In my previous post, I summarized William Nordhaus’ new article, “Economics of the disintegration of the Greenland ice sheet,” which showed that standard estimates of the “social cost of carbon” were not much changed even if we explicitly include the “tipping point” of a melting Greenland ice sheet. I showed that the pushback from some in the scientific community—who warned of sea level rise that wouldn’t occur for centuries—demonstrated how baseless these warnings really were.

After defending his DICE model’s earlier estimates of the social cost of carbon, Nordhaus goes on to argue that government intervention is important, because a “no-action baseline,” even if followed by aggressive “geo-engineering,” would be too late. In response, in the present post I sketch out a scenario in which completely voluntary actions, without any government intervention at all, could achieve just as satisfactory an outcome regarding the Greenland ice sheet as Nordhaus’ recommended carbon tax.

A Much More Relevant Geo-Engineering “Experiment,” Relying on the Private Sector

To reiterate, Nordhaus in his new paper leads the reader to believe that it is important for humanity to embark on mitigation procedures—for example, by implementing his recommended carbon tax profile very soon—in order to slow the disintegration of the Greenland ice sheet. If humanity were to foolishly do nothing and then try to use geo-engineering down the road to clean up the mess, it would be too late. Here is Nordhaus’ specific analysis, from page 12267 of the official PNAS version of the article:

The geoengineering experiments in DICE-GIS are roughly the same as the results from the three modeling studies [summarized earlier in his paper—RPM]. Consider a scenario with a temperature anomaly of 6 °C global for 300 y, which leads to 2 m of SLR [sea level rise—RPM] at that time. If the temperature is reduced to zero in a geoengineering experiment, the GIS is estimated to rebuild by only 0.2 m after 1,000 y, or about 0.2 mm/y.

The reasons for the strong asymmetry is straightforward. An ice sheet melts when the temperature is elevated, and the decumulation (melting plus glacial discharge) exceeds accumulation (precipitation). However, there is no negative meltingin the cold phase. Rather, to build an ice sheet requires not just cold temperatures, but also precipitation. When temperature declines, precipitation tends to asymptote to low levels, and the ice sheet buildup is consequently extremely slow. This implies that there is a sharp asymmetry in the response of the ice sheet to positive and negative temperature shocks.

The conclusions of the geoengineering simulations have important implications for climate policy. The results apply to mitigation and carbon removal as well as solar-radiation management. They suggest that the disintegration of the GIS is essentially irreversible on a relevant societal timescale. The GIS will rebuild when temperature is reduced, but the growth is so slow that, from a human perspective, disintegration should be considered irreversible. [Nordhaus, bold added.]

Yet it is hardly a fair test of a “no-policy followed by geo-engineering” policy to let global temperatures rise a cumulative 6 degrees Celsius (relative to preindustrial times) and then wait another 300 years before doing anything about it. That would be a bit like saying building hospitals is a waste of time, because if you dispatch ambulances 3 weeks after someone calls 911, the patient is usually dead.

Murphy’s Private Sector “Experiment,” Step 1 of 3

In the remainder of this article, let me sketch out a much more relevant “test” or “experiment” of this strategy regarding the ice sheet. And note that I will rely on the same data and assumptions as Nordhaus in his paper.

First, let’s assume human-caused emissions follow the highest path in the IPCC’s latest batch of scenarios, namely RCP8.5, which actually involves more emissions than a standard “business-as-usual” projection. In other words, I am being generous by picking RCP8.5 as the starting point.[1]

As the authors explain on page 49 in this paper, “Cumulative CO2 emissions in RCP8.5 amount to about 7300 GtCO2 over the course of the entire century,” where “Gt” is gigaton, or one billion tons, of carbon dioxide. Thus, total human emissions over the 21st century in this high-end scenario amount to some 7.3 trillion tons of carbon dioxide.

Now in this thought experiment that I’m developing, let’s suppose that by the year 2100, normal market developments in battery technology, wind and solar electric generation, etc., combined with privately funded initiatives for planting trees and other such measures, render a net-zero human contribution to atmospheric carbon dioxide, going forward from that point. In other words, in this thought experiment, I’m assuming the “worst case” of governments doing nothing about climate change, up through the year 2100. Then at that point, I assume normal technological developments will have gotten humanity to the point where refraining from adding more carbon dioxide (on net) each year is so relatively easy that it can be achieved through voluntary, private sector means.

This is surely not a stretch for me to assume,[2] in light of all of the assurances from cheerleaders for wind and solar that these technologies will outcompete coal and natural gas even without government help, any day now. At the very least, if the critic wants to say that it would still take widespread government intervention in the year 2100 to keep humans from releasing net emissions, then that should tell you something about how economically destructive it would be to force that type of outcome fifty years earlier—which is what the now-popular goal of 1.5°C would require (see the executive summary of Chapter 2 of last fall’s IPCC document).

Murphy’s Private Sector “Experiment,” Step 2 of 3

Now for the next component of my thought experiment, let’s further assume that humans do nothing at all about the accumulated stock of CO2 they’ve pumped into the atmosphere, and that they maintain this posture for an additional century. That is, except for whatever sopping up measures they deploy (through the private sector) to achieve net-zero emissions starting in the year 2100, humans don’t do anything extra to deal with the legacy of the past. They merely say, “From this point in 2100 going forward until the year 2200, we will just make sure that we are no longer contributing carbon dioxide to the atmosphere, but we won’t lift a finger to do anything about what we already put up there.”

So of the 7.3 trillion tons of carbon dioxide lingering in the atmosphere[3] that is the accumulation from human emissions during the prior century, the question is: How much will natural causes alone remove, from 2100 through 2200?

Well, despite the true-but-possibly-misleading statements about carbon dioxide emissions contributing to global warming “for thousands of years,” actually about 70 percent of human-caused atmospheric CO2 would be removed naturally in a century. So of our stipulated 7.3 trillion tons of “excess” (human-caused) CO2 in the atmosphere as of the year 2100, only 30 percent—in other words, 2.2 trillion tons—would remain by the year 2200. The rest would have been dissolved into the ocean and absorbed by other natural “carbon sinks” over the course of the 22nd century.

Murphy’s Private Sector “Experiment,” Step 3 of 3

Finally, let’s suppose that starting in the year 2200, humans embark on geo-engineering in order to remove the excess carbon dioxide still remaining in the atmosphere. Now to be clear, there are many proposals for geo-engineering involving various ways to reflect sunlight back into space, but here I am merely talking about techniques for rapidly removing CO2 from the air. Further suppose that this geo-engineering program is given a 50-year timeframe for completion.

Assuming they could pull this off, the potential problem would be easily solved. By the year 2250, the atmospheric concentration of carbon dioxide would be back (at least) to year-2000 levels. Because of its elevated temperature, the Earth would be emitting more heat energy than it was taking in from the sun, leading to long-run cooling (according to the conventional climate models endorsed in UN reports).

As the scientific literature summarized in Nordhaus’ paper indicates, this scenario I’ve outlined would come nowhere near to a dangerous “tipping point.” For example, Nordhaus’ Table 1 (on page 12266 of the paper) shows that it would take a sustained human-caused warming of more than 6 degrees Celsius for 500 years before melting the first 20 percent of the Greenland ice sheet. In contrast, in our hypothetical scenario—using standard UN-endorsed models—global temperatures would only flirt with such extremes for at most a century, even with pessimistic assumptions about the climate system’s sensitivity to CO2.

Indeed, I’ll reproduce Nordhaus’ crucial figure showing the results from three different strategies:

SOURCE: Figure 8 (p. 18) of the longer version of Nordhaus, but also Figure 5 (p. 12267) of the official PNAS version, where it is currently missing the “base-geo” line.
SOURCE: Figure 8 (p. 18) of the longer version of Nordhaus, but also Figure 5 (p. 12267) of the official PNAS version, where it is currently missing the “base-geo” line.


Eyeballing the figure, it should be clear that if we slid the “Base-geo” green dotted line upward, so that it began in the year 2250, rather than the year 2515, then the volume of the Greenland ice sheet would barely have dipped below the blue dotted line, which shows its fate under Nordhaus’ “optimal” scenario. (And eventually, under the Murphy proposal, the GIS volume would be higher than the Nordhaus level.)

In the figure, the red arrow shows the range of volume change, given different assumptions about the ice sheet’s melt rate. Even if we assume the fastest melt rate, meaning we use the lower tip of the red arrow, nonetheless the Greenland ice sheet’s volume in the year 2250 will remain above 85 percent of its year-2000 level. (Just slide the red arrow up along the curve to the year 2250, and see where its bottom tip ends up on the y-axis.)

And so we see that if humanity could pull off my proposed task, they would be fine, at least regarding the threat of a collapsing Greenland ice sheet. Not only would they stay well above the 80% tipping point, but they would (eventually) have more ice sheet than what Nordhaus projects would occur under his optimal path.

Let’s now spell out the specifics, to see if my proposal is feasible: If the humans in the year 2200 start with a hangover of 2.2 trillion tons of excess atmospheric CO2, and they give themselves 50 years to remove it, then that means they will have to remove 44 billion tons per year. (For simplicity, I’m now ignoring the natural carbon sinks that would continue to help chip away at the excess stock, side-by-side with the new human techniques to speed up the removal.)

Now we have to wonder: How hard would humans in the year 2200 find it, to begin a program of removing 44 billion tons of CO2 annually from the atmosphere? Well right now, we have the technology to use machines to directly suck out carbon dioxide at a cost of $100-$150 per ton. Let’s say they figure out how to knock it down to $22 per ton by the year 2200. That means we’ve now got an annual cost of 44 billion x $22 = $1 trillion per year.

Yikes! That sounds like a huge expense, doesn’t it? But we have to remember: This 50-year program doesn’t start until the year 2200. Very conservatively let’s assume global real GDP grows at a long-run average rate of 2% per year until then. (Right now projections through 2050 are for 3% annual growth.) So if global GDP right now is around $85 trillion, then by the year 2200 it will have risen to some $3,060 trillion. Thus, to the people living at that time, our CO2 removal program would cost them 0.03% of global GDP per year, and that percentage would gradually fall as the global economy continued to grow. For some perspective, right now humans spend more than that—specifically, 0.04% of global GDP—on chewing gum.

For another data point, consider that in 2018 Americans gave about $430 billion to charity, which was 2.1% of US GDP that year. So if humanity as a whole in the year 2200 has the same degree of philanthropy as Americans do right now, then they would have to earmark (0.03 / 2.1) = 1.4% of their annual charitable giving toward the carbon dioxide removal program. Is it really farfetched “denialism” to think that humans could solve the potential problem of a melting Greenland ice sheet through voluntary means?


I have written two relatively long posts in order to demonstrate that those skeptical of government intervention do not need to rely on skepticism of the UN-endorsed physical science. As William Nordhaus’ new paper shows, the standard estimates of the “social cost of carbon” do not rise significantly, even when explicitly modeling the Greenland ice sheet. The extreme activists trying to throw out traditional cost-benefit analysis are shown, once again, to be bluffing.

Furthermore, even Nordhaus himself places far too much reliance on the need for government coercion. Using his own framework and estimates, I sketched out a scenario in which humanity “does nothing” until the year 2100, and then relies purely on voluntary programs to gradually eliminate any possible threat of a melting Greenland ice sheet, and to (eventually) have more volume in the Greenland ice sheet than occurs under Nordhaus’ “optimal carbon tax” scenario. And at no point in my story did I assume humans devoted more financial resources to this project than what they currently spend on bubble gum. The extreme climate activists will no doubt shift the goalposts to “ocean acidification” or other downsides to a voluntary approach, but I have clearly demonstrated that their warnings about London being underwater are absurd.





[1] Indeed some experts wonder if the RCP8.5 scenario is even technically possible, because it involves far more emissions than the current estimates of available fossil fuels and relies primarily on demand growth, rather than modeling the supply-side constraints as humanity consumes vast amounts of coal and oil.


[2] To avoid confusion, let me elaborate about the assumptions I’m making: In the RCP8.5 scenario, carbon dioxide emissions are still very high, even in the year 2100. (They level off at a bit over 80 GtCO2 per year at the end of the century.) What I’m doing in my thought experiment in the text above, is taking the advocates of wind and solar at their word, and saying surely it should not be very difficult for humans to achieve net-zero emissions in the year 2100, giving allowance for philanthropists to fund private mitigation efforts to sop up any stragglers. (For example, even with the RCP8.5’s assumed ~80 Gt of annual emissions, it would only require spending about $50 million annually on planting trees [with some lead time if desired because the trees need to mature] to counteract that, according to the figures I discuss in this article.) Then, in order to estimate how much carbon dioxide has been released by that point, I am using the emissions trajectory from the RCP8.5 scenario, to err on the side of caution. As a final clarification, let me mention that strictly speaking, I am documenting the emissions from 2000 – 2100, whereas humans also (of course) released emissions before 2000. But the anthropogenic contribution to climate change up through the year 2000 is not going to cause a crisis, in anyone’s book, and so it is fine for our purposes in this article to just deal with the anthropogenic contribution from 2000 onward.


[3] This figure represents another generous assumption on my part. My figure is the total emissions during the 21st century, which would be higher than the stock of anthropogenic CO2 in the atmosphere as of the year 2100. Yet the rising atmospheric concentrations during the 21st century would cause the oceans and other carbon sinks to increase their absorption rate. For example, the human emissions in the year, say, 2060 would already have had 40 years of natural processes to help remove them from the atmosphere, by the time stage 2 of my thought experiment begins. But to keep things simple, and to err on the side of caution, I’m disregarding that aspect of the analysis.

Print Friendly, PDF & Email