Abstract

If I were reviewing this manuscript for an economics journal, I would recommend revise and resubmit, with only minor revisions. Overall, the manuscript is well written in the sense that, since the reader understands immediately what is being estimated and why, that they could be lulled into the trap of thinking that the research is “obvious.” The final line before the conclusion is "We conclude that climate change poses a substantial threat to the world economy." But it is the careful empirical construction of the argument, and the identification that a literature has been mechanically missing (netting out) the effect of global temperature shocks while focussing exclusively on local temperature shocks, that gives this manuscript its evident worth. The manuscript consists of two parts. First, a reduced form impact of global temperature shocks on economic activity at the world and country level. Second, a structural model to convert those estimates into welfare losses and a valuation of the social cost of carbon. The first is the “heart” of the paper in my estimation, while the second half is where many articles stop short - the “so what.”

Summary Measures

We asked evaluators to give some overall assessments, in addition to ratings across a range of criteria. See the evaluation summary “metrics” for a more detailed breakdown of this. See these ratings in the context of all Unjournal ratings, with some analysis, in our data presentation here.1

Overall assessment (See footnote2)

Journal rank tier, normative rating (0-5): On a ‘scale of journals’, what ‘quality of journal’ should this be published in?3 Note: 0= lowest/none, 5= highest/best.

Written report

Manager’s note: I made some very small formatting adjustments and fixes, italicizing a phrase, adding paragraph breaks, etc. I also added the bracketed subsection headers.

[Overview and discussion of strengths and implications]

This is a reviewer report for "The Macroeconomic Impact of Climate Change: Global vs. Local Temperature." The version reviewed is labelled November 2024. If I were reviewing this manuscript for an economics journal, I would recommend revise and resubmit, with only minor revisions.

The manuscript's main claim is that macroeconomic damages from climate change are six times larger than previously thought (around 12% instead of 2% reduction in world GDP following 1 degree Celsius of warming). This is because previous work does not account for (or indeed misses the effect of) global shifts in warming. The effect of global (rather than local) warming is proposed to occur through many mechanisms (considered via the top-down rather than bottom-up analysis), while the warming increases extreme climate events mechanism is specifically evidenced. A suite of robustness checks, at all stages of the comprehensive analysis, assuage many of the concerns the applied economics reader may have.

These previously underestimated damages, when used to inform a standard neoclassical growth model, paint policy makers a dire portrait. The 2024 world GDP is around 20% lower than it would have been absent the already observed warming. Further, 2040's GDP will be an estimated 25% lower than it would have been under business-as-usual projections. The related Social Cost of Carbon, which is often used as a benchmarking for whether an emissions-reducing policy is “worth it” is $1,367 per ton of abated CO2. This is striking when compared to the fact that most decarbonization policies cost around $80 per ton.

Overall, the manuscript is well written in the sense that, since the reader understands immediately what is being estimated and why, […] they could be lulled into the trap of thinking that the research is “obvious.” The final line before the conclusion is "We conclude that climate change poses a substantial threat to the world economy." But it is the careful empirical construction of the argument, and the identification that a literature has been mechanically missing (netting out) the effect of global temperature shocks while focussing exclusively on local temperature shocks, that gives this manuscript its evident worth.

The manuscript consists of two parts. First, a reduced form impact of global temperature shocks on economic activity at the world and country level. Second, a structural model to convert those estimates into welfare losses and a valuation of the social cost of carbon. The first is the "heart" of the paper in my estimation, while the second half is where many articles stop short - the “so what.”

In my estimation, the result - that global temperature shocks have large and persistent negative effects on GDP - has the potential to reframe a literature that began with Dell, Jones, and Olken (2012)[1]. I hope that this work has the same effect as Goodman-Bacon for the DID literature.4 Empirically, it reads as the inevitable next step of the climate change - temperature - extreme event - GDP literature.

The manuscript contains numerous robustness tests, including [some…] involving much work (such as re-estimation with a temperature measure independently created than the primary one), and clear explanations about why each was implemented.

The data and methods used in the estimation of warming damages are state-of-the-art and conducted in a clear and reasonable manner. The use of the Berkeley Earth Surface Temperature Database is compared against using National Oceanic and Atmospheric Administration data, to little difference. Literature standard data are taken from ISIMIP, the Penn World Tables, and Jordà-Schularick-Taylor Macrohistory database. The methods used are approachable and appropriate - and robust to alternative research choices such as, when appropriate, not using the Hamilton projections.

[Requests to authors]

In section 5.2, at the moment the range for the SCC [social cost of carbon] which is reported are the 95% confidence intervals. However, policy makers often focus on just one number, and may choose to focus on the lower bound if they are sceptical about climate change impacts. From an application perspective, therefore, it might be better to focus on a different, but equally conventional, bound: the 90% confidence interval. Naturally, the 90% range will be narrower –possibly by quite a bit if the tails are relatively thin – and it will ensure that the (reported) lower bound of the SCC is higher. This is a taste-based, and normative comment, with a view to how this crucial work may be applied

My primary request of the authors is to provide a comparison between […] business-as-usual and the other pathways ahead. Give us, the readers, a section which tells us just “how bad” things can be, yes, but also under different IAM's [Integrated Assessment Models, presumably] and pathways, such as in Burke, Hsiang, Miguel (2015)[2] Figure 5.

My secondary request is to increase the lags applied throughout (but most specifically in section 2.2) to at least 7 (the appendix goes to 4) given the cited geo-physical literature noting cyclicality of solar cycles being as long or longer.

A taste request is to characterize - up front - just how out-of-sample a 1C change truly is. The largest [change] in the sample is 0.3C. Calibrating response damages to this largest, but not particularly uncommon, shock might add to the credibility of the paper rather than take away from its message.

References

[1]Dell, Melissa, Benjamin F Jones, and Benjamin A Olken. 2012. ‘Temperature Shocks and Economic Growth: Evidence from the Last Half Century’. American Economic Journal: Macroeconomics 4 (3): 66–95. https://doi.org/10.1257/mac.4.3.66.

[2]Burke, Marshall, Solomon Hsiang, and Edward Miguel. 2015. ‘Global Non-Linear Effect of Temperature on Economic Production’. Nature 527 (7577): 235–39. https://doi.org/10.1038/nature15725.

Evaluator details

1. How long have you been in this field?

   • 5 complete years

2. How many proposals and papers have you evaluated?

   • More than 30