The ergodicity assumption in statistical approaches to climate damages
Projections of the economic impacts of climate change are essential to determine efficient levels of mitigation and adaptation. One common approach to projections is based on statistical models, regressing historically observed economic outcomes on climate or weather variables, without imposing explicit assumptions about the underlying mechanisms that relate the variables to each other. However, using these regression-based statistical models for projection relies on an implicit assumption of ergodicity, namely, that estimates can be generalized across space and time and that the state space of possible outcomes has been adequately explored in historical data. This paper explains how the ergodicity assumption enters the statistical approaches and discusses why it may be inappropriate in the context of damages from climate change. This article is part of the theme issue ‘Ergodicity and ergodicity breaking in the sciences’.
Authors
- Emilie Rosenlund Soysal (ORCID: https://orcid.org/0000-0001-7597-2718)
Institutions
- London Mathematical Laboratory (GB)
Publication Details
- Journal
- Philosophical Transactions of the Royal Society A Mathematical Physical and Engineering Sciences
- Published
- 2026-11-19
- DOI
- https://doi.org/10.1098/rsta.2025.0165
- Primary Topic
- Climate Change Policy and Economics
- Type
- article
- Field-Weighted Citation Impact
- 0.00