In April 2024, the prestigious journal Nature published a study that found that climate change would cause far greater economic damage by the end of the century than previous estimates had assumed. The conclusion made headlines around the world and was quoted and incorporated into central bank risk management scenarios.
On Wednesday, Nature withdrew the proposal, intensifying the debate about the extent of climate change's impact on society.
The decision came after a team of economists found problems with the data for one country, Uzbekistan, that significantly skewed the results. If Uzbekistan were excluded, they found, the damage would be similar to previous studies. Instead of a 62 percent decline in economic output by 2100, in a world where CO2 emissions continue unabated, global production would fall by 23 percent.
Of course, wiping out more than 20 percent of global economic activity would still be a devastating blow to human well-being. The paper's critics emphasize that climate change is a major threat, as recent meta-analyses have shown, and that more should be done to address it – but unusual findings should be viewed with skepticism, they say.
“Most people for the last decade thought that a 20 percent reduction in 2100 was an insanely large number,” said Solomon Hsiang, a professor of global environmental policy at Stanford University and co-author of the critique published in August. “So the fact that this paper states 60 percent is completely exaggerated.”
Retractions have become more common in recent years, according to Retraction Watch, an organization that tracks corrections in academic journals. However, they are still rare, accounting for about one in every 500 published articles.
Economists have long struggled with incorporating granular, sometimes subtle impacts of climate change into models that forecast far into the future, especially when combining them with something as complex as the global economy.
The study was led by Leonie Wenz, an economist at the Potsdam Institute for Climate Impact Research in Germany, and Maximilian Kotz, a postdoctoral fellow at the institute. The team developed several novel techniques to more comprehensively capture the economic impacts of climate change.
They used a carefully curated data set of economic conditions in geographic areas smaller than countries, such as states and provinces. They took into account not only average temperatures, but also a range of climatic conditions such as rainfall and heat waves. And they considered the effects of extreme weather events over a decade rather than assuming they dissipated quickly.
“We were trying to understand how long we could observe these effects in the data,” Ms. Wenz said. “This resulted in higher levels of damage compared to work that did not take these more persistent effects into account.”
It also produced a powerful comparison with the cost of avoiding catastrophic warming. Damage substantially contained over the next 25 years will cost six times the money needed to cut emissions enough to limit global warming to 2 degrees Celsius, the goal of the Paris climate agreement.
The paper's ambitious scope attracted the Network for Greening the Financial System, a network of mostly European central banks and financial regulators, as it updated a guide used to stress test whether banks would remain healthy in the face of increasing climate damage. After questions were raised, the organization added a disclaimer to the guide and said it would rely on a broader range of other research for future updates.
The article was also cited by the Organization for Economic Cooperation and Development and was in the top 5 percent of journal articles tracked by Altmetric, a tool for measuring research impact. Carbon Brief, a climate-focused news outlet, noted that it was the second most cited climate paper in 2024.
But the criticism increased. A few days after the researchers published their comment in the summer, another came from one of the authors' colleagues at the Potsdam Institute for Climate Impact Research, Christof Schötz. He focused on a common problem when working out the effects of one variable for many different geographical points: nearby places often behave similarly to one another. The authors' failure to take these correlations into account made the results so uncertain that they were essentially worthless, Mr. Schötz argued.
“The paper does not provide additional evidence of economic damage from climate change, nor can it serve as a basis for reliable future projections,” he said in an email.
In response to the counterarguments, Mr. Kotz and Ms. Wenz made corrections that they said changed the results only slightly, with a slightly larger margin of uncertainty and slightly smaller economic impact by the end of the century. “There are small quantitative changes, but these general qualitative messages are largely the same,” Mr. Kotz said.
They plan to revise and resubmit the paper.
Some researchers still believe it is possible that climate change could cause as much damage as the Potsdam researchers originally found. Scientists in the field are finding ways to incorporate more and more ripple effects—for example, the effects of wildfire smoke on respiratory health or how sea level rise affects home values.
Timothy Neal, an economist at the University of New South Wales, recently published an article noting that climate change will cause global disruptions that limit economic growth even in countries where the weather is less extreme. The results suggest a loss of 40 percent in 2100 if CO2 emissions remain high.
“Damage is most likely underestimated in the existing literature,” Mr. Neal said.
But climate economists are worried about maintaining their credibility, especially as the Trump administration defunds climate science and abandons efforts to cut emissions.
Lint Barrage, chair of energy and climate economics at ETH Zurich, believes the retracted paper has other methodological problems having to do with country-level inflation calculations that bias the results upward. Mr Kotz and Ms Wenz's correction note indicates that their results would be “significantly lower” if they used Ms Barrage's preferred approach, which they find frustrating.
“Sometimes, depending on your audience, it feels like big quotes are expected,” Ms. Barrage said. “If your goal is to advocate for climate change, you have crossed the line from scientist to activist, and why should the public trust you?”
As a remedy, some researchers recommend not trying so much in the first place. Noah Kaufman, a senior research scientist at Columbia University's Center on Global Energy Policy who worked in the Biden White House, believes studying specific questions — such as how to decarbonize while keeping electricity affordable — is more useful than forecasting macroeconomic impacts decades down the road.
“There are just a lot of examples in the world where we recognize that there are big risks, but we don't act like we can optimize our response to them,” Kaufman said. “We just try to avoid them in a sensible way.”