Humanity does not play DICE: A sensitivity analysis of technical progress
Abstract
The DICE model claims that technological progress can prevent global warming at low cost. However, its conclusions are predicated on highly optimistic and contested assumptions regarding the costs and effectiveness of backstop technologies, as well as the magnitude of climate damages. This study assesses the robustness of the DICE model's recommendations by exploring their sensitivity to these core assumptions. We employ a large-scale scenario-based analysis to account for the possible range of the climate damage function and the level of efficiency of future backstop technologies, as well as their financial costs.Our results demonstrate that relying on technology alone is a very risky bet ; under a wide spectrum of plausible, less optimistic assumptions, 82 % of scenarios lead to a climate-induced economic collapse.To address this fragility, we introduce a social lever into the model: economic sufficiency, defined as the policy option to produce less than the maximum production capacity. The inclusion of this lever dramatically alters the outcomes. Our study concludes that purely technological solutions are insufficient to tackle the climate crisis under uncertainty and that a dual strategy, combining both technological and social levers, offers a more resilient pathway.
Related
- AI and humanity's long conversation
- EQUIVARIANCE VERSUS FIRM NON-EXPANSIVENESS IN INCREMENTAL PROXIMAL GRADIENT METHODS FOR PLUG-AND-PLAY DENOISING
- Learning controls in mean field games: the fictitious play and related gradient descent
- Plug-and-Play Methods Provably Converge Even with Improperly Trained Denoisers: Convergence by Architectural Design
- On the Role of Tie-Breaking Rules in the Convergence of Fictitious Play for Symmetric First-Price Auctions
BibTeX
@article{ronsain2026humanity,
title = {{Humanity does not play DICE: A sensitivity analysis of technical progress}},
author = {Antoine Ronsain and Alexandre Gondran and Estelle Malavolti},
journal = {Ecological Economics},
year = {2026},
month = nov,
doi = {10.1016/j.ecolecon.2026.109089},
url = {https://enac.hal.science/hal-05854604v1},
note = {CC BY},
}
RIS
TY - JOUR
TI - Humanity does not play DICE: A sensitivity analysis of technical progress
AU - Antoine Ronsain
AU - Alexandre Gondran
AU - Estelle Malavolti
PY - 2026
DA - 2026/11/06
JO - Ecological Economics
DO - 10.1016/j.ecolecon.2026.109089
UR - https://enac.hal.science/hal-05854604v1
AB - The DICE model claims that technological progress can prevent global warming at low cost. However, its conclusions are predicated on highly optimistic and contested assumptions regarding the costs and effectiveness of backstop technologies, as well as the magnitude of climate damages. This study assesses the robustness of the DICE model's recommendations by exploring their sensitivity to these core assumptions. We employ a large-scale scenario-based analysis to account for the possible range of the climate damage function and the level of efficiency of future backstop technologies, as well as their financial costs.Our results demonstrate that relying on technology alone is a very risky bet ; under a wide spectrum of plausible, less optimistic assumptions, 82 % of scenarios lead to a climate-induced economic collapse.To address this fragility, we introduce a social lever into the model: economic sufficiency, defined as the policy option to produce less than the maximum production capacity. The inclusion of this lever dramatically alters the outcomes. Our study concludes that purely technological solutions are insufficient to tackle the climate crisis under uncertainty and that a dual strategy, combining both technological and social levers, offers a more resilient pathway.
SN - 0921-8009
SN - 1873-6106
ER -