The paper has now received two anonymous reviews plus one comment from M. Verbitsky, followed by another round of discussion / answers.

Focusing on what have been the major comments, two issues have caused concern from reviewers.

First is that the LGM climate would be an unrealistic straw man.
The objective of the authors is that using well-used standard damage functions to a cool climate similar as experienced during the last glacial maximum yields obviously wrong economical damage estimates.
The experiment is meant to highlight aspects of climate change that experts believe would be costly, yet currently overlooked by standard economic models. The dispute evolves around the argument that reaching LGM conditions in ca. 100 years is absurd anyway, which weakens the strength of the argument that the computed cost would be absurd, too.

Apart from acknowledging the caveat and further clarify the objective (which seems to have satisfied reviewer 2), I do not see any other way to fully resolve the issue at this point, but the point of the view of the authors deserves to be published. I would therefore authorize publication given the modifications proposed by the authors after their interactions with M. Verbitsky and Reviewer #2.

The second issue, mainly raised by reviewer #1 relates to the fact that the BHM model has in itself serious caveats ("equation 1 is wrong"; the BHM model is "known to be wrong"). The authors respond that the BHM model has been widely used, and that the author's specification match the BHM publication. They would have welcomed further references or information about the BHM semantic inconsistencies which reviewer #1 alludes to, but unfortunately the discussion turned short. The author's study is further complemented by an unpublished cost function by Newell et al. 2018.

I believe the only way to resolve this issue is for authors to identify the general traits of the damage function which they speculate are problematic, such that their study has some applicability beyond the particular case of the BHM and NPS models.

From the interactive discussion, it appears that the authors are well-prepared to implement these modifications and I would therefore accept the article under this condition.

An additional important technical comment: the web address of the Newell et al. model is currently not complete in the current version of the article. Fixing it will not be enough. Some solution should be found to guarantee long-term access to the model specification. One way is to convince Newell et al. to get a DOI reference. A better approach would be to make the computer code used for the current contribution available on a long-term archive like Zenodo, or as supplementary material of the present issue.

Dear authors,

I am glad to advise publication of your article.

I believe you have done the most that you could to justify the experiment, even though the non-physical character of an immediate transition to an LGM-climate, cold in spite of lacking ice sheets remains, I must admit, somewhat disturbing. Yet, the article definitely deserves being published.

Before sending the electronic version to the publication office, please take some time to double-check the consistency of the use of past and present. For example, in the sentence
"The most striking feature of the last glacial world was the existence of large and thick ice sheets in the northern hemisphere", the present must be used (was -> is), because it is striking now.

Yours sincerely,

Michel Crucifix.