A recent Nature article, Rennert et al. 2022, estimates the social cost of CO2 summed through 2300. The authors find a total cost of $185 per ton of CO2, more than three times the value of $51 used in current U.S. regulatory decisions. $90 of that is due to increased mortality from higher temperatures, $84 to reduced agricultural output, $2 to sea level rise and $9 to energy costs for residential and commercial buildings.
This is the opening paragraph of David Friedman, “Critique of ‘Comprehensive evidence implies a higher social cost of CO2,” David Friedman’s Substack, July 30, 2023.
David submitted his critique to Nature, which was the publisher of the Rennert et al article. Unfortunately, and not surprisingly, Nature rejected it.
David’s critique is lengthy but here are some highlights.
The mortality calculation in Rennert is based on regional figures for increased mortality per degree of temperature rise from Cromar et al. 2022. Temperature-related mortality depends, among other things, on income since richer people can afford air conditioning and better insulated homes and have less need to go out in unfavorable weather. The economic model in Rennert implies per capita GNP roughly tripling by 2100, increasing about eleven-fold by 2300, but since Cromar does not include income in the relation between temperature and mortality Rennert ignores the effect of that increase on temperature-related mortality. Socioeconomic conditions are mentioned in Cromar as a factor to be considered in future work but the implicit assumption of the two articles taken together is that, despite the large projected increase of income, the relation between temperature and mortality will remain at the level of the recent past.
Yet, as David points out, it’s absurd to think that if people’s incomes triple on average, they would not take steps to protect themselves from higher temperatures. Indeed, an article by T. Carlton et al in the Quarterly Journal of Economics did just that. David writes:
Carleton et al. 2022, a more recent and more sophisticated calculation of the contribution to the Social Cost of Carbon from temperature-related mortality, uses the same time period and discount rate as Rennert but takes account of the effect on mortality of both income and the temperature distribution. It found a value of $36.6 for a high emissions scenario (RCP 8.5) and $17.1 for a moderate emissions scenario (RCP 4.5). The latter is much closer than the former to the assumptions in Rennert.
Focusing just on mortality, the difference between a social cost of carbon dioxide of $90 per ton and a social cost of $36.6 or $17.1 per ton is substantial.
David then goes to to consider the other components of social cost, showing that Rennert et al overstate the social cost for each.
If you click on the Rennert et al article, you’ll see multiple authors, many of whom are well-known economists at name brand schools. That makes their basic economic errors all the more stunning.
Friedman ends with this:
Rennert sums costs over the next three centuries, with about two-thirds of the total coming after 2100. Their solution to the problem of predicting technological change over that period is, with the exception of their estimates of CO2 production and energy costs, to ignore it, implicitly assume technological stasis. That is the wrong solution — but any projection of technological change that far into the future would be science fiction not science.
What they claim to do cannot be done.
David’s critique is well worth taking the time to read.
READER COMMENTS
Jon Murphy
Jul 31 2023 at 8:48am
Good stuff here. This post also highlights why I’m skeptical about carbon taxes even within their own logical framework. The administrative costs associated with altering the carbon tax as we take into account those charges in technology, agriculture, etc likely will dwarf any net gain in benefit from the carbon tax itself.
Thomas L Hutcheson
Jul 31 2023 at 3:12pm
Is this quantitative conclusion supported by a model of some kind?
Jon Murphy
Jul 31 2023 at 5:04pm
Yes. I’m working on a model now where I combine the current literature on administrative costs in Pigouvian models (Schell is the main guy, I think) with the issues discussed above and a Hicks demand curve. It’s a dynamic model. I’m hoping to present it at the SEA meetings but it might not be debuted until the Spring.
Thomas L Hutcheson
Jul 31 2023 at 9:10pm
Great! 🙂
Kevin Dick
Jul 31 2023 at 3:40pm
I think this is the best argument against carbon taxes. The difference between theory and practice is that, in theory they’re the same, but in practice they’re different. Or in more formal terms, there’s a meta cost-benefit problem.
Now, I think there’s a reasonable response to this reasonable objection. A carbon tax that is intentionally set low is still better than both no carbon restrictions and the incredibly distortionary measures we have now.
Of course, there’s yet another level or response. A carbon tax that starts out low is still an opportunity for abuse through capture and also potentially crowds out private solutions.
Then it’s political economy all the way down.
Jon Murphy
Jul 31 2023 at 5:26pm
That is a potential objection. Although I’d argue it’s irrelevant to the problem discussed. The issue is that estimates of social costs are too high. So, the proposed carbon taxes off those estimates will also be too high. That they could be lower and not cause as many problems doesn’t matter of the estimate is too high and you don’t know that lower estimate.
David Henderson
Jul 31 2023 at 6:24pm
You write:
True, and you know that I’m not an advocate of carbon taxes, but David does give the lower estimate.
Jon Murphy
Jul 31 2023 at 8:08pm
Yes he does give the lower estimate. My apologies for being unclear. I was making my comment in the context of policymakers and advisors ignoring David’s points.
Andrew_FL
Aug 1 2023 at 9:34am
He gives some quantitative examples of components of the calculation that are substantially lower, but as I read it, he does not calculate his own estimate. Which I think is right, “Can’t calculate it exactly, but we know it’s significantly lower than this” is good enough.
Kevin Dick
Aug 1 2023 at 3:12am
My comment was on your comment, not the original post. You said:
“The administrative costs associated with altering the carbon tax as we take into account those charges in technology, agriculture, etc likely will dwarf any net gain in benefit from the carbon tax itself.”
It seems like you had already assumed in arguendo that one could effectively set the original level of the because you prefaced this statement with, “… within their own logical framework.”
In my response to you, I was also assuming only in arguendo that we could satisfactorily make an initial estimate of the cost.
Jon Murphy
Aug 1 2023 at 6:28am
Ah! I see your point now. My mistake for misreading you. Thank you for the clarification.
steve
Jul 31 2023 at 10:38am
For. the countries with lowest incomes a tripling would leave them with incomes comparable to South Africa or Colombia. In the case of Afghanistan or Congo it puts them up in the range of Nepal. You are going to see better electrification though probably not for rural or poor areas. If you look at access to air-conditioning even relatively rich, in comparison, countries like Mexico have have poor air conditioning rates. So while the richer people will be safer the poor will be at more risk. The other part is that with increased wealth people live longer and older people have worse temperature regulation and more vulnerable.
I totally agree that they should not assume there will not be advances in geoengineering. There is also a strong tendency to ignore the rapid advances in the increasing efficiencies in Solar, wind and battery tech. Maybe SMRs work out. Maybe, probably not, someone solves fusion. However, they would open themselves up to criticism if they suggested these new technologies can solve this problem. You should just look at their estimates as a high side estimate. Look at Friedman’s as a low side estimate with his assumptions.
Steve
Thomas L Hutcheson
Jul 31 2023 at 4:13pm
A modeler would not assume that technology “solves” the problem of CO2 accumulation, but they certainly need to make assumptions about the costs of technologies that will be encouraged by taxation of net CO2 and methane emissions. Tax rates on net emissions would need to be constantly updated as scientific understanding of the effects of CO2 and methane accumulation increases and changing cost of reducing net emissions and mitigation.
Todd K
Jul 31 2023 at 10:00pm
“For. the countries with lowest incomes a tripling would leave them with incomes comparable to South Africa or Colombia. In the case of Afghanistan or Congo it puts them up in the range of Nepal.”
The poorest countries have per capita GDPs (PPP) of $1,000 to $1,500. South Africa and Columbia are at $16,000 and $19,000 respectively. Afghanistan already has the same GDP as Nepal, $5,000.
” You are going to see better electrification though probably not for rural or poor areas.”
95% of Nepal have access to electricity. 91% have access to the internet.
David Friedman
Aug 12 2023 at 5:07pm
I don’t give a low end estimate of actual cost, just arguments showing that, using their approach correctly, it’s much lower than they think. In my view the calculation of net cost is hard enough that we don’t even know its sign, let alone its size. The part of the effect on agriculture that we can be most confident of is positive, the large increase in yield from CO2 fertilization — everything else is much more uncertain.
That, uncertain sign, was also my conclusion on the effects of population growth some fifty years ago.
It’s worth noting that Rennert et. al. have a graph showing a (low) probability that the net cost of carbon is negative, even with all the things they do to inflate it.
