table. If you do, you are already ahead of others, most of whom focus on averages, here depicted as "median Δ°C" (eventual changes in global average surface temperatures). The median is what we would expect to exceed half the time, given particular greenhouse gas concentrations in the atmosphere. And it's bad enough.
But what really puts the "shock" into Climate Shock is the rapid increase in probabilities of eventual temperatures exceeding 6°C, the bottom row. While average temperatures go up steadily with rising concentrations, the chance of true extremes rises rapidly:
That 6°C is an Earth-as-we-know-it-altering temperature increase. Think of it as a planetary fever. Normal body temperatures hover around 37°C. Anything above 38°C and you have a fever. Anything above 40°C is life-threatening.
Global average warming of 3°C wouldn't be unprecedented for the planet as a whole, in all of it geological history. For human society, it would be. And that's where we are heading at the moment—on average, already assuming some 'new policies' to come into play that aren't currently on the books.
It's the high-probability averages rather than low-probability extremes that drive the original $40 figure. Our table links greenhouse gas concentrations to worryingly high probability estimates for temperatures eventually exceeding 6°C, an outcome that clearly would be catastrophic for human society as we know it.
Instead of focusing on averages then, climate ought to be seen as a risk management problem. Some greenhouse gas concentration thresholds should simply not be crossed. The risks are too high.
This kind of focus on temperature extremes is far from accepted wisdom. We argue it ought to be.
Think of the atmosphere as a giant bathtub. There’s a faucet—emissions from human activity—and a drain—the planet’s ability to absorb that pollution. For most of human civilization and hundreds of thousands of years before, the inflow and the outflow were in relative balance. Then humans started burning coal and turned on the faucet far beyond what the drain could handle. The levels of carbon in the atmosphere began to rise to levels last seen in the Pliocene, over three million years ago.
What to do? That’s the question John Sterman, an MIT professor, asked two hundred graduate students. More specifically, he asked what to do to stabilize concentrations of carbon dioxide in the atmosphere close to present levels. How far do we need to go in turning off the faucet in order to stabilize concentrations? Here’s what not to do: stabilizing the flow of carbon into the atmosphere today won’t stabilize the carbon already there at close to present levels. You’re still adding carbon. Just because the inflow remains steady year after year, doesn’t mean the amount already in the tub doesn’t go up. Inflow and outflow need to be in balance, and that won’t happen at current levels of carbon dioxide in the tub (currently at 400 ppm) unless the inflow goes down by a lot.
That seems like an obvious point. It also seems to get lost on the average MIT graduate student, and these students aren’t exactly 'average'. Still, over 80 percent of them in Sterman’s study seem to confuse the faucet with the tub. They confuse stabilizing the inflow with stabilizing the level.
Watch this video to avoid making the same mistake: