Researchers Valentina Bosetti, Laura Diaz Anadon, and Stefan Feuerriegel published a study in Nature Energy on September 8, 2026, detailing how causal machine learning can assess clean technology policies. The peer-reviewed work outlines methods to evaluate interventions such as carbon pricing and public subsidies. According to the study, causal machine learning enables policymakers to compare alternative policy scenarios and identify where, how, and for whom cleantech measures deliver results.
The authors argue that designing targeted, evidence-based rules requires understanding how specific interventions affect both environmental and socioeconomic outcomes. The research brings together specialists across Europe and the United States. Participating bodies include Bocconi University’s Department of Economics and the Euro-Mediterranean Center on Climate Change in Milan, Italy. Additional contributors represent the University of Cambridge, the Harvard Kennedy School’s Belfer Center, and LMU Munich alongside the Munich Center for Machine Learning.
Nature Energy credited Bernardino D’Amico for contributing to the peer review of the work, and the authors declared no competing interests. The paper cites earlier methodological and empirical work across energy policy, econometrics, and causal inference, including studies by Susan Athey, Stefan Wager, Donald Rubin, and Guido Imbens.
Springer Nature charges $39.95 for instant access to the standalone article PDF. A 30-day online access pass through Nature+ costs $32.99, while an annual subscription providing 12 digital issues costs $119.00, or $9.92 per issue.