Thomas L Hutcheson
Jul 31 2023 at 3:31pm
The big picture here is that estimates of the cost of CO2 accumulation have to take account of the positive effects of investment to mitigate costs AND the costs of those investments. For example if CO2 accumulation is modeled as raising sea levels but costal cities can build seawalls to prevent any damage from the sea-level change then the cost of the CO2 accumulation is only the cost of the seawall. If the deadweight loss of a tax on net CO2 emissions prevents the sea level rise is less than the cost of the seawall, it makes sense.
As I read Nordhaus (I suppose other modeler do the same), he does properly count only the net of mitigation cost + cost of mitigation ins the cost benefit analyses. The paper cited seems not to have been as careful or at least as complete.
john hare
Jul 31 2023 at 7:46pm
Technological stasis across centuries is a sticking point as I see it. A century ago very few had air conditioning and a high percentage of the population worked outside. This being before a 40 hour work week was common. I doubt there can be accurate forecasts for half a century, much less three. What’s the cell phone ownership rate in Afghanistan compare to 1990 US?
Todd Kreider
Jul 31 2023 at 10:08pm
“What’s the cell phone ownership rate in Afghanistan compare to 1990 US?”
0.5% of Americans had a cellphone in 1990.
65% in Afghanistan have a cell phone today.
Mark Brady
Jul 31 2023 at 8:31pm
Words can be defined anew, but it should be noted that the 2022 article in Nature explains that “the social cost of carbon dioxide (SC-CO2) measures the monetized value of the damages to society caused by an incremental metric tonne of CO2 emissions and is a key metric informing climate policy.”
Yet when economists use the expression “social cost” – and when I teach the topic in a principles course – it means the sum of private and external costs, and not just “the monetized value of the damages to society.” As Cowen and Tabarrok write in their principles textbook, “The social cost is the cost to everyone: the private cost plus the external cost.” Mankiw in his principles book uses a similar definition.
Thomas L Hutcheson
Aug 1 2023 at 5:30pm
But either way, for purposes of calculating the optimal taxes on net CO@ and methane emissions, the cost includes the net cost of mitigation.
Zeke5123
Aug 1 2023 at 7:22am
David Friedman’s piece was all well and good (seriously it is a solid piece). However, I came to the conclusion the original nature paper was bogus when I saw they were projecting costs in 2300.
278 years is just way too big of a time frame to model anything. The United States didn’t exist 278 years ago.
We probably cannot accurately model the next 27 years let alone ten times that. Nature should’ve rejected this fatal conceit with the epistemological humility needed to do science. Instead, we end up with something worse than no map; a made up map.
Todd K
Aug 1 2023 at 11:23am
“We probably cannot accurately model the next 27 years let alone ten times that. Nature should’ve rejected this fatal conceit with the epistemological humility needed to do science. ”
Nature and Science haven’t been serious journals with respect to much of Covid and climate change for many years. Not only did Nature accept the article but 24 economists at supposedly very good universities worked on a paper going out 278 years. Most bright junior high students would see this doesn’t make sense.
At this point, these types of forecasts break down after just 10 to 15 years, if that. What will solar, nuclear fusion, battery storage and AI look like around 2033 to 2038?
Thomas L Hutcheson
Aug 1 2023 at 5:35pm
Well modeled or not decisions take today will have consequences N years in the future.
Should we just impose a tax on net emissions of CO2 and methane and dismiss the concerns of skeptics because the are too hard to model? I don’t think so. It seems we shoud make the best decision we can today and revisit it tomorrow.
Zeke5123
Aug 1 2023 at 8:16pm
You should not model. Knowing you have no map is better than having a wrong map. You can make decisions in light of uncertainty. But thinking you have a map when you don’t is a big problem.
Floccina
Aug 1 2023 at 10:18am
Did they deduct for fewer cold deaths?
David Friedman
Aug 12 2023 at 4:53pm
Yes.
I am suspicious of their numbers, because total deaths from cold are much larger than from heat but they use numbers for which marginal deaths from cold are substantially smaller than from heat. But I haven’t looked at the article they are basing that on carefully enough to critique it.
Comments are closed.