My poster presentation for “Online probabilistic estimation of carbon beta and carbon Shapley values for financial and climate risk” at the 18th FINANCIAL RISKS INTERNATIONAL FORUM (organized by Institut Louis Bachelier/Fondation du Risque/Europlace Institute of Finance). This year’s focus Shaping Financial Research: Data, AI and New Challenges.

This study explores climate-related financial risks and introduces innovative methods for estimating carbon beta and carbon Shapley values. Carbon beta quantifies how stock returns react to a portfolio called Brown Minus Green (BMG), which takes a long position in less climate-friendly (brown) stocks and a short position in climate-friendly (green) stocks. Shapley values, derived from game theory, provide an additive framework to interpret how input factors influence a supervised model’s output. The study presents novel approaches for time-varying estimation of carbon beta and carbon Shapley values. These methods are forward-looking, nonparametric, nonlinear, and adaptive. Unlike traditional approaches, they do not assume a fixed “true” value for carbon beta based on a linear relationship between individual stock returns and global market returns, and they leverage conformal prediction to quantify uncertainty in the estimates of carbon beta and carbon Shapley values:

https://www.researchgate.net/publication/389953574_Online_probabilistic_estimation_of_carbon_beta_and_carbon_Shapley_values_for_financial_and_climate_risk