Photo: soupstock.in
Errors in data related to Uzbekistan significantly inflated estimates of climate change–related economic losses worldwide. This conclusion was reached by researchers at Stanford University, whose commentary was published by Nature on August 6 and reported by The Washington Post.
The commentary is the result of a review of a widely discussed scientific study published a year earlier. That study claimed that, due to climate change, global GDP would decline by 19 percent by 2050 and by 62 percent by 2100 compared to a hypothetical scenario without warming. These figures were three times higher than previous estimates and sparked significant debate. The article became the second most cited research in the media in 2024, and its data were used by the World Bank, the U.S. government, and other organizations for strategic planning.
However, Stanford researchers found that these drastic projections were largely the result of errors in the data for a single country — Uzbekistan.
Solomon Hsiang, director of Stanford’s Global Policy Laboratory and one of the commentary’s authors, explained that he and two graduate students discovered the error by deleting data country by country from the model. Most countries had only a minor effect on the projections. But when Uzbekistan was removed, the forecast shifted dramatically: the predicted drop in global GDP by 2100 fell from 62 percent to 23 percent, and by 2050 from 19 percent to 6 percent.
The team then closely examined the Uzbekistan data. According to the original study, the country’s GDP plummeted by nearly 90 percent in 2000, while in 2010, some regions showed growth of over 90 percent. Other years also showed extreme fluctuations. But data from the World Bank indicate that Uzbekistan’s economic growth over the past 40 years has been relatively stable, ranging from a 0.2 percent decline to growth of 7.7 percent.
These anomalies had a disproportionate impact on the model’s output because it linked changes in temperature and precipitation to economic growth. As a result, the model projected catastrophic climate impacts on global GDP.
Hsiang called the findings unexpected: “When you have so much data, it’s hard to imagine that a small country could have such a large effect.” He emphasized the importance of rigorous data verification.
The original study’s authors, from the Potsdam Institute for Climate Impact Research (PIK) in Germany, acknowledged the error in the Uzbekistan data. They attributed it to problems in processing the raw information and carried out an alternative analysis. In the revised model, the projected global GDP loss by 2050 was adjusted to 17 percent instead of 19 percent. The researchers stressed that “the study’s core conclusions remain valid and the changes in estimates are minor.”
Meanwhile, critics argue that the changes in methodology cast doubt on the original findings.
“Science doesn’t work by changing the conditions of the experiment to get the result you want. That approach contradicts the scientific method,” Hsiang noted.
Nature stated that it will take “appropriate editorial action” following the completion of its review.
