Current projects
Downscaling of physical risks for climate risk scenario design.
In this strand of work we use machine learning to downscale global climate model outputs to spatial resolutions of interest to market participants and asset owners in different regions and urban clusters of the world. We aim to deliver simple, yet effective risk scores based on aerial surface experiencing changes in the average or tail risk profile of target variables. Work in collaboration with the Monetary Authority of Singapore and the founding partners of the Singapore Green Finance Centre applies the methodology to Southeast Asian exposures.
Climate risk scenarios for asset allocation and stress testing.
We use a combination of AI and statistical tools to refine our understanding of climate risk disclosures and financed emissions with a view toward improving disclosures and detecting greenwashing. In addition to estimation of current indirect emissions, we study the projection of indirect emissions along different climate pathways, thus offering invaluable insights into the effectiveness and robustness of net-zero strategies.