Nature journal retracts high-profile climate study after major data errors found for Uzbekistan

In its retraction published on 3 December, the journal stated that incorrect figures for Uzbekistan substantially skewed the study’s forecast of global economic losses.

A post-publication review revealed that the results were highly sensitive to the inclusion of Uzbekistan in the dataset, and that the country’s preliminary economic data for 1995–1999 contained serious inconsistencies. The authors acknowledged that these errors made it impossible to maintain the article in its current form.

The study’s alarming predictions of large-scale worldwide losses were ultimately traced to errors in statistics for just one country. Solomon Hsiang, director of the Global Policy Laboratory at Stanford University, identified the anomaly during experimental checks. He found that excluding Uzbekistan from the dataset fundamentally altered the final results.

The flawed model had projected that Uzbekistan’s GDP collapsed by nearly 90 percent in 2000, followed by extraordinary growth of more than 90 percent in some regions by 2010 – swings that contradict all official records. According to World Bank data, Uzbekistan’s economic performance over the past 40 years has been relatively stable, fluctuating between a 0.2 percent decline and 7.7 percent growth.

After recalculations, the study’s headline conclusions changed dramatically. The projected decline in global GDP by 2100 dropped from 62 percent to 23 percent, while the forecast for 2050 fell from 19 percent to 6 percent. Revised estimates now suggest that climate change may reduce global GDP by 17 percent by mid-century, rather than the previously claimed 19 percent.

The retracted article generated significant attention and was the second most-cited study in media outlets in 2024. Its findings had been referenced in high-level strategic planning, including by the U.S. government, the World Bank and other international organizations.