The IPCC has published an erratum for our chapter in the Fifth Assessment Report.

Four data points were changed. Two relate to a paper by Roberto Roson and Dominique van der Mensbrugghe. Roson's key contribution was to introduce the impact of climate change on labour productivity into the analysis of the total cost of climate change. The concluding section of that paper presents two estimates per scenario: the total impact, and the share of labour productivity in that total. Michael Mastandrea, the co-head of the Technical Support Unit double-checking the numbers in our IPCC chapter, thought that Roson instead presents the impact of labour productivity and its share in the total. Mastrandrea checked his reading with Roson, who confirmed, and the estimates were changed. This is one of the discrepancies between the IPCC chapter and my paper in the Journal of Economic Perspectives (and the forthcoming paper in REEP).

Later on, Robert Kopp was checking the numbers again, asked Roson for the underlying data, and found that my original reading was correct. Mastandrea and Roson were wrong. JEP was right, IPCC was wrong. The erratum sets the record straight: The correct estimates by Roson are lower than the incorrect ones.

The other two changes relate to a paper by Robert Mendelsohn, Michael Schlesinger and Larry Williams. Mendelsohn presents his results for population-weighted temperature changes. Everybody else in this literature uses area-weighted temperature changes, and Mendelsohn duly reports those numbers as well. Double-checking our results, Mastandrea insisted that the population-weighted temperatures are used -- these show positive impacts at a lower temperature because the world population is concentrated in the tropics which are projected to warm more slowly than the globe. Violating IPCC procedure, Mastandrea ignored our protests. This is another of the discrepancies between JEP and IPCC. The erratum sets the record straight. The numbers shown for global warming for different studies are comparable to one another. Mendelsohn's estimates show benefits at a greater warming.

In sum, the Technical Support Unit of IPCC WG2 introduced four errors into the Fifth Assessment Report. All four errors exaggerate the impact of climate change.

Update (12 Oct 2016): Chris Field, former chair of IPCC WG2, submitted a call for an erratum to the erratum, reverting the changes made to the Mendelsohn estimates. Field's argument is that Mendelsohn's area-weighted temperatures are land-only. There is no dispute there. Field overlooks, however, that Mendelsohn's populated-weighted temperatures are land-only too (as rather few people live in the ocean). Mendelsohn's area-weighted temperatures are therefore less incomparable to other studies than his population-weighted temperatures.

Update (16 Dec 2016): Two months later, we're still going back and forth. Field continues to dispute our reading of Mendelsohn. We offered to show both estimates in an amendment to the erratum, but Field just wants the erratum gone.

Update (29 Aug 2017): Unable to find agreement between the co-Coordinating Lead Authors and the former Working Group Chair, the current Chairs appointed a committee of three to adjudicate. They found against Field's interpretation: Both estimates will be shown in an amendment to the erratum. However, in an apparent attempt to save Field's face, there will be a vaguely worded footnote that is likely to cause confusion rather than clarity.