Navigation auf uzh.ch
Navigation auf uzh.ch
Felix Kübler, Professor of Financial Economics, is improving the predictive accuracy of climate-economic models
A study improves calibration of key climate-economic models, potentially leading to more accurate predictions of climate change impacts and better-informed policy decisions.
Climate change poses a significant threat to our planet. Policymakers rely on economic models; these are integrated assessment models that are based on so-called climate emulators, as a basis for decisions about mitigation strategies. These climate emulators are simplified models used to simulate the relationship between CO2 emissions and global warming. They are a computationally efficient way to incorporate complex and resource-intensive climate calculations into economic models.
However, the predictive accuracy of these models relies on the individual setting of free parameters used for calibration to match observed data from more complex models. Setting these parameters is crucial for ensuring that the emulator accurately reflects the dynamics it aims to simulate.
Felix Kübler from the Department of Finance at the University of Zurich, together with researchers from the ETH Zurich and the University of Lausanne, and the University of Munich proposes a new calibration strategy for these climate emulators. Their approach aims to ensure that economic models of climate change are more closely aligned with the complex global climate models used by climate scientists. Their study was recently published in the Review of Economics Studies.
The researchers developed tests evaluating how well climate emulators perform in predicting the carbon cycle response to emissions and temperature changes due to increased CO2 concentrations. They applied the tests to the widely used Dynamic Integrated Model of Climate and the Economy (DICE-2016) by William Nordhaus and found that, while the model's basic structure is sound, its calibration needs significant improvement. For example, small variations in the choice of free parameters lead to significantly different predictions. When estimating the social cost of carbon (SCC), the externalized marginal discounted future cost of an extra ton of carbon emission, a crucial metric for determining optimal carbon taxation levels, the predictions varied by a factor of four, depending on calibration.
The authors propose a comprehensive and transparent calibration strategy for climate emulators used in economic models, based on two tests for calibrating the carbon cycle and temperature response and two tests for evaluating model performance. These allow for more accurate projections of climate change impacts. "Our research significantly improves the quality of climate-economic model predictions by ensuring they are properly calibrated against state-of-the-art climate science,” Felix Kübler explains.
For calibration the authors matched the outputs of earth system models in CMIP5 with the outputs of climate emulators, by implementing two of the most frequently used tests in CMIP5, the instantaneous release 100 gigatons of carbon to present day atmosphere and the instant quadrupling of C02 in the atmosphere.
By providing a more robust framework for calibrating climate-economic models, this research contributes to the ongoing effort to develop effective strategies for mitigating climate change. As policymakers, businesses, and individuals grapple with the challenges posed by a warming planet, having more accurate and reliable models will be crucial in making informed decisions and implementing effective policies.
Folini, Doris; Friedl, Aleksandra; Kübler, Felix; Scheidegger, Simon (2024). The Climate in Climate Economics.
Review of Economic Studies
More about the DICE-2016 Models by Nobel Laureate William Nordhaus: https://williamnordhaus.com/dicerice-models
More about the CMIP5 model, a product by the World Climate Research Program first established in 1995:
https://pcmdi.llnl.gov/mips/cmip5/
Photo Source: Nasa by Unsplash (Earth), Carolin Krajcir (Portrait)
