Will Climate Change Make Us Poorer? Lessons From a Retracted Nature Paper

The story of climate change is long and complex. Scientists have known for more than a century that carbon dioxide is a greenhouse gas, and by the late 1930s researchers were assembling the idea that human-driven increases in atmospheric CO2 could warm the planet. But it was not until the 1990s that climate change grew into the major political, economic and social issue it is today.

In 2024 a high-profile paper by Kotz, Levermann and Wenz, published in Nature, drew wide attention by estimating very large long-run economic damages from climate change. The paper argued that global GDP per capita would be about 19% lower by 2050 and as much as 60% lower by 2100 under a high-emissions pathway (RCP8.5) compared with a counterfactual without climate change.

From Influence To Retraction

Because of its stark findings, the study was quickly cited by major fiscal and financial institutions — including the U.S. Congressional Budget Office, the OECD, the World Bank, the UK Office for Budget Responsibility and the Network for Greening the Financial System (a consortium of central banks). Those organisations used the paper's damage estimates to inform policy analysis and prudential guidance.

On 3 December 2024, Nature retracted the paper, saying the data and methodological problems were "too substantial for a correction." Retractions in leading journals are rare and raise questions about peer review and editorial safeguards.

What Went Wrong — And What Remains Unclear

My review of the article and its available data suggests several issues worth distinguishing:

• Authorship and approach: The authors are primarily physicists and a mathematician who applied econometric techniques to large economic datasets. Their methods relied on statistical model-fitting rather than standard economic modelling developed by applied economists.
• Data and coding errors: Nature's retraction cited substantive data and methodological errors. Those specific errors were judged severe enough that a simple correction was insufficient.
• Modelling choices matter: Reasonable, defensible decisions in empirical work — for example, whether to convert GDP per capita using market exchange rates or purchasing-power-parity (PPP) rates, whether to treat historical patterns as persistent into the future, and how to adjust for major shocks such as the 2008 financial crisis — can materially change projected outcomes.
• Mechanisms were under-explained: The paper used multiple climate indicators (mean temperature, daily temperature variability, precipitation measures and extremes) but offered limited plain-language explanations for how those variables would produce the dramatic economic declines it projected, especially given that global temperatures have already risen roughly 1°C while economic growth has continued.

Why The Results Were Contested

Good empirical research makes explicit how sensitive results are to alternative, reasonable assumptions. The retracted paper produced remarkably large damage estimates, but those numbers did not appear to be uniquely determined by the data — different modelling decisions could plausibly yield much smaller damages, or very different spatial and temporal patterns of impact. That does not mean the alternatives are correct, only that the original conclusions were not robustly established.

More broadly, the episode highlights two important lessons for policy-relevant science: (1) high-impact claims require exceptionally rigorous data and transparent code; and (2) peer review and editorial processes must be attentive to cross-disciplinary methods where specialists from different fields apply unfamiliar tools.

What This Means For Policy And Public Debate

This single paper does not settle the question of climate change's economic costs. A large and growing literature uses diverse methods — integrated assessment models, sectoral analyses, regional case studies and reduced-form econometrics — and arrives at a range of estimates. Policymakers should consider the full body of evidence, evaluate assumptions explicitly, and stress-test conclusions against alternative, defensible modelling choices.

Finally, scientific integrity depends on following the evidence rather than fitting evidence to a preferred narrative. The retraction is an important corrective: it reminds readers and decision-makers to demand transparency, robustness checks, and clear explanations of mechanisms before adopting dramatic policy conclusions.