Updating my prior: How recent research has changed my views of climate risk (and how it hasn’t)

Become a paying subscriber of Unpacking Climate Risk to read in full👇

Whenever I find a new research paper suggesting that 3°C will decimate the global economy, my initial reaction is always skeptical.  There’s an old adage that extraordinary claims must be backed by extraordinary evidence.  I argue that the suggestion that centuries of economic progress will be lost because of three lousy degrees – and that adaptation will be impossible – counts as an extraordinary claim.

But extraordinary evidence is starting to emerge.  Back in April, I wrote about an excellent paper by Katz et al (2024) that found economic damage of between 11% and 29% for a business-as-usual scenario.  For that paper I pointed out that forecasting backtests would have been useful and that the research may underweight the potential for humanity to adapt.

More recently, I’ve been reading the exemplary new working paper by Bilal and Känzig (2024) that reaches very similar conclusions to the Katz paper.  

The key methodological development pursued by Bilal et al is to combine global and local temperature shocks to explain changes in economic output.  They point out that global temperature correlates much more closely to the occurrence of extreme weather events than the local temperature shocks that had previously been the focus of the industry.  Incorporating these global shocks in a neoclassical economic damage model allows them to attribute a 12% GDP loss to each degree of global warming.  A “hothouse world” business-as-usual scenario would lead to a 29% present value welfare loss if this estimate is accurate.

Like with Katz et al, the econometric techniques employed in this paper are of the highest caliber.  I failed to find signs of thumbs anywhere near the scales.

For readers that may have missed it, I commented on the Bilal et al paper in my recent article on old-school climate predictions.  I suggested that Bilal and Känzig should have used emissions shocks as the basis for their analysis because we know that not all temperature variations are acts of God.  At that time I didn’t realize that this issue was actually covered in the Bilal et al article – they point out that emissions shocks are too slight to have any meaningful empirical effect on GDP.

Clearly evidence is emerging of significant economic damage caused by global warming.  Like any scientific evidence, of course, the findings are tentative and subject to review.  That said, my prior view that economic damage would more closely align with earlier, more modest findings – like those proposed by Nordhaus (2018) – is on increasingly shaky ground.