Thursday, January 12, 2023

Wanted: Criticism

As I have mentioned here before, I have been working on a critique of a recent Nature article that finds a very high value for the social cost of carbon, the net cost from now to 2300 produced by an additional ton of CO2. I argue that the article greatly exaggerates the cost by errors which should be obvious to a careful reader, one of them being the implicit assumption that there will be no improvement in most of the technologies relevant to the costs, including medicine, for the next three centuries.


The implication of my critique, if it is correct, is that the enterprise of producing estimates of the costs of climate change is broken. The important point is not that this particular article makes mistakes but that it makes obvious mistakes, should never have gotten through peer review in its present form. If an article with obvious mistakes, all of which increase its estimate of the cost of carbon, can get published in a high status scientific journal and be considered by the EPA as a possible basis for regulatory decisions, it follows that the mechanisms that are supposed to filter out bad scientific work are not working in this context, hence that none of the published estimates can be trusted.


That is a strong claim and it might be wrong, but if it is right should be a serious problem for people who accept the current orthodoxy with regard to climate. I am therefor inviting anyone interested, especially anyone inclined to accept current views on the effects of climate change, to look over my critique and tell me what is wrong with it. 


Hostile reviewers are more useful, for my purposes, than friendly reviewers.


Dustin said...

It sounds like you're saying that once a mistake (publishing an obviously incorrect paper (just accepting your claim here)) is made, then nothing else can be trusted.

This seems questionable to me. Has Nature never published an obviously incorrect paper before? If it has, did you, and should have others, decided that no other prior papers ever published anywhere on the same subject should be trusted?

Saying that the whole enterprise is broken because of this paper seems like too strong of a claim from the data presented in your post. It should lessen your confidence in such papers, but saying they shouldn't be trusted is too strong without further data and verification.

georgesdelatour said...

Is the distribution of bad-papers-which-Nature-approves random? I suspect it isn’t.

Alan said...

Do your active commenters include any real mainstream climate scientists? I expect most such scientists would be unable to stand reading your blog regularly. I hope you're also looking elsewhere for criticism.

David Friedman said...


I put my request for comment on FB.

More relevantly, I wrote a shorter version of the article to submit to _Nature_, which has a 1200 word limit for articles commenting on articles they publish. They strongly suggest that someone writing a critical article should first send it to the authors of the article being criticized to get their response. I accordingly sent it to the corresponding author for Rennert, who responded that he was forwarding it to the other authors. I also sent a copy to the lead authors of the two papers that feed into Rennert, since some of my criticisms are of them. Unfortunately the lead author of Moore et. al. is now working for the Council of Economic Advisors and doesn't seem to be responding to academic matters, so probably won't read it.

I have, however, also corresponded with one of the other authors of Moore, who persuaded me that one of my criticisms was probably wrong, so I removed it.

I also plan to send a version of the article to the EPA in response to their request for comment, but am waiting for the response from the authors of Rennert before doing so.

I don't think "mainstream climate scientists" are necessarily the most relevant people to respond, since none of my criticism is of the climate science. I am accepting Rennert's figures on what will happen to temperature over the next three centuries and disagreeing with the implications for human welfare.

Part of my criticism is that the article ignores both the implications of climate science, in particular the pattern of warming in the IPCC projections, and the implications of their own model. Rennert predicts large increases in per capita GNP over the period it covers but ignores them in its estimate of the effect of temperature on mortality.

David Friedman said...


The key word is "trust." I expect some things written on the subject would stand up to criticism. But all of us are dependent for our beliefs about such issues on second hand information, so distinguishing between sources of information you can trust and sources you can't is important. It's natural to assume that respectable scientific publications can be trusted, meaning not that they never make mistakes but that they are honest attempts by competent people to find out what is true and report it.

When a high status scientific journal publishes an article which contains obvious mistakes, that's a reason to lower your trust in such articles. When the mistakes produce a result which many people, especially many academics, would approve of, that's a reason to suspect that the whole filtering mechanism in that field is badly biased, hence cannot be trusted. It's worth noting that Rennert spends two paragraphs on reasons why the cost of carbon might be higher than they project, not a single sentence, I think, on why it might be lower.

Two other points. The first is that it should lower your confidence in the reliability of published scientic articles in general, in particular ones that deal with issues that many people care about, hence have an incentive to bias the results in the direction they prefer. The second is that while this is the only evidence I am offering in this post for distrust of work in the field, I have quite a lot more. For example:

I am focusing on this one in the hope of persuading someone who disagrees with me to actually do the work of going over my article and seeing if he can find things wrong with it.

Glen said...

I suggest posting your query/request to the RealClimate blog. About once a month they post an "open thread" where there's no blog content, just an opportunity for the community to discuss climate stuff in the comments. The latest such link is:

(Be sure to save a copy of your comment before posting it as there's a chance any consensus-critical comments might be held in review a while or even outright deleted. Still if they DO post it, the denizens there should include people with the appropriate qualifications and background.)

David Friedman said...


Thanks for the suggestion. I have just put my post up as a comment on the RealClimate thread.

Unknown said...

The link to the critique is broken

David Friedman said...

Many thanks. I fixed it.

Nunzarius said...

A depressing amount of peer reviewed science is garbage. I would definitely say it is an over reaction to believe that mainstream climate science is wrong based on a bad paper. Especially one trying to predict anything out to 2300 which is a doomed exercise from the start.

David Friedman said...

The point is not just that it is a bad paper but that it is obviously bad, got published in a high status journal, and is being seriously considered by the EPA. That doesn't imply that all mainstream climate science is wrong but it does imply that the results of at least this part of it, the calculation of costs of climate change, cannot be trusted. Some might be right, but the fact that a conclusion is accepted by people in the field is very weak evidence that it is true.

I don't think the IPCC reports are nearly that bad. The Summary for Policy Makers presents information in a biased way, occasionally dishonest, as when it truthfully reports that the proportion of high end tropical cyclones is projected to increase but doesn't explain that the reason is a decrease in the number of low end cyclones, with the number of high end staying about the same. I don't think I have found anything that bad in the main body of the report, just a tendency to pay much more attention to the negative consequences of hotter summers than to the positive consequences of milder winters and the like.